Gelderman2003PhDthesis PDF

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/254760062
A portfolio approach to the development of differentiated

purchasing strategies
Article · January 2003
CITATIONS READS
50 282
1 author:
Cees J. Gelderman
Open Universiteit Nederland
62 PUBLICATIONS 2,257 CITATIONS
SEE PROFILE
All content following this page was uploaded by Cees J. Gelderman on 23 December 2020.
The user has requested enhancement of the downloaded file.

A Portfolio Approach to the Development
of Differentiated Purchasing Strategies
A portfolio approach to the
development of differentiated
purchasing strategies
Cornelis Johannes Gelderman
Eindhoven University of Technology

ISBN 90-386-1678-3
NUR 800
CIP-DATA LIBRARY TECHNISCHE UNIVERSITEIT EINDHOVEN
Gelderman, Cornelis J.
A portfolio approach to the development of

differentiated purchasing strategies/
by Cornelis Johannes Gelderman. — Eindhoven:
Technische Universiteit Eindhoven, 2003.
— Proefschrift. —
Key words: Purchasing/Purchasing strategy/Portfolio

management/Supplier relations
Printed by: Universiteits Drukkerij/Technische

Universiteit Eindhoven
Cover design: Paul Verspaget, UD/TU/e
© 2003 C.J. Gelderman, Maastricht
All rights reserved. No part of this publication may be

reproduced, stored in or introduced into a retrieval
system or transmitted, in any form, or by any means,
electronic, mechanical, photocopying, recording or
otherwise, without the prior permission of the author.
A Portfolio Approach to the Development
of Differentiated Purchasing Strategies
Proefschrift
ter verkrijging van de graad van doctor

aan de Technische Universiteit Eindhoven,
op gezag van de Rector Magnificus, prof.dr. R.A. van Santen,
voor een commissie aangewezen door het College voor Promoties
in het openbaar te verdedigen op vrijdag 12 september 2003 om 16.00 uur
door
Cornelis Johannes Gelderman
geboren te Schiedam
Dit proefschrift is goedgekeurd door de promotoren:
prof.dr. A.J. van Weele

en
prof.dr. R.S.J. Tuninga
Contents
Acknowledgements 13
Chapter 1 Introduction to the study 15

1.1 Developments in purchasing management 15
1.2 Need for differentiation 18
1.3 The portfolio concept 20
1.4 From Pareto to Kraljic 22
1.5 Statement of the problem 25
1.6 Relevance of the stated problem 28
1.7 Methodology 29
1.8 Structure of the thesis 33
Chapter 2 Portfolio approaches in related disciplines 35

2.1 Investment theory: the roots of all portfolio management 35
2.1.1 The Markowitz model 35
2.1.2 Relevance for purchasing: what can we learn? 37
2.2 Portfolio models in strategic management 38
2.2.1 The BCG-approach 38
2.2.2 Assumptions, criticism and adoption 41
2.2.3 Other, multifactor portfolio models 43
2.2.4 Conclusions 46
2.3 Portfolio models in marketing management 48
2.3.1 New product portfolio models 48
2.3.2 Classifying products: from product
evaluation to product positioning 50
2.3.3 Classifying customers: from segmentation
to customer portfolios 52
2.3.4 Customer profitability models 54
2.3.5 Customer relationships models 57
2.4 Concluding remarks 61
Chapter 3 Portfolio approaches in purchasing management 63

3.1 The Kraljic portfolio approach 63
3.2 Discussion of the Kraljic approach 67
3.2.1 Dimensions 67
3.2.2 Categories 70
3.2.3 Strategic recommendations 71
3.2.4 Use issues 73
3.3 Elaboration of purchasing strategies for all categories 75
3.4 Supply positioning analysis 78
3.5 Cunningham’s interaction approach to portfolio analysis 82
3.6 The Olsen and Ellram approach 83
3.6.1 Description of the model 83
3.6.2 Discussion of Olsen and Ellram's model 86
3.7 Bensaou's specific investments approach 88
3.7.1 Description of the model 88
3.7.2 Discussion of Bensaou's model 91
3.8 Conclusions of the literature study: addressing
the research questions 93
3.8.1 Answers to the first research question:
differences and similarities 93
3.8.2 Elaboration of the second research question:
explaining the use of a portfolio approach 97
3.8.3 Elaboration of the third research question:
the practice of experienced professionals 103
3.8.4 Elaboration of the fifth research question:
power and dependence in the matri 104
Chapter 4 A power-dependence perspective 107

4.1 Rotating the Kraljic-matrix: new perspectives 107
4.2 Power and dependence in buyer-supplier relationships 110
4.2.1 Dependent: to be or not to be? 110
4.2.2 A resource perspective: RBV and RDT 113
4.2.3 Power: definitions and associations 115
4.2.4 Relative power and total interdependence 118
4.3 A mutual dependence-based purchasing portfolio model 120
4.4 Determinants of dependence 122
4.4.1 Conceptual studies to the determinants of dependence 123
4.4.2 Empirical studies to the determinants of dependence 126
4.4.3 Towards a conceptual model 128
4.5 Managing dependence 133
4.5.1 Resource dependence theory 133
4.5.2 Transaction cost theory 134
4.5.3 The Kraljic approach 135
4.5.4 Critical views on power-dependence strategies 137
4.6 Answers to and further elaboration of the fifth research question 141
Chapter 5 The cases studies 145
5.1 Methodology 145
5.2 DSM 148
5.2.1 Organization of purchasing 149
5.2.2 Purchasing and supply strategies 150
5.2.3 Purchasing portfolio analysis: dimensions,
measurement and use 153
5.2.4 Portfolio-based strategies 154
5.3 Akzo Nobel Coatings 162
5.4 TE STRAKE - Engineering & Production 170
5.5 Reactions to the dependence model 178
5.6 Answers to the third research question: comparison
of the case studies 180
5.6.1 Specific business context 181
5.6.2 Measurement and use 183
5.6.3 Strategic directions 187
5.7 Conclusions 194
Chapter 6 The survey 197

6.1 Introduction and objectives 197
6.2 The conceptual models 198
6.3 Relationship-dependence profiles: a condition-
seeking research strategy 200
6.4 Population and sample 202
6.5 Design and procedure 203
6.6 Measurement and operationalization 206
6.6.1 Variables and constructs in the use-model 207
6.6.2 Variables and constructs relating to buyer's
and supplier's dependence 208
6.6.3 Variables and constructs relating to other conditions 210
6.7 Methodological concerns 214
6.7.1 Validity 214
6.7.2 Buyer’s perceptions 215
6.7.3 Nonresponsev 215
6.8 The results of the survey 217
6.8.1 Response 217
6.8.2 Answers to the second research question:
explaining the use of the portfolio analysis 223
6.8.3 Users versus non-users 228
6.8.4 Answers to the fourth research question:
conditions for the selection of strategies 234
6.8.4.1 Profiles in the bottleneck quadrant 237
6.8.4.2 Profiles in the non-critical quadrant 239
6.8.4.3 Profiles in the leverage quadrant 241
6.8.4.4 Profiles in the strategic quadrant 244
6.8.5 Answers to the fifth research question:
power and dependence in the matrix 248
6.8.5.1 Power and total interdependency in the matrix 249
6.8.5.2 Determinants of dependence 251
6.9 Conclusions 254
Chapter 7 Conclusions and recommendations 257

7.1 Conclusions and main findings 257
7.2 Reflection on the methodology 268
7.3 Recommendations for business 272
7.3.1 Adopting a portfolio approach 272
7.3.2 A 5-step approach to the application of
purchasing portfolio analysis 275
7.3.3 Conditional dynamics in the matrix: managing
power and dependence 279
7.4 Recommendations for future research 284
References 291
Appendices 315
A Interview guide for the case studies 315
B Sources of the case studies 317
C Reactions to the mutual dependence model 319
D A 5-step approach to the application of purchasing
portfolio analysis 322
E Action research: using the outline in practice 323
F Elaboration and overview of the 9 scenarios 337
G Questionnaire (in Dutch) 338
H Measurement and operationalizations of
organizational dependence 355
Summary 361
Samenvatting (in Dutch) 368
About the author 375

13
Acknowledgements
Over the past four years I have been in the fortunate position to conduct research in my
favourite area: purchasing management. It was time for me to take the next step in my
personal and intellectual development. I would like to thank everyone who has
stimulated and supported me to initiate and complete this study. I wish to acknowledge
the helpful contributions of many people and companies.
First of all, I would like to thank professor Arjan van Weele for his indispensable
encouragements and for his continuous inspiration. There is an old saying in negotiation:
if you aim higher, you come out better. Behold, Arjan’s supervision in a nutshell. I
enjoyed the freedom and responsibility he bestowed on me, to pursue the research my
own way. In addition, Arjan acted as a gatekeeper, introducing me to many interesting
purchasing academics and purchasing practitioners. I would like to thank Arjan for the
many lessons and for his worldly wisdom. It was an honour and a privilege to be
supervised by Arjan, one of the founding fathers of modern purchasing and supply chain
management. Thank you, Arjan!
Secondly, I wish to thank professor Ron Tuninga for expressing his confidence in me at
the initial phase of the research project. During the project Ron was the ultimate
stimulator, contributing his research expertise and infectious enthusiasm. Thank you,
Ron!
Also, I wish to thank professors Janjaap Semeijn and Martin Wetzels for their valuable
contributions in the final stage of the research project.
Furthermore, I would like to express my gratitude towards the research council (NRS) of
the Dutch Association of Purchasing Management (NEVI) for their substantial support to
this research project. In particular I would like to thank Arno de Schepper (chairman of
NRS), Geert Brouwer (NRS), and Jan Willem van der Meer (chairman NEVI). A special
thanks goes to Yvonne Peter and Ella Gelauf (both NEVI) for their aid and assistance in
administering the initial and follow-up mailing.
During the past four years many purchasing practitioners have invested valuable time
and energy in sharing their experiences. I wish to thank Willem van Oppen (ex-DSM),
Dick Bartelse (Akzo Nobel Coatings), and Wilfred van der Made (ex-Te Strake) for their
roles as key informant in the case studies. The study to the practice of experienced
practitioners would not have been possible without their friendly collaboration.
Interviewing these purchasing experts has contributed significantly to my understanding
of the possibilities of a portfolio approach in purchasing management. I have felt a deep
respect for these pioneers, their insights have proved to be invaluable for the entire
research project.
The survey procedure included a pilot test to pre-test the questionnaire by a limited
number of respondents. I would like to thank the key informants of the pilot study for
their efforts and friendly cooperation which contributed significantly to the success of the
survey study: Marcel Roost (GTI), Edwin Noordman (ex-EMTEC), Fréderique van der
Burg-van Baarle (DSM) and Bert van den Heuvel (Akzo Nobel). We have recognized that
the questionnaire was extremely long, containing more than 175 questions. Therefore, I
would like to thank all the respondents who were willing to complete my ‘absurd’
questionnaire.
There are several fellow academics who supported me in various ways, providing
inspiration, professional experience, and most of all friendship. In this respect I would
like to thank Bé Albronda for his support in the early phases of the project, Luitzen de
Boer for collaborating on our practitioner paper, Wim Braakman for his refreshing sense
of perspective, Ivo de Loo and Frits Schreuder for their assistance with some of the
statistical procedures, and Paul Ghijsen and Marjolein Caniëls for being the best
colleagues one could ever wish for.
A special mention to Søren Vammen, director of the Danish Purchasing & Logistics
Forum, who invited me twice to submit a paper for publication in his journal. Thanks for
your e-mail in which you asked me questions like ‘do you feel lucky – punk?’ In reply:
yes, Søren, sometimes you have ‘made my day’.
Sadly, it is not possible to share the excitement and the delight of taking my PhD with my
parents, Cis and Wim. They would have been very proud of their son, as proud as
parents could ever be. I thank them for their continuous support.
Writing a dissertation is practically impossible without a supportive home base to fall

back on. For that I am most grateful to my wife Catharina. She convinced me constantly
that I was on the right track, making progress at a proper pace. And last but not least, I
would like to thank Sander and Eline, our wonderful kids, for their infectious zest for
life.
Cees J. Gelderman
Maastricht, April 2003
15
1 Introduction to the study
For long, purchasing was considered as a relatively unimportant clerical function in

business. Not earlier than the latter half of the 20th century were there any indications
that the importance of purchasing started to be recognized. Especially driven by the 1973
Oil crisis, top management started taking an interest in the importance of suppliers and
supply management (e.g. Kraljic, 1983; Leenders and Fearon, 1993). Over the last 30
years, the role of purchasing has changed dramatically, from a clerical, administrative
function into a strategic function that contributes to the competitive advantage of
companies. This ‘revolution in purchasing’ has lead to tremendous changes in the scope,
the impact and the responsibilities of purchasing management (Van Weele and
Rozemeijer, 1996). Major developments in purchasing management all point at the
importance of the management of supplier relationships.
In accordance with an increased focus on relationship management, there is a growing
acceptance and use of purchasing portfolio approaches, such as the Kraljic approach.
Purchasing portfolio analysis had subsequently become “the dominant approach to what
the profession regards as operational professionalism”, according to Cox (1997: 270).
Boodie (1997) for instance reported a peak of 80% portfolio users in mass production
companies. However, most publications on the portfolio approach are conceptual or
anecdotal by nature. In contrast to an increased adoption, there is a lack of theoretical
foundation and empirical evidence. This study addresses the gap between the lack of
academic research into purchasing portfolio models and their growing adoption by
practitioners.
1.1 Developments in purchasing management

A large body of research supports the idea that professional purchasing has changed in
the last four decades. Several authors have forecasted and described changes for the
purchasing role in companies. Reviewing all these contributions is beyond the scope of
this introduction. There are studies to the history and evolution of:
- the purchasing function (Syson, 1992; Monckza et al., 1998)
- purchasing and supply management (Trent and Monckza, 1998)
- strategic purchasing (Ellram and Carr, 1994)
- purchasing positions (Pooley and Dunn, 1994).
The act of purchasing can be considered as one of the oldest basic types of human
behavior. Webster’s Third New International Dictionary (1993: 1844) defines ‘to
purchase’ as “to get into one’s possession by paying money or its equivalent”. In a
16
business context, purchasing activities are indispensable for any company. However, it
was not until the end of the Early Years of purchasing history (1850-1900) that
Purchasing started to be organized as separate corporate functions (Monckza et al., 1998).
For long purchasing was looked upon primarily as an administrative clerical function.
The American Management Association for instance, recommended in 1931 that the
model purchase department should address the following aspects: administration,
ordering, payments, accounting, inspection, and salvage (Syson, 1992). Purchasing
gained importance during times of scarcity in the Twentieth Century, emphasizing the
need for obtaining the required materials. The purchasing function was seen as a part of
the umbrella of materials management. The scope of purchasing activities was limited to
meet the needs of the manufacturing function or other internal function for which it was
buying. It was not the responsibility of purchasing to question those needs, forge long
term relationships with suppliers, or to understand the needs of the end customer
(Ellram, 1998).
The field of purchasing has evolved significantly over the past 30 years. More and more it
is expected that a purchasing department could and should contribute to a firm’s
efficiency and competitiveness. Just-in-time management entered western industry,
providing a completely new view on the role and management of suppliers. The need for
flexibility and customization has promoted component and modular sourcing (Syson,
1992). The outsourcing trend, focussing on core activities and core competences, has
added to the increasing importance of a competitive supply base management. Suppliers
are considered as critical sources of product and process technology. The pressure to
innovate, by including the latest technology in product designs, makes supplier
contribution increasingly vital (Trent and Monczka, 1998). Tully (1995: 46) described
“purchasing’s new muscle” in terms of leveraging buying power (centralized buying),
but especially in terms of forming enduring partnerships with suppliers. A new type
purchasers “show they can add millions to the bottom line”.
In the nineties there is much support for the idea of shifting from a traditional
antagonistic approach towards a more collaborative approach of suppliers (Matthyssens
and Van den Bulte, 1994). Partnerships sourcing is said to be superior to adversarial
competition, because it leads to long-term collaboration based on trust (MacBeth and
Ferguson, 1994). Partnership Sourcing Ltd., the U.K. body explicitly promotes the idea of
‘partnership sourcing’: “…where customer and supplier develop such a close relationship
that the two work together as partners. (…) Partnership sourcing will lead to a win-win
situation, because both partners have an interest in each other’s success”. The
management of relationships across the supply chain is increasingly being referred to as
supply chain management (Lambert and Cooper, 2000). Supply chain analysis is closely
related to network analysis, structuring inter-organizational relationships (Lazzarini et
al., 2001). The emphasis on interorganizational relationships and their embeddedness in
A Portfolio Approach to the Development of Differentiated Purchasing Strategies 17
networks has its roots in the work of a group of mainly European researchers in the area
of industrial marketing and purchasing, the IMP Group (Håkansson, 1982).
Obviously, partnership sourcing is only one concept creating change in purchasing
(McIvor and McHugh, 2000). Extensive change is also created by lean production
(Womack et al., 1990), lean supply (Lamming, 1993), buyer-supplier process integration
(Christopher et al., 1992), early supplier involvement (Håkansson and Erikkson, 1993),
outsourcing (Quinn and Hilmer, 1994), supply base reduction (Homburg, 1995), total cost
management (Ellram, 1996), supplier development (Hahn et al., 1990; Krause and Ellram,
1997), cross functional teams (Maltz and Ellram, 1999), supply base rationalization
(Cousins, 1999), corporate advantage in purchasing (Rozemeijer, 2000), supply chain
management (Lambert and Cooper, 2000), supply chain integration (Frohlich and
Westbrook, 2001), and e-procurement (Neef, 2001; Essig and Arnold, 2001).
Nowadays, there is a growing consensus that purchasing is an important contributor to a

firm’s competitive advantage. Carter and Narasimhan (1996) posited that the purchasing
function has a strategic impact on the firm, contributing to both the overall corporate
performance and the performance of other functions. In their words: “Purchasing is
indeed strategic”. An appropriate sourcing strategy can contribute significantly to
profitability, market share, and technological innovation (Hartmann et al., 2001).
According to Monczka et al. (1998) there are three major benefits that can be achieved by
a stronger focus on purchasing: (1) cost savings, (2) impact on quality, and (3)
contribution to technology development and improvement of product and process
designs. Cost-reduction programs should result in cost reduction and/or cost avoidance
(Ellram, 1998). In general, there will be more possibilities for cost reduction for
companies with relatively high purchasing cost. In manufacturing industries the value of
purchases is significantly higher than in the service sector (Ellram, 1998). In the
automotive and electronic industries, typically between 60% and 80% of the product
value has been outsourced to suppliers (Schary and Skjøtt-Larsen, 2001). Mol et al. (2001)
documented the main trend in the Netherlands over the 1990’s to source more from
external suppliers. While process-based industries increased external sourcing only
slightly, the OEM industries, like electronics, vehicles and machinery, strongly increased
their dependence on external suppliers. The rising share of purchasing cost indicates a
natural area for potential cost savings. Purchasing has an important impact on quality
too, since companies concentrate more on their core competences (Prahald and Hamel,
1990). Shortened life cycles and global competition are making business too complex and
expensive for one firm to go it alone (Ellram, 1998). The need to concentrate on core
competences has lead to an increased emphasis on outsourcing (Venkatesan, 1992), which
at the same time increased the supplier’s influence on quality.
18
Managing supplier relations is increasingly considered as a strategically important

activity for the firm (Mol, 2002). Moreover, the management of the supply chain has
become one of the key issues for many companies. The increased need for buyer-supplier
integration is provoked by end markets that require reduced development and delivery
times. Caglinano et al. (2002) clearly distinguished two different dimensions of buyer-
supplier integration: technological integration and operational integration. Technological
integration refers to collaboration in designing and developing new products. Many firms
realize that early supplier involvement in new product development is beneficial with
regard to the costs and the quality of new products. Early supplier involvement can
reduce development times, which can be crucial for getting to the market quickly.
Consequently, more and more suppliers are becoming involved in their customer’s
development projects (e.g. Wynstra and Ten Pierick, 2000). The supplier thus becomes a
strategic resource in the design process. Operational integration instead refers to activities
such as planning, production and delivery. There is a wide set of techniques that can be
traced back to the principles of Just-In-Time management, obtaining operational and
logistical integration between customers and suppliers (e.g. Lamming, 1993). In
competitive end markets, companies are pressed for short and reliable production and
delivery schedules (fulfillment), which implies the disposal of a lean and flexible supplier
network. As a consequence of the reduced times allowed by end-markets there is an
increased need for supply chain integration. The reduction of the number of suppliers
can be necessary in light of the logistical complexity and customer’s demands.
Purchasing should develop and maintain a world class supplier network, since a firm’s
ability to compete on its end markets is considerably influenced by its suppliers’
capabilities (e.g. Hahn et al., 1990).) To conclude, purchasing’s new role is to contribute to
supplier development and relationship management, identifying, developing and
managing new and existing suppliers. Purchasers can focus their efforts on managing
suppliers as opposed to placing orders and expediting (Lambert and Cooper, 2000).
In theory there are many different types of buyer-supplier relationships. In practice
companies need a variety of relationships, each providing its different benefits.
Purchasing professionals should have the capacity to cope with a variety of relationships,
to be handled in differentiated ways. Purchasing portfolio analysis is considered as a
particularly useful tool, developing and selecting differentiated purchasing and supplier
strategies.
1.2 Need for differentiation

Matthyssens and Van den Bulte (1994) observed a general shift in organizational buying
behavior, from an antagonistic mode to a more co-operative mode. Since the nineties,
long term collaborative relationships with suppliers are generally considered to be the
(next) source of competitive advantage. Then again, there are critical remarks too,
referring to the ‘myth of partnership’ (Van Weele, 2001: 164) and the ‘case against
partnerships’ (Ramsay, 1996a:13). A partnership is the result of continuous effort on both
sides, it is not a technique which can be adapted. This could explain the rather small
number of really successful partnerships (Van Weele, 2001). Ramsay (1996a) suggested
that partnerships are frequently only appropriate for a minority of a company’s
purchases and that it is arguable that partnerships are only advisable for very large
companies. Lamming (1993: 238) and Lamming and Harrison (2001: 597) have observed
that in practice “the so-called partnership often relies on customer dominance”.
Competitive relationships do not necessarily involve lower trust and adversarial
behavior, according to Parker and Hartley (1997). ‘Competition’ may be more effective
than ‘cooperation’ in many buyer-supplier relationships (Forker and Stannack, 2000). It
can be argued that companies should pursue both competitive and cooperative strategies
simultaneously (Cox, 1995; Lado et al.,1997; Parker and Hartley, 1997). The main point
here is that a firm should develop long-term relationships with a relatively small group
of key suppliers, not with all suppliers.
Nowadays, there is a strong believe and a general consensus that world class purchasing
includes building and sustaining strategic partnerships with superior suppliers. This
conclusion is supported by a number of empirical studies comparing Japanese business
practices to those of the rest of the world (see for instance Cusmano and Takeishi, 1991;
Dyer and Ouchi, 1993; Helper and Sako, 1995). Dyer (1996) found empirical evidence for
a positive relationship between interfirm specialization and performance in the Japanese
and the American auto industry. A conclusion might be that Japanese firms manage their
suppliers primarily by partnerships, in contrast with their American colleagues. Contrary
to this belief, Bensaou (1999) empirically showed that in the auto industry Japanese firms
conduct their business with a smaller ratio of strategic partnerships (19 percent) and
make extensive use of market exchange (31 percent). Similarly, some 25 percent of U.S.
automakers engage in market exchange, and another 25 percent have developed
mutually committed relationships. Strategic partnerships create new value, however they
are costly to develop, nurture and maintain. In addition they are risky, given the
specialized investments they require. Moreover, benefits from a relationship could
decrease over time. The relationship benefits may peak when both parties are working on
new solutions, but diminish once the new solution has been put into place (Christopher
and Jüttner, 2000).
Companies need a variety of relationships, each providing its different benefits, where no
general ‘best’ type of relationship exists (e.g. Young and Wilkinson, 1997; Gadde and
Snehota, 2000). Axelsson et al. (2002) have stated that much of the debate in the area of
purchasing and supply management has focused on two opposite purchasing
approaches: transaction-oriented and relationship-oriented behavior. However, the
authors emphasized that the two approaches are complementary: a firm can adopt
different approaches for different suppliers. This underlines the proposition that
20
differentiation is needed in managing supplier relationships. Relationships require

different mixes of cooperation and competition.
From a supply chain perspective Hines et al. (2000) described seven types of value:
customer responsiveness, timely supply, high quality goods and services, efficient
operating processes, lower prices, impact on profit and highly innovative. The
management of supplier relationships should be tailored to those values that are in line
with the overall business value strategy (Treacy and Wiersema, 1993). The outcomes of
relationships may range from cost savings through joint product development. Once the
focus (output) is decided, the appropriate relationship can be developed (Cousins, 2002).
Obviously, not all suppliers are to be dealt with in the same way. This places purchasing
managers for the task of developing and executing a set of differentiated supplier
strategies. The need for differentiated supplier strategies requires some sort of
classification (Lilliecreutz and Ydreskog, 1999). Therefore, in advance a portfolio model
for supplier relationships appears beforehand to be a useful tool. Effective purchasing
and supply management requires the selection of strategies that are appropriate to the
prevailing circumstances.
1.3 The portfolio concept

A portfolio refers to a collection of different but connected items. The items may be
objects or subjects. In general, the portfolio concept focuses on the interdependencies
among management decisions and emphasizes an integrated approach (Turnbull, 1990).
The portfolio concept stresses the importance of the whole rather than the parts. It reflects
the importance for balance in a collection of individual elements. As a consequence, it
allows for differentiation and diversification, in our aim for balance and an optimal use
of limited resources. The portfolio concept has its roots in financial investment in the
1950s. For business purposes, portfolio approaches have been developed for applications
in investment theory, strategic management, marketing, and purchasing management.
Nowadays, portfolio thinking, including different kinds of portfolios, is enjoying an
increased popularity. For instance, Donaldson (2000) proposed an organizational
portfolio theory which treats the organization as a portfolio of causes of organizational
performance; Kirchhof et al. (2001) and Graves et al. (2000) discussed some models for
the management of R&D portfolios; Archer and Ghasemzadeh (1999) discussed a
framework for project portfolio selection, and McLagan (2000) promoted personal
portfolio management, mixing and managing the outputs and competences of employees.
When limited to more main stream practices, portfolio theory is essentially concerned
with the collective returns from the use of assets together with their possible
redistribution over various options at the discretion of management. How ‘assets’,
‘returns’ and ‘options’ are defined depends on the area of application (Yorke, 1984: 9). In
this study we define a portfolio as:
‘a collection of different items, objects or subjects that are connected to each other’
Portfolio methods are widely taught and widely used by consultants and decision
makers. There is a broad consensus that portfolio models should be used with an
understanding of their limitations and perhaps in combination with other tools (Olsen
and Ellram, 1997). In much literature matrix portfolio methods are viewed as a useful
starting point for (strategic) analysis. Still, many practitioners and researchers have
strong reservations about the use of and the premises behind portfolio methods. They
point out that there is a general lack of theoretical and empirical support for these
techniques. Not many authors would go as far as Armstrong and Brodie (1994: 84) who
concluded that: “Until contrary evidence is produced, we advise against using matrix
methods under all circumstances.” On the contrary, there is a general feeling that
portfolio methods form a useful tool for various management decisions.
At this point we would like to stress that any criticism of a technique should not be that it
simplifies but rather that it focuses on unimportant factors. After all, the logic for
(portfolio) techniques is in the first place that it constitutes a tool for management so that
complex problems can be simplified and solved in an acceptable way. Coate (1983)
concluded that the usefulness of any strategy generated by a portfolio model, depends
critically on the validity of the assumptions. Wensley (1994) posed two main questions
for evaluating management tools:
1. To what extent is the approach based on assumptions which themselves are
empirically valid?
2. To what extent does the approach help to improve the quality of decision-making?
Portfolio approaches are used for management problems in various fields and disciplines.
All portfolio models have their roots in investment theory. The work of Markowitz in the
early 1950s is the origin of modern portfolio theory for investment purposes. Balancing
the objectives of high yield and low risk, the portfolio approach focuses on the efficient
allocation of limited resources. In the 1970s and 1980s a great number of portfolio models
were developed in other areas of business administration, notably strategic management
and marketing management. In business administration a portfolio approach is a way of
looking at and dealing with (management) problems by focussing on a small number of
important factors. The basic idea is the simplification of a complex problem.
In this study we define a portfolio model as:
‘a tool that combines two or more dimensions into a set of heterogeneous categories for which
different (strategic) recommendations are provided’.
22
Three basic elements are to be recognized in this definition: (1) dimensions, (2) categories,
and (3) strategic recommendations. The use of portfolio model implies the classification
of objects/subjects, usually presented in the form of a two dimensional matrix. The basic
idea is that the positions of the units on the grid or in the matrix should determine the
formulation of the most appropriate strategy (Yorke and Droussiotis, 1994). However,
models and tools that do not provide guidance for management decisions are merely
classification schemes, not portfolio models. In purchasing, classification tools can be
seen as the predecessors of the actual portfolio models.
1.4 From Pareto to Kraljic

Organizations usually have to deal with a large number of products and a variety of
suppliers, to be treated in different ways. For long, the ABC-analysis (or Pareto-analysis)
was the only tool for differentiating between important and less important purchases. In
a Pareto-curve items are classified according to the cumulative number of purchasing
orders and their cumulative value. The A-category contains 20% of the numbers of
orders, which typically accounts for 80% of the total value. The B-category, the next
category of 30% of the items accounts for less than 20% of the spending. The remaining
50% of the items accounts for less than 2% of the total spend (C-category). The ABC-
analysis is considered to be helpful in situations where the majority of purchase spend is
caused by only a few material categories (Corsten, 1996). Another volume characteristic is
the number of parts, which is especially important in discrete production. Instead of a
volume-dimension, it is also possible to use the cumulative number of suppliers. This
classification differentiates suppliers with significant spend from the mass of suppliers
with only small purchase volume (Hartmann et al., 2001).
The ABC-analysis concentrates on the financial value of items. It ignores the cost of poor
quality (Burt, 1989), performance risk, social risk and other components (Hartmann et al.,
2001). In addition, ABC analysis fails to discriminate between the methods which should
be used to obtain different item categories (Steele and Court, 1996). The ABC-analysis
does not provide strategic recommendations for the categories, but it provides
information on the concentration of purchase spend. It is a classification tool, not a
portfolio model.
The commodity analysis is another example of a classification tool in purchasing. A
commodity analysis divides the total purchasing volume in percentages for all
combinations of ‘product groups’ and ‘principal users’. It reveals key users and the
commodities that are most important to them (Bauer, 1977). Commodity analysis
identifies critical procurement areas, it is helpful for setting priorities, and it provides
recommendations with respect to the organization of purchasing (assignment of
responsibilities and centralized or decentralized purchasing). Just as the ABC-analysis,
the commodity analysis should be classified as a classification tool, since it does not
provide (differentiated) purchasing and supplier strategies. They are both examples of a
spend analysis which is limited to the classification of items and suppliers according to
their financial value.
Kraljic (1977, 1983) introduced the first comprehensive portfolio approach for the use in
purchasing and supply management. Some twenty years ago he advised managers to
guard their firms against disastrous supply interruptions and to cope with changing
economics and new technologies. His message was that ‘purchasing must become supply
management’. In this context Kraljic (1977, 1983) developed a convenient portfolio
approach for the determination of a comprehensive strategy for supply. Kraljic’s
approach includes the construction of two portfolio matrices. The fist matrix engages a
classification of products on the basis of two dimensions: profit impact and supply risk.
Each variable has two possible values: ‘low’ and ‘high’. The result is a 2x2 matrix and a
classification in four categories (see also figure 1.1):
Categories Values Main tasks/strategies

strategic items high on both dimensions diversify, balance, or exploit
bottleneck items low profit impact, high supply risk volume assurance
leverage items high profit impact, low supply risk exploitation of purchasing power
non-critical items low on both dimensions efficient processing
A second matrix is used for the strategic items. This matrix shows the relative power
position of the company in the corresponding supply markets. Three general purchasing
strategies are determined, depending on the balance of power in the buyer/seller
relationship: exploit (in case of buyer dominance), balance (in case of a balanced
relationship), and diversify (in case of supplier dominance). Kraljic’s approach elaborates
and focuses on the strategic items. For the other item categories Kraljic merely formulated
a number of ‘main tasks’. Kempeners and Van Weele (1997) have emphasized that the
upper-left area of the (first) Kraljic matrix refers to a buyer-dominated segment, while the
lower-right area corresponds to a supplier-dominated segment. The balance of power
obviously is a key issue for the classification of items in the Kraljic framework and for the
selection of differentiated strategies.
24
FIGURE 1.1 Kraljic's categories and strategic recommendations
Other authors have used Kraljic’s basic ideas for the development of similar portfolio
models, see for instance Elliott-Shircore and Steele (1985), Syson (1992), Van Weele
(1992), Olsen and Ellram (1997),Lilliecreutz and Ydreskog (1999), Gelderman (2000), and
to a certain extent Bensaou (1999). In general, purchasing portfolio models aim at
developing and implementing differentiated purchasing strategies. Recently, some new
specific applications have been introduced, notably the determination of the number of
suppliers (Homburg, 1995), the selection of cost management tools (Ellram, 1996),
supplier involvement in product development (Wynstra, 1998), supplier selection (De
Boer, 1998), supplier development (Handfield et al., 2000), web-based procurement of
MRO-items (Croom, 2000), specification process (Nellore and Söderquist, 2000),
engineering-purchasing-supplier interaction (Nellore and Taylor, 2000), facilitation of an
internal process of change (Axelsson et al., 2000), interorganizational competence
development situations (Møller and Momme, 2000), recruitment and competences of
purchasers (Vammen 2000), strategic structuring of suppliers in a supply network
(Åhman, 2002), and the positioning and implementation of E-Procurement (Caldwell et
al., 2002 and Leonard and Spring, 2002).
In the course of time the Kraljic approach has entered many textbooks on purchasing and
supply management. Purchasing portfolio models have gained ground in both research
and practice (Nellore and Söderquist, 2000). Kraljic (1983) made a reasonable case for the
usefulness of the portfolio approach by describing the experiences of four large industrial
companies: a welding materials producer, a manufacturer of electrical equipment, a
chemical company and a heavy-equipment maker. Now, many years later the purchasing
portfolio approach is being used by several other large companies, such as Shell, Alcatel,
Philips, DSM, and Siemens (e.g. Van Weele, 2000). Especially in Western Europe Kraljic
has received large-scale recognition and has gained an increasing degree of adoption.
Gradually Kraljic has gained acceptance in other countries, notably in the USA, Canada
and Northern Europe.
The Kraljic portfolio approach can be considered as an important breakthrough in the

development of theory in the field of purchasing and supply management. Syson (1992)
characterized it as a powerful tool, to be used for diagnostic and prescriptive purposes,
and that goes far beyond the well-known, rather simplistic ABC-analysis. For long
Kraljic’s model was the only available purchasing portfolio model. In the late nineties,
portfolio models received more attention in professional and academic publications. In
general there is a growing interest in the possibilities of portfolio models in purchasing
and supply management.
1.5 Statement of the problem

We have observed a growing acceptance and utilization of purchasing portfolio models,
especially Kraljic’s approach. However, some unanswered questions remain. For
instance, the choice of dimensions is to a certain extent plausible, but it is not clear why
these dimensions are appropriate and how they might be related to a more
comprehensive theory. What might be the theoretical foundations supporting the choice of
dimensions? In general, decisions based on portfolio models are proven to be sensitive to
the choice of dimensions, factors, and weights. Day (1986) concluded that measurement is
considered to be the Achilles’ heel for all portfolio models. What is exactly meant by
‘profit impact’ and ‘supply risk’? How could or should we measure them in practice?
Theory does not provide prescriptions or procedures for measurement, leading Ramsay
(1996a: 15) to conclude that these concepts are “actually made up of a number of
nebulous concepts without operational dimensions”. Olsen and Ellram (1997)
emphasized that the weighting of each factor is the most important part of the
implementation process, but at the same time very subjective. The decision-makers must
come to an agreement on the relative importance of each factor. Besides, what is the exact
distinction between ‘a high’ and ‘a low’ supply risk? If we have problems discriminating
between categories, than the classification of products will be arbitrary and so will be the
provided recommendations. Homburg (1995) for instance concluded that
recommendations should be applied with reserve, especially if a product is positioned
near a demarcation line. De Boer (1997) suggested a fully customized approach:
organizations should determine their own criteria as well as their own specific threshold
values. But, perhaps more importantly, how do purchasing professionals handle such
26
issues in practice? In general, what could we learn from their experience? Little is known
about the actual use of portfolio models in purchasing. Most publications are conceptual
or anecdotal by nature. Leonard and Spring (2002) concluded that there is a need for
research on the way in which managers actually use portfolio models, how they are
operationalized in complex organizations, and on the political process within
organizations where the classification takes place. More questions arise. How useful is a
portfolio approach? What measures of success should be used in the evaluation of a
purchasing portfolio approach? How many firms actually use portfolio techniques and
for what reasons? Are they using the Kraljic approach or are other models used or
developed in practice? Does the portfolio approach take into account the possible
strategies and strategic intentions of the supplier? What company-specific factors will
influence the management of the company’s relationships? Olsen and Ellram (1997)
suggested that future research should include case studies to capture important aspects
of the implementation process.
To conclude, in contrast with the increased adoption of purchasing portfolio models,

there are many unanswered questions. There is a lack of knowledge with respect to a
number of issues, such as the actual use (frequency and implementation), theoretical
foundations, empirical evidence and empirical testing. This brings us to the research
objectives and the research questions.
The objectives of this research project are to gain a better understanding of:
- the theoretical and conceptual foundations of purchasing portfolio models,

- the actual use of purchasing portfolio models in practice, and
- how they could be used by purchasing professionals in order to pursue differentiated
purchasing strategies.
These research objectives imply a conceptual study of the various portfolio models in
combination with an empirical study of practical experiences with portfolio approaches
in purchasing. The research objectives are elaborated in five major research questions,
which are:
1) What are the differences and similarities between the various purchasing portfolio models?
Portfolio models are generally used for decisions with respect to the allocation of limited
resources. These models are developed in various fields and disciplines, such as
investment theory, strategic management and marketing management. This study also
includes a discussion and a review of portfolio models in related business disciplines,
drawing from areas with a longer experience in the use of such models.
Kraljic (1983) introduced the first comprehensive portfolio approach in purchasing which
is still the dominant approach in the profession. Since then, some variations of the Kraljic
approach have been introduced, building more or less on the points of departure of the
original model. This study will investigate the various portfolio models in purchasing,
identifying differences and similarities. According to the three basic elements of any
portfolio model, the portfolio approaches in purchasing will be discussed and evaluated
on (1) dimensions, (2) categories, and (3) strategic recommendations. In addition, a
comparison will be made on (4) use issues (acceptance and adoption).
2) Which factors would explain the utilization of the purchasing portfolio analysis?
A number of sources has reported on the adoption and growing utilization of the
portfolio analysis in practice (for instance Van Weele, 1997; Cox, 1997; Boodie, 1997, and
more recently Kamann, 2000; and Lamming and Harrison, 2001). In advance it seems that
the popularity of the portfolio tool has to do with the attractive visual display, the
convenient classification of items and the face validity of the strategic recommendations.
However, little is known about the actual use of the portfolio analysis, including number
of users, use intensity, reasons for (non)use, perceived (dis)advantages, satisfaction and
attributed results.
The second research question of this study refers to possible explanations of the
utilization of the portfolio analysis. Based on literature study a number of sub-questions
will be formulated and elaborated. A use model will be specified, including variables to
explain the actual use of a purchasing portfolio approach. The results of the survey will
be used for quantifying variables and relationships, and for testing hypotheses.
3) How are portfolio models employed by experienced purchasing professionals?
As stated earlier, little is known about the actual use of purchasing portfolio models.
There are unanswered questions, addressing measurement issues and strategic issues.
How do experienced professionals handle such issues in practice? What could we learn
from their experience? These tentative questions will be elaborated and articulated
during the research project, including theoretical and empirical findings. The literature
study will not result in final answers to these questions, because most publications are
conceptual or anecdotal by nature. Therefore, an exploratory field study is needed to
address the gap between conceptual problems and practical solutions, identifying and
describing advanced uses of a purchasing portfolio approach.
4) Under which conditions are the various portfolio-based strategies selected in purchasing
management?
28
We would like to gain more insights in the possibilities for selecting and developing a set
of differentiated purchasing strategies, based on a portfolio approach. Which factors
would explain the selection of different strategic choices? Answering research question 3
will lead to an overview of portfolio-based strategies for purchasing management. The
fourth research question is directed towards the conditions that produce the selection of a
certain portfolio-based strategy. Under which conditions are the various strategies
selected? The possibilities for selecting purchasing strategies are obviously limited by
external conditions, for instance by conditions regarding mutual dependence and the
power of balance. For the explanation of strategic choices we will focus on the
determinants of (buyer and supplier) dependence and other relationship conditions,
producing different strategic choices.
5) What is the role of power and dependence in the Kraljic approach?
It can be argued that power and dependence are very important in understanding
buyer/supplier relationships. However, the Kraljic approach does not explicitly deal with
issues of power and dependence. Some of the recommendations obviously refer to the
power structure (‘exploit power’), others do not. Some are aimed at reducing the
dependence on suppliers (‘diversify’), others are not. In addition, the recommendations
for the strategic items are largely determined by the balance of power, while it is not clear
in what way the dimensions ‘profit impact’ and ‘supply risk’ are related to the relative
power position of the buying company. At the same time Kraljic (1983: 112) posited that
the general idea of the portfolio approach is to “minimize supply vulnerability and make
the most of potential buying power”. To conclude, power and dependence do play a
significant part in the Kraljic approach, although in an unclear way.
1.6 Relevance of the stated problem

In the last two decades there has been a strong development of professional purchasing
in many organizations. Many new methods and concepts have been developed in the
field of purchasing and supply chain management (Van Weele et al., 1998). Purchasing
portfolio management is one of them. Organizations are increasingly using portfolio
approaches in purchasing. Purchasing professionals tend to use the Kraljic portfolio
model or similar approaches as a tool for classifying their products and for formulating
differentiated purchasing strategies. As we have observed, little is known about the
actual use of portfolio models in practice. Despite their weak theoretical and empirical
foundations, portfolio models do lead to prescriptions or strategic recommendations.
There are some unanswered questions with respect to the measurement of dimensions
and factors, and the nature of strategic recommendations.
This research project aims at a significant relevance for business as well as for science.
With the results of this research project the management of companies should have a
better understanding of the possibilities and limitations of purchasing portfolio
approaches. Insights will be gained in the actual use in practice. Purchasing management
tools, such as guidelines and checklists, can be derived from the results of the research
project. Practitioners may benefit from this study, finding indications for a more
advanced utilization of a portfolio approach in purchasing management.
The popularity of portfolio approaches can be explained by their relative ease of use and
the straightforward strategic recommendations with a high face validity. The strategic
options all seem very logical. But, from a scientific point of view, the models, the choice
of dimensions and the recommendations lack theoretical and empirical foundation. In
general little research has been conducted with respect to purchasing portfolio models. In
addition, most studies are conceptual by nature, pointing out different factors and
approaches which are in affect variations of the original Kraljic-matrix. The lack of
proven knowledge is problematic since portfolio models are increasingly adopted by
purchasing practitioners. This research project attempts to respond to these omissions
and give more insights into the usefulness and possibilities of purchasing portfolio
models.
1.7 Methodology
Three major research methods are successively being used in this research project: an
extensive literature study, a series of explorative case studies, and a large scale survey.
The literature study covers three areas: portfolio models in related business disciplines
(1), portfolio models in purchasing management (2), and power and dependence in
buyer/supplier relationships (3). Each research method has its own characteristics and its
own strong points, which make it more appropriate for answering certain types of
research questions. Figure 1.2 provides an overview of these successive research steps.
30
FIGURE 1.2 Overview of the successive research steps
We will amplify on the three main research methods, describing and explaining the
various research steps. In addition, we will clarify in more detail the relationships
between the research steps. Obviously, the research steps are not arranged in a random
order. Any output of a research step will result either in an answer to a research question
or in a useful input for a next research step. Figure 1.3 provides a comprehensive view of
the intermediate research results (input) and the end results (output) of the various
research steps.
(a) Literature study

We have started the literature study with a review of portfolio models in related business
areas, namely investment theory, strategic management, and marketing management. The
main reason for starting with other areas than purchasing management was that we
wanted to learn from disciplines with a longer tradition and experience in the use of such
models. The lessons and insights should be valuable for the next step of the literature
study: the review of portfolio models in purchasing management. This step will answer the
first research question (‘similarities and differences’). The critical review of the portfolio
models will result in an overview of the main issues and unanswered questions. Most
prominently, measurement and strategic issues will be identified and used as input for
the case studies. The case studies aim to reveal how experienced professionals handle
these issues in practice, providing an answer to the third research question (‘handling of
issues’). Additionally, we will explore the literature in search of factors and variables that
might explain the use of a portfolio approach in purchasing. The variables enter a use
model, to be quantified and tested by means of the survey data. During the research
project we have stayed closely connected with state-of-the art literature. Since the start of
the research project in 1998, we have observed a growing interest in purchasing portfolio
management, which has resulted in a corresponding growth of academic and
professional publications on this matter. Obviously, we have used the additional insights
in the literature study.
Our study of purchasing portfolio models induced us to go deeply into issues of power
and dependence in buyer-supplier relationships. It was apparent that these issues are key
elements for understanding the foundations and possibilities of a portfolio approach to
the management of supplier relationships. Much has been written about power and
dependence in business relations, although never related to the purchasing portfolio
models. We have addressed this gap and formulated a number of hypotheses regarding
the (expected) power and dependence structure in the Kraljic matrix. Additionally, we
have looked for possible determinants of buyer’s dependence and determinants of
supplier’s dependence. This has resulted in a preliminary model of determinants, which
has been adjusted after consulting purchasing practitioners who were involved in our
case studies.
(b) Case studies

Case studies will be used to identify and to describe advanced current practices with
respect to purchasing portfolio models. The main purpose of the case studies is to answer
the third research question, which refers to the employment of portfolio models by
experienced purchasing professionals. How do they handle measurement issues and
strategic issues? Which solutions have been developed in practice? It will become clear
that portfolio analysis enables purchasing professionals to differentiate their purchasing
strategies, not only between the item categories of a matrix, but also within each category.
A further analysis of the case studies will result in an overview of portfolio-based
strategies and the conditions under which they are selected in practice. This
comprehensive model of strategic directions enables us to develop different scenarios, to
be used in the survey study.
The case study method is appropriate because of the limited research on the actual
employment and possibilities of a portfolio approach in purchasing and supply
management. Most publications are conceptual or anecdotical by nature. Additionally,
this research step is of an explorative nature, aimed at identifying advanced practices,
dealing with ‘how’ and ‘why’ questions.
The sample consists of a selection of Dutch companies, restricted to production
industries, where purchasing is by nature an important business area. We have selected
three case companies mainly on their ongoing use of purchasing portfolio analysis. The
cases were studied sequentially, one after another. Because we wanted to explore
different possibilities of the portfolio approaches, we have included different levels of
32
analysis: the corporate level of a large company, the level of a multinational business unit
with many plants all over the world, and the level of a business unit of a fairly small
(single plant) company. The variety in levels enables us to reveal different kinds of
practices, although we are aware of the limitations of this approach. The case studies are
not supposed to give an exhaustive treatment of portfolio methods, and they do not allow
for any statistical generalization.
Respondents were interviewed (face-to-face) on the basis of a semi-structured
questionnaire, allowing for elucidation, elaboration and clarification. The case studies
entail the use of a key-informant method in combination with a snow-balling technique
whereby the key informant is nominating other informants. Several rounds of interviews
were conducted with the respondents, at each stage reporting back the tentative analysis
and conclusions from earlier rounds, giving them the opportunity to check and recheck
interim reports, to improve the match with the intended information, and to explore
issues in more detail.
(c) Survey
In the third and last stage a survey will be conducted. Principally, the questionnaire is
aimed at measuring the variables and relationships in the conceptual models, which are
based on the insights from the literature study and the case studies:
- the conditions for the selection of portfolio-based strategies,
- the power and dependence structure in the Kraljic matrix, and
- the determinants of buyer’s and supplier’s dependence in the Kraljic matrix.
The survey method requires the development of scenarios which describe a number of
situations in terms of the Kraljic dimension (profit impact and supply risk) and in terms
of the selection of a corresponding specific (portfolio based) purchasing strategy.
Respondents will be asked to assume the role of a purchasing expert in their own
company where they are actually employed. The respondents are asked to evaluate a
series of questions relating to the different scenarios. The design of the study can be
characterized as a repeated measures design, because the same respondent participates in all
conditions of the experiment.
Different sources of knowledge and expertise are being used for the development of the
questionnaire. The construction of items is largely based on literature review
(operationalization) and on the insights gained from interviewing practitioners during
the case studies. The first draft questionnaire will be discussed with a focus group of
academics. The procedure includes a pilot study aimed at enhancing the reliability and
the validity of the questionnaire. A small pre-test group will be asked to review the
questionnaire for the clarification of questions, instructions, lay-out and other text
elements. Finally, the questionnaires will be administered in three rounds to a large
number of purchasing managers who are employed by manufacturing companies who
are member of the Dutch Association of Purchasing Management (NEVI).
The survey data allow for quantifying and testing the relationships between variables in
the conceptual models. The subsequent analysis of the collected data will result in the
rejection and confirmation of the formulated hypotheses. The survey instrument is
appropriate for answering the second research question (portfolio use), the fourth
research question (conditions), and the fifth question (power and dependence).
FIGURE 1.3 Intermediate and end results of the successive research steps
We have presented and discussed our findings during several IPSERA-conferences, one
of the leading international research conferences on purchasing and supply chain
management (see Gelderman and Van Weele, 2000, 2001, 2002a, and 2003).
1.8 Structure of the thesis

In this chapter we have outlined recent (revolutionary) developments in purchasing
management, we have introduced the study, explained the background, and described
the design and methodology of the research project. In the next chapter we will explore
different portfolio approaches that are used in related business disciplines, such as
strategic management and marketing management. We will start with the roots of
portfolio management, which lie in investment theory. Chapter Two concludes with the
34
most important lessons and experiences, as background information for the investigation
of purchasing portfolio models. In chapter Three we will review the main portfolio
approaches in purchasing and supply management. Kraljic’s fundamental ideas and
concepts appear to dominate the discipline. However, there are a number of problems
and unanswered questions. Building on the findings and conclusions in chapter Four, we
will develop some new perspectives on Kraljic’s original portfolio matrix. A theoretical
foundation is found for Kraljic’s portfolio approach in the resource dependence theory
(Pfeffer and Salancik, 1978). A new mutual based dependence model is introduced by
adding a (resource) dependence-perspective to the original Kraljic matrix. Elaborating on
‘power and dependence’, we will explore the determinants of buyer’s and supplier’s
dependence (Chapter Four). The field research part starts with Chapter Five which
reports on the in-depth case studies. Measurement issues and strategic issues are
addressed, providing insights into the various ways experienced professionals deal with
a purchasing portfolio approach in practice. Chapter Six explains the design of the survey
and summarizes the findings. In Chapter Seven we will present the main conclusions and
recommendations of this study. We will reflect on the research project, providing
suggestions for further research.
35
2 Portfolio-approaches in related
disciplines
In this chapter we will discuss and review portfolio models which are developed in
related disciplines, learning from areas with a longer tradition and experience in the use
of such models. We will begin with the roots of all portfolio management: investment
theory. Then we will review portfolio models for strategic management, developed in the
early seventies. They have received a great deal of attention in strategic planning
(Armstrong and Brodie, 1984). Therefore we expect to find a mass of experience with
portfolio-approaches in strategic management. In this section we will look for valuable
lessons and learning experiences for the use of portfolio analysis in purchasing
management. Another point of interest is the role of purchasing in (corporate) portfolio
models. If purchasing is of a strategic nature and is said to have a significant impact on
the competitive position of companies, then obviously purchasing should be included in
any corporate portfolio model. Finally, we will review marketing portfolio models which
are of special interest to the central object of this study, being purchasing portfolio
models. After all, both marketing and purchasing professionals are trying to manage
exactly the same buyer/supplier relationships. A customer focus in marketing implies
that marketing models are likely to pay much attention to the interests and positions of
customers (the buying perspective).
2.1 Investment theory: the roots of all portfolio

management
Portfolio theory has its roots in financial investment and the desire to balance the
conflicting objectives of high yield and low risk (Yorke, 1984). In common parlance a
‘portfolio’ refers to stock, bonds and other financial investments. Webster’s Third New
International Dictionary (1993: 1768) defines ‘portfolio’ as “the securities held by an
investor or the commercial paper held by a bank or other financial house”. The work of
Markowitz in the early 1950s is generally seen as the origin of modern portfolio theory
for investment purposes. Further development from Markowitz’s portfolio theory gave
rise to what is known as the Modern Portfolio Theory (Sharpe, 1963).
2.1.1 The Markowitz model
In the early 1950s the investment community talked about risk, but there was no specific
measure for this key concept. Investors had to quantify ‘risk’ for investment decisions.
Markowitz (1952) derived the expected rate of return for a portfolio of assets and an
36
expected risk measure. He showed that the variance of the rate of return was a useful
measure of portfolio risk and he derived a formula for computing the variance of a
portfolio. Markowitz showed that:
- the expected rate of return of a portfolio is the weighted average of the expected return
for the individual investments, and
- the standard deviation of a portfolio is a function not only of the standard deviation for
the individual investment, but also the covariance between the rates of return for all the
pairs of assets in the portfolio. In a large portfolio, these covariance’s are the important
factors. Markowitz’s formula not only indicated the importance of diversifying
investments to reduce the total risk of a portfolio, but also showed how to diversify.
The Markowitz model is based on several assumptions regarding investor behavior:

- Investors consider each investment alternative as being presented by a probability
distribution of expected returns over some holding period.
- Investors maximize one-period expected utility, and their utility curves demonstrate
diminishing marginal utility of wealth.
- Investors estimate the risk of the portfolio on the basis of the variability of expected
returns.
- Investors base decisions solely on expected return and risk, so their utility curves are a
function of expected return and the expected variance of returns only.
- For a given risk level, investors prefer higher returns to lower returns. Similarly, for a
given level of expected return, investors prefer less risk to more risk.
(e.g. Reilly and Norton, 1999)
Under these assumptions, a portfolio of assets is considered to be efficient if no other

portfolio offers higher expected return with the same or lower risk, or lower risk with the
same (or higher) expected return. Markowitz conclusion is that rational investors would
(or should) want to select efficient portfolios, i.e. portfolios with a minimum of risk
(variance) for a given expected return, or with a maximum expected return for a given
risk (variance).
Nowadays Markowitz formula is still being used as the basis for modern investment
theory and investment practice. The consequences are far-reaching and provide much
guidance for investment decisions. The most important implication is that diversification
reduces variability and risk. The greatest payoff to diversification comes when stocks (or
other assets) are negatively correlated (Brealy and Myers, 1996). When there is a perfect
negative correlation (-1) between two assets, the overall variance of the portfolio is zero
(0). This would be a risk-free portfolio (Reilly and Norton, 1999). Wise investors do not put
all their money into just one stock. They want to reduce their risk by diversification.
Investors should be interested in the effect that each stock has on the risk of the portfolio
as a whole. It is therefore not decisive how risky an investment is, but what the impact
will be on the risk of the entire portfolio. One of the simplest ways for an individual to
diversify is to buy shares in a mutual fund which holds a diversified portfolio. Software
programs, called ‘optimizers’, are used to determine ‘efficient portfolios’. Financial
planners use information on past returns and manager performance, in addition to
optimizers, to make recommendations to their clients (cf Reilly and Norton, 1999).
2.1.2 Relevance for purchasing: what can we learn?
Investment theory basically deals with the choice of investments between an infinite number
of potential portfolios, resulting in a ‘yes’ or ‘no’ kind of recommendations. It is all about
the composition of an investment portfolio by reducing risks and optimizing returns. In
purchasing management however it is a matter of choice of strategies for a given number of
items (usually products), resulting in a set of differentiated purchasing and supplier
strategies. This means that there are important difference in scope and perspective.
Another major difference regards the existence of another party and the factor of social
interaction (Yorke, 1984). Here lies a sharp contrast with the application of portfolio
theory to investment purposes and to purchasing purposes. There is always an
unpredictable element in a purchasing context, due to the dynamics of business and
human behavior. The risk-factor is of an other order. In the investment theory the
variable ‘risk’ is measurable in a rather easy and unambiguous way, based on a
mathematical formula. It is clear that such a formula does not exist for the determination
of risk in a purchasing context.
The main similarity however is that any portfolio approach focuses on the efficient
allocation of limited resources. In a general sense this problem is relevant for investment
decisions and for purchasing decisions. A general objective of Kraljic’s model is to
minimize supply risk (Kraljic, 1983). Another similar characteristic refers to the trade-off
between risks and rewards. For instance, there are benefits in spreading purchases
among a number of suppliers. However, there is a trade-off involved. The net effect of
dealing with a large number of suppliers, can be to shift the balance of power towards
the seller rather than towards the buyer. Just as in the financial world, ‘risks’ and
‘rewards’ are closely linked. Buyers should therefore be aware of the balance between
risk and return. Nicholson (1993) points out that buyers should ask themselves:
- If the overall business risk increases due to a purchasing strategy, is that increased risk
matched by an increased return or benefit?
- If the buyer obtains an increased reward for the firm, has it been achieved by exposing
the business to a higher level of risk?
Smeltzer and Siferd (1998) argued that proactive purchasing management is concerned
with risk management. It should mitigate risk and, at the same time, provide a higher
return. To conclude, purchasing management has to deal with issues of ‘risks’ and
38
‘rewards’, albeit that the context of decision making is different in comparison to

investment problems.
2.2 Portfolio models in strategic management

In seeking answers to questions of strategic planning for the diversified organization,
management has a strong need for tools that assist in allocating resources among
business units (or products). A number of portfolio models have been proposed for this
purpose. First and best known is the Boston Consulting Group’s growth share matrix
(Henderson, 1970, 1972, 1973). The model is widely used, even though it has received
considerable criticism and there are some major limitations. Other multifactor models
have been developed with the intention of overcoming some of the limitations of the
BCG-approach.
The portfolio models for strategic management are mainly developed to support resource
allocation decisions among strategic business units (SBU’s). However, any portfolio
analysis can be conducted at different levels of operation (Wind et al., 1983):
- a portfolio of strategic business units at the corporate level,
- a portfolio of product lines at the SBU-level, and
- a portfolio of products at the product group level.
In this section we will concentrate on the portfolio analysis at the corporate level.
Portfolio models with respect to products lines and products will be discussed in section
2.3 when dealing with portfolio models in marketing management.
2.2.1 The BCG-approach
The BCG-model uses two key variables: relative market share and business growth. The
positions of SBU’s can be portrayed in the growth-share matrix, that also shows the
turnover for each SBU by the diameter of the circles. Figure 2.1a shows an example of a
growth-share matrix that consists of eight SBU’s. The quadrants of the growth-share are
connected with expected cash flow results: stars, problem children, cash cows and dogs (see
figure 2.1b).
FIGURE 2.1a An example of a growth-share matrix
FIGURE 2.1b Cach flows in a growth-share matrix

40
In the matrix the market growth is measured on a linear scale, the relative market share
on a logarithmic scale. The growth rate variable is divided into high and low, usually at
the level of the nominal GNP growth, which in those days was between the range of 10 to
12 percent. In the upper half of the matrix we would expect SBU’s with products that are
positioned in the early phases of their product life cycle. For new products it is assumed
that they need high expenses for which in turn relative small earnings are gained.
The relative market share is determined by dividing the market share by the market
share of the largest competitor, on the horizontal axis. A vertical cross hair is placed at a
relative share of 1, which is the dividing line for a market leader position. Another line is
often placed at a relative share of 1.5 on the grounds that dominance is not established
unless the business is at least 50 percent larger than the closest competitor (Day, 1986).
Being market leader has the advantage of moving faster on the experience curve. As a
consequence, SBU’s with a higher relative market share, produce in a more efficient mode
with lower average costs, and with relative high cash flows. The marks for both variables
can be viewed as natural threshold values, this in contrast with most other portfolio
models with more arbitrary cross hairs.
A relative high market share implies relative high cash revenues. In contrast, to be
competitive in fast growing markets the company needs relatively high cash
expenditures. When market growth declines, stars can change into cash cows under the
condition that no relative market share is lost. Problem children on the contrary need a
great deal of investment. Cash cows finance the problem children so that they will
eventually grow to be the future stars. Less potential problem children can be sold,
terminated or harvested.
The ‘cash flow’-concept plays a central role in the BCG-approach: in view of continuity
every corporation needs to have continuous positive cash flows to finance new,
promising activities and business units. Eventually this new business will replace the
activities that are currently profitable. In terms of the BCG-approach this means that a
well-balanced portfolio of SBU’s is necessarily for the survival of the corporation.
On the basis of the growth-share matrix the following strategic recommendations are
provided:
1. The strategic conclusion for dogs is to disinvest or withdraw. Earnings will be low and
little or no profit will be made now and in the future.
2. The recommendation for stars is to invest for growth. There are good prospects for
SBU’s with a relative high share in a growing market.
3. There is a potential for problem children, although there is a low market share. There
could be possibilities to make stars out of them, but there is the risk that problem children
degrade into dogs.
4. The strategic implication of cash cows is to exploit their strong positions, but not to
spend much money on them. The positive cash flows should be invested in stars and
occasionally in problem children.
Saunders (1997: 94) summarizes: “Milk the cows, invest in the stars, divest the dogs and
analyze the problem children to determine whether they can be growing into stars or
whether they will degenerate into dogs.” These recommendations show an emphasis on
interdependence of elements within the portfolio-concept: managing the whole, rather
than making separate allocation decisions for separate SBU’s.
2.2.2 Assumptions, criticism and adoption
The growth share matrix consists of two dimensions: market growth and relative market
share. These dimensions were chosen because market growth served as a proxy for the
need for cash, while relative market share was a proxy for profitability and cash generating
ability.
The relative market share is an indication for the company’s competitive position: its
market share relative to its largest competitor in the same market. High relative market
shares are assumed to be more profitable, especially if the experience curve applies. The
experience curve effect states that as the accumulated experience of manufacturing
doubles, total unit costs can decline by 20-30 percent. In effect, the greater the market
share, the greater is the cost-saving accrued from the experience effect (Turnbull, 1990).
The market growth-variable was chosen as a rough proxy for the product life cycle,
because it is believed to have known and predictable consequences for cash requirements
(Day, 1986). In a high-growth market, sales volume can more easily be maintained or
increased, because of new users and new uses. However, in a static or declining market,
shares can only be gained in a struggle with competitors. As a consequence, marketing
costs are likely to be high.
It is agreed that, although industry growth and relative market share are very important
variables for strategic planning, the complexity of the business environment does not
allow for simple recommendations. More factors should be incorporated in the strategic
decision planning process. Morrison and Wensley (1991) suggested that the scope of the
BCG matrix ignores other relevant strategic issues. In addition, the model uses data from
the past with respect to current business units. It can be said that the BCG model is not
designed to deal with the development of new business opportunities and may even
inhibit creative thinking about these questions (Day, 1986). But there is more criticism.
In the BCG-model dogs are undesirable business units: they should be disinvested or
eliminated. This drastic recommendation seems to be not always the wise thing to do. Most
markets are mature and show a relative low growth rate. Since there can only be one
market leader per market, most strategic business units are ‘dogs’. In contrast with the
42
theory it is certainly possible that dogs are profitable business units, even in the long run.
There is empirical evidence that ‘dogs’ perform much better than would be predicted by
the BCG-model. Hambrick et al. (1982) empirically explored the performance of business
units in the four cells of the matrix. The data used in this study were drawn form the
Profit Impact of Marketing Strategies (PIMS) project. They found that ‘problem children’
had negative cash flows, ‘cash cows’ were net cash generators and the average ‘star’
generated as much cash as it spent, all according to the BCG-principles. The ‘dogs’
outperformed the expectations placed on them by BCG. Many ‘dogs’ earned higher
profits than other categories in the growth-share. The main conclusion is that the results
of this study do not support the BCG’s advice that ‘dogs’ should be harvested or
liquidated.
There are also questions on the validity of the fundamental assumptions of the BCG-
model. Several studies (e.g. Jacobson and Aaker, 1985; Cook, 1985; Jacobson, 1988) have
questioned the validity and generalizability of the market share-profitability relationship.
It is claimed that there is little empirical evidence to support a causal relationship
between market growth and profits.
However, there are sources which express strong support for the assumed relationships.
Buzzell et al. (1975) found a significant relationship of market share and return of
investment: the greater the share, the greater the cash flow. Szymanski et al. (1993)
performed a meta-analysis on 276 market-share profitability findings from 49 studies.
They found that, on average, market share has a positive and significant effect on
business profitability.
In addition to the questions on the assumptions, there are difficulties in measurement of

market growth and relative market shares. There are no natural market boundaries and
undebatable market definitions. Different market definitions will produce different
market leaders, different growth numbers, different matrices and different strategic
recommendations. Wind et al. (1983) proved that decisions based on portfolio models are
quite sensitive to the choice of the methods of measuring market share and market
growth. According to Day (1986) the Achilles’ heel of the growth-share matrix is the
measurement. If the share and the growth estimates are dubious, so are the
interpretations and the recommendations.
The BCG-approach is also based on two assumptions with respect to cash (balance):
- there is a need for companies to achieve and maintain cash balance;
- the interdependence among business units is limited to the generation and use of cash.
The strategic recommendations of the BCG-matrix emphasize the balance of cash flows.
Some business units have to supply cash so that other business units are able to sustain
growth in the future. The underlying premise is that companies do not have possibilities
for raising outside capital for their activities. Wensley (1981: 176) called it “a strategic
mistake in seeing the corporation as an independent cash recycling entity”.
According to Day (1986) there is an implicit and potentially dangerous premise of the
growth share matrix that interdependencies among businesses are limited to their
generation and use of a common resource: cash. This assumption neglects the possibilities
of sharing techniques, skills, and knowledge across business units that might give the
company a long-term competitive advantage. Prahalad and Hamel (1990) introduced the
powerful term ‘core competence’, referring to underlying forces and powers that
constitute competitive advantages.
In defense of the BCG-model Morrison and Wensley (1991) concluded that much of the
academic criticism is misplaced. Too often the matrix is being treated as if it were a
‘comprehensive’ theory of markets and company performance. Morrison and Wensley
(1991) argued that it is more sensible to think of BCG’s approach as a tool, rather than a
theory. A hammer is good for banging in nails, but not with your eyes closed. In other
words, any strategic planning technique needs to be thoughtfully applied.
While the academic community was dissecting the model and other matrices, and
evaluating their failings and shortcomings, the business community was adopting the
BCG-model in the late seventies and early eighties (Zallocco et al., 1983; Verhage and
Waarts, 1988; Morrison and Wensley, 1991). In 1978 and 1979 Haspeslagh carried out
research into the adoption of portfolio planning techniques by major US corporations. He
found that the technique had spread across a wide range of companies and was still
being increasingly introduced. Most companies introduced it under conditions of crisis
and capital constraint, in situations of uncertainty and competitive pressure (Bowman,
1974; Ansoff, 1984). It was perceived as being a tool for communication and influence from
the corporate centre to its diversified business units, and particularly as a framework for
resource allocation. Hamermesh (1986) found that roughly 75 percent of the Fortune 500
companies practiced some form of portfolio planning. Capon et al. (1987) presented
evidence that it is the most widely used portfolio method in US firms. Morrison and
Wensley (1991), in their survey of teachers at 34 business schools in the UK, found that
the BCG matrix is taught in all schools.
2.2.3 Other, multifactor portfolio models
Characteristic for the BCG-approach is the use of single-variable dimensions. Business

growth and relative market share are to a certain extent measurable in an unambiguous
and rather objective sense. This can be seen as a strong point, but also as a serious
limitation. Managers would want more flexibility in determining dimensions, factors and
weights. In addition to the BCG framework, others have been developed for similar
purposes: the multifactor portfolio models that use composite dimensions to designate
44
the matrix axes. In most books on strategic management the following models are
included:
- the General Electric Business Screen
- the Shell directional policy matrix
- the strategic condition matrix.
The General Electric ‘Nine-cell’ Strategic Business Screen is probably the best known
alternative for the BCG-model (Morrison and Wensley, 1991). The model produces
strategic recommendations comparable to those of the BCG-model, such as ‘invest’,
‘protect’, ‘build’, ‘harvest’ and ‘divest’ (see figure 2.2). In the GE-portfolio model the
business units of a company are positioned in a nine-box against two (composite)
dimensions: industry attractiveness and business strength (Hofer and Schendell, 1978).
Both dimensions are constructed from factors selected and weighted by management.
The main question is: what makes a market ‘attractive’ and what makes a business
position ‘strong’? These factors are usually listed under five major headings: market
factors, competition, financial and economic factors, technological factors, and socio-
political factors.
Protect position Invest to build Build selectively

- invest to grow at - challenge for leadership - specialise around limited strengths
maximum digestible rate - build selectively on - seek ways to overcome
high
- concentrate effort on strengths weaknesses

maintaining strength - reinforce vulnerable areas - withdraw if indications of
sustainable growth are lacking
Build slectively Selectivity/manage for Limited expansion or harvest

Market attractiveness
- invest heavily in most earnings - look for ways to expand without

attractive segments - protect existing programme high risk; otherwise, minimise
medium
- build up ability to counter - concentrate investments in investment and rationalise

competition segments where profitability is operations
- emphasise profitability by good and risk is relatively low
raising productivity
Protect abd refocus Manage for earnings Divest

- manage for current - protect position in most - sell at time that will maximise cash
earnings profitable segments value
- concentrate on attractive - upgrade product line - cut fixed costs and avoid
segments - minimise investment investment meanwhile
- defend strengths
low
strong medium weak

Competitive position
FIGURE 2.2 Generic strategy options in the GE-matrix

Source: Day (1986: 204)
The Shell directional policy matrix has business sector prospects and the company’s
competitive capabilities as two dimensions (Hughes, 1981). This portfolio model appears
to be used infrequently in practice, although it is a potentially rich and valuable tool,
according to McDonald (1990). The strategic condition matrix of AD Little Inc is a ‘business
profile matrix’ that identifies stages of industry maturity and competitive position (Patel
and Younger, 1978).
Some less-known models and matrices are: BCG’s Growth/Gain Matrix (Abell and
Hammond, 1979), the Competitive Advantage Matrix (Lockridge, 1981), Matrix for
Market Definition (Day, 1981) and a 27 option Share/Strategy Matrix (Catry and
Chevalier, 1974). We will not discuss these techniques here because we consider them to
be variations on the original BCG portfolio method. For an overview and comparison of
portfolio models for strategic planning, we refer to Wind and Mahajan (1981), Coate
(1983) and Day (1986). We will focus on the limitations of the multifactor models, in
comparison with the BCG-model.
Unlike the BCG-approach, multifactor models rely heavily on managerial judgment to

identify the relevant factors and determine their relative importance. Because of the
subjective evaluations there is a distinct possibility that the outcome of a multifactor
approach could simply be a tautological recommendation to ‘invest preferentially in those
areas of greatest market attractiveness and strongest competitive position’ (Day, 1986).
Wind and Mahajan (1981) have dealt with another problem that is endemic to any
composite rating based on multiple factors. Two businesses can be assigned identical
positions in the portfolio matrix, despite enormous differences in the underlying factors.
The summed scores could still be the same, ignoring important differences and losing a
great deal of useful information.
The models share a number of important characteristics, approaching strategic problems

in a comparable way. All models suggest variables that can be used to measure the
business strength of each unit and other variables to estimate the attractiveness of the
unit’s industry. Although the choice of weights is subjective, multifactor models can be
considered a generalization of the BCG-model. The BCG weights relative market share
and industry growth with 1, and all other variables with 0 (Coate, 1983). Multifactor
models take a broader view of business strength and market attractiveness. However, the
choice of factors and the weightings are based on managers’ subjective judgment, not
specified by any objective procedure. In comparison with the BCG-model, theoretical and
empirical support for the multifactor models is low (Hooley and Saunders, 1993). Figure
2.3 summarizes the main differences and similarities between the BCG-matrix and the
GE-matrix, the latter as a representative for all multifactor models. In conclusion, there
46
seems to be a trade-off between ‘completeness’ and ‘ease of use’, while the theoretical
and empirical support to a multifactor approach is assessed as ‘lower’, compared to the
BCG approach.
Model Focus Measure Complete- Theoretical Empirical Ease Strategic Operational

ness support support of use value value
Market
Share
Boston Cash vs Low Moderate Mode-rate Good Moderate Good

Box flow
Market
growth
Business
strength
GE ROI vs High Low Low Mode- Moderate Good

matrix rate
Market
attrcct.
FIGURE 2.3 Comparison of the BCG-matrix and the GE-matrix

Source: Hooley and Saunders (1993: 63)
2.2.4 Conclusions
We have conducted this literature study mainly for two reasons: to see whether
purchasing is included in corporate portfolio models and what might be learned from the
many years of experience in this discipline. The discussion has made it clear that
purchasing is completely missing in the various models. We subscribe to the conclusion
of Day (1986) that (corporate) portfolio models have been virtually synonymous with
strategic market planning and that they are particularly used as a framework for resource
allocation based on end market considerations, such as market growth, market
attractiveness, and competitive position. In fact these portfolio approaches should be seen
as marketing models at the corporate level.
Reviewing the different portfolio models, it became clear that every model is a tool, and
that recommendations should never replace common sense and sound managerial
judgments. The overall conclusion is that portfolio models are useful in initiating
corporate planning and strategic change. Much criticism is misplaced because it treats a
matrix as a comprehensive theory and not as a tool (Morrison and Wensley, 1991).
Moreover, there are no simple solutions to complex problems. The clear cut strategic
recommendations for cash cows (‘milk’), stars (‘invest’) and dogs (‘divest’) should not
simply be followed without additional considerations. Referring to the measurement issue,
it was found that decisions based on portfolio models are proven to be quite sensitive to
the choice of the methods of measuring variables and dimensions. For all portfolio
models fundamental issues apply with respect to measurement and validation. Based on
a comparative empirical study, Wind et al. (1983) concluded that the classification of any
business into a specific portfolio position depends on four factors:
- the operational definition of the dimensions used;
- the rules used to divide a dimension into low and high categories;
- the weighting of the variables constituting the composite dimensions that are used, and
- the specific portfolio model used.
Given these results the authors advised to avoid a single portfolio model. They further
suggested the need to test sensitivity of the portfolio classification of businesses to
various definitions, cut-off rules, weights, and models. Especially multifactor models
using composite dimensions rely heavily on managerial judgement which is inherently
subjective.
The rapid adoption of the portfolio models is attributed to the appealing visual displays
and the immediate comparisons and recommendations which it offered (Brown, 1991).
On psychological grounds it fulfils a human desire for taxonomy, classifying a complex
mix of different businesses. Furthermore, it is easy to grasp, uses catch terms and phrases
which are easy to memorize and have a link to strategy (Hooley and Saunders, 1993).
The customization of any portfolio model enhances the strategical thinking of decision-
makers. The process of using might be more important than the derived classifications
and recommendations. Despite all the known limitations and problems, there is a
consensus that portfolio models can improve the corporate management’s ability to
allocate resources across businesses. The BCG-matrix is the best known and probably the
most frequently used portfolio model. It might be argued however, that multifactor models
are to be preferred to the rigid BCG-approach. An important benefit is that the actual
using and customizing of portfolio models will lead to a better understanding of the
strategic issues at hand. The process of customizing and using the model are in itself a
strategic thinking process (Wind and Mahajan, 1981). Portfolio models provide a
structure for analysis that facilitates the communication and sharing of judgments and
assumptions about strategic issues (Day, 1986). Haspeslagh (1982) reported a sizable to
dramatic improvement in corporate management’s understanding of individual
businesses, by using portfolio models. Olsen and Ellram (1997) emphasized that the
process of categorizing the items is even more important than the classification itself.
During the process of categorizing, the decision-makers have to discuss inconsistencies
among themselves and agree on the importance of the different elements of the portfolio.
48
2.3 Portfolio models in marketing management

For marketing management we found three kinds of portfolio approaches:
1. models for the selection of R&D projects for new product development,
2. models for classifying products, and
3. models for classifying customers.
The main issues that will be addressed in this section, are:
- Which portfolio models are developed for marketing management?
- What are the results of empirical studies to the use of marketing portfolio models?
- How is the buyer’s perspective included in these models?
2.3.1 New product portfolio models
The prioritization of new products or R&D projects is said to be vital to successful

business performance (Roussel et al., 1991; Matheson et al., 1994). Technology choices
determine what the business will be in the next years. Obviously, the sales of today come
from investment decisions of the past. Firms should decide on the allocation of resources,
how to spend scarce engineering, R&D and operations resources. New product portfolio
management focuses on project selection, and on which new products or development
projects are to be chosen from the many opportunities firm usually have. The basic issue
of new product or R&D project portfolio models has much resemblance to the issues
raised in selecting a financial portfolio of investments. The main concern in both cases is
the allocation of scarce resources and the selection of elements (investments and R&D
projects) in a balanced portfolio. It is not about finding recommendations in terms of
differentiation in strategies. Portfolio models plot new products (that do not even exist)
on a map and allocate resources to these (Cooper et al., 1998).
Literature study resulted in the following new product portfolio models (cf. Cooper et al.,
1999):
- Financial models
In case of financial models projects are judged and rank-ordered on the basis of financial
criteria, such as net present value, internal rate of return, and payback time (Matheson et
al., 1994). There are also probabilistic financial models, which include Monte Carlo
Simulation and decision trees (Cooper et al., 1998). Faulkner (1996) described a method
that uses the options pricing theory concepts in the R&D valuation process. This method
treats each stage of the new product project much like purchasing an option on a future
investment.
- Strategic approaches
Here the selection of the portfolio of projects is largely driven by the strategy of the
business. The strategy decides the split of resources across different categories to create
strategic buckets.
To make sure the portfolio is consistent with business strategy, strategic buckets are
designated by top management, and funds are allocated to each bucket based on strategic
considerations. An example is Allied Signal that uses three buckets, platform projects,
new products, and minor projects, and separately manages a portfolio within each bucket
(Cooper et al., 2000).
- Mapping approaches or bubble diagrams
These are essentially extensions of the BCG- and the GE-model. In new product mapping
models, various parameters are plotted against each other in a bubble diagram format,
such as ‘reward’ versus ‘profitability’ or ‘ease of undertaking’ versus ‘project
attractiveness’ (Matheson et al., 1994; Roussel et al., 1991).
- Scoring models and checklists
Projects are rated and scored on a variety of qualitative questions. The questions or items
often capture drivers of new product success, such as product advantage, market
attractiveness, synergy with the base business, and familiarity (Hall and Nauda, 1990;
Roussel et al., 1991; Yorke and Droussiotis, 1994).
Cooper et al. (1999) observed that there is very little evidence regarding the transfer of
these models and techniques into management practice, or whether these approaches had
positive results. Matheson et al. (1994) reported on a benchmarking study that was
undertaken in order to learn how industry leaders make strategic R&D decisions and
integrate technology with strategic objectives. The study revealed that there is much
room for improvement in the quality of strategic R&D decisions (organizational
learning). Cooper et al. (1999, 2000) performed an extensive study of portfolio
management in industry. The study reported the portfolio practices and performance of
200 (large) U.S. companies, active in product development in North America. It appeared
that almost every business in the survey used multiple methods or techniques for
portfolio management. These techniques, in order of popularity, are as follows: financial
methods (77%), business strategy methods (65%), bubble diagrams (41%), scoring models
(38%), and checklists (21%). The performance of the various methods was assessed on the
basis of three performance metrics: project value, number of projects, and time to market.
Strategic methods, along with scoring approaches, produced the best portfolios. Financial
methods resulted in the poorest performance results.
50
2.3.2 Classifying products: from product evaluation to product

positioning
The development of a strategy for the product line should be an important part of every
company’s marketing planning activities. The product portfolio of a company is more
than the sum of the individual products. There are always interactions and
interdependencies in an assortment of products. Kotler (1971) described a quantitative
approach for measuring interactions and interdependencies of products in an assortment.
He distinguished three kinds of interactions:
- product interaction in which the marketing mix for each product influences sales and or
cost of the other products;
- sales covariance in which major environmental forces, such as seasonal or cyclical
demand, cause two or more products to move in the same way;
- risk covariance in which the returns on different products are subject to uncertainty and
the measures of uncertainty are intercorrelated.
Kotler (1971) adopted the original Markowitz (1952) portfolio model for the case of
selecting an optimal product mix in terms of an ‘efficient portfolio set’. Under the
assumptions that:
- the average rate of return on each product can be estimated,
- the variance of these returns can be estimated,
- the covariance of return for each pair of products can be estimated,
- the mean and variance of return are constant for the planning horizon,
- the company can define a return-risk indifferent curve,
management is able to choose an efficient subset of products that produces the highest
expected return with a given variance, or that produces a given expected return with the
lowest variance.
This mathematical approach deals with a basic marketing problem: which products
should the company have in its product line? However, the adaptation of the investment
theory has merely conceptual value rather than practical use and it presents a specific
way of thinking about choice problems. Due to the strict assumptions, this portfolio
approach is not applicable in practice.
In the early sixties several attempts were made to use product sales or the stages in the
product life cycle as a guideline for marketing strategy (see for example Mickwitz, 1959).
Marvin (1972) drew attention to the need for dynamic product portfolio management. He
recommended that each product of a company should be positioned both to its own life
cycle and the objectives of the product portfolios (product lines). However, much severe
criticism has been passed on the product life cycle (Cox, 1967;Polli and Cook, 1969;
Dhalla and Yuspeh, 1976; Wood, 1990). A major problem is that the recommended
strategies are given with little concern for the product’s profitability, its market share
position and other relevant factors. In the product life cycle-model, ‘time’ is the only
explanatory variable for the development of a product’s sales. Henderson (1970, 1973)
introduced the earlier discussed BCG-approach which could at the SBU-level serve as a
product portfolio model that could overcome this major limitation. The same holds for
the other multifactor portfolio models. In all these models factors with respect to market
attractiveness and competitive position are included in the analysis of the product
portfolio of a company.
Brown (1991) described the General Electric Business Screen as the most comprehensive
model for the overall assessment of competitive capability against an overall assessment
of market attractiveness. As Hofer and Schendel (1978) pointed out, for strategy
formulation purposes, the matrix is more appropriately divided into diagonal zones
rather than horizontal/vertical boxes. Brown (1991) developed this idea to a five-zone
business screen (see figure 2.4). For each zone matching strategies are elaborated:
- zone 1: ‘build’ by gaining market share or expanding the market;
- zone 2: ‘hold’ by defending market share or defending margins;
- zone 3: ‘build, hold, or harvest’, depending on the circumstances;
- zone 4: ‘harvest’ in a rapid or slow way;
- zone 5: ‘terminate’ by liquidating or divesting.
FIGURE 2.4 The five-zone business screen product portfolio

Source: Brown (1991: 104)
52
In marketing the eighties were the period of the rise of ‘positioning’ theory. While
segmentation identifies (homogeneous) groups of potential customers, positioning
studies show how these customers perceive the competing products. Ries and Trout
(1981) made clear that a product’s position is the way the product is defined by
customers on important attributes: the place the product occupies in customer’s minds
relative to competing products. Positioning shifted the emphasis of marketing theory
from the product to ‘the battle for your mind’. The number of products, affiliated with a
brand and the variance in quality among those products represent two basic properties of
a brand portfolio (Dacin and Smith, 1994). Adding products to a brand (brand extension) is
believed to weaken that brand. A brand becomes diluted by offering extensions.
However, extensions of existing product lines have accounted for over 90% of the new
consumer packaged goods offered every year (DelVecchio, 2000). If there is an acceptable
fit between a brand and the extension category, then it might be very beneficial to
capitalize on the brands’ risk reducing capabilities.
Positioning has obtained a central role within most textbooks on marketing management,
providing a bridge between the company and its target customers, describing how the
product differs from competing products (Hooley and Saunders, 1993). The positioning of
a product can be seen as an interplay of three major factors: market segmentation, the
competitive advantage, and the competitors’ offerings. The targeting of customers
determines where a firm shall compete; the competitive advantage determines how it
shall compete, in light of course of the competitors’ offerings. Product or product line
decisions should be taken in line with these strategic marketing factors. In the nineties,
the emphasis shifted again, but this time from classifying products to classifying
customers.
2.3.3 Classifying customers: from segmentation to customer

portfolios
The problem of classifying and dealing with different groups of (potential) customers is
one of the oldest and still one of the most important issues in marketing management:
marketing segmentation. Frederick (1934) used this fundamental marketing concept in one
of the earliest textbooks on marketing. Yet, it was not until the 1950s and 1960s that more
detailed research papers on this subject began to appear. Consumer marketing was the
first and main area of application of the new principles of market segmentation. Because
consumer marketing often had to deal with large numbers of (potential) customers, a
system of market segmentation with differentiated strategies and approaches for the
‘segmented’ groups appeared to be very useful. For long, an individual approach of
customers was not feasible in the case of large consumer markets. Nowadays, by means
of modern technology and new management methods, there are more possibilities for
creating variety and customization through flexibility and quick responsiveness (Pine,
1993): mass customization. By combining the principles of mass production and individual
consumption, the ultimate in market segmentation is achieved: one-to-one marketing

(Peppers and Rogers, 2001). In business markets it has always been clear that some
customers are more important than other customers. They are usually called ‘accounts’ or
‘key customers’. For analyzing the customer data base and for strategy development, the
portfolio approaches are known as account or (key) customer portfolio analysis.
Fiocca (1982) introduced an account portfolio analysis as a two-step classification,

tailored to industrial companies. The first step is aimed at identifying key accounts, based
on an assessment of the strategic importance and the difficulty of managing the account.
The output of the first step is a matrix on which the industrial seller can decide which
accounts need special attention and, as a consequence, deserve a more in-depth analysis.
The latter is performed in the second step. Each key account is analyzed on the following
two dimensions:
- customer’s business attractiveness, and
- the strength of the buyer/seller relationship.
The result is a matrix with much resemblance to the GE-matrix. The position in the 3x3
matrix is connected with the marketing strategy (improve, hold, or withdrawal) and the
profitability.
A similar approach is described by Homburg and Daum (1997) who also presented a
two-step customer portfolio model. Their first matrix is built up by a ‘customer
attractiveness’-dimension and a ‘business position’-dimension. A second customer
analysis is suggested aimed at identifying and enhancing the profitability of different
customer groups. A problem area is indicated for customers with a relative low
willingness to pay for relative high costs to serve. The best customers, from the marketing
point of view, are so called ‘passive customers’ who do not mind to pay much money in
combination with low costs to serve.
In the remaining of this chapter we will focus on the main types of customer portfolio
models. An examination of the various dimensions used in customer portfolio analysis
highlights the different approaches. Considering the dimensions and the purpose of the
various customer portfolio approaches, roughly two types can be distinguished with
corresponding main focuses:
1. profitability, balancing cost and (potential) revenues of different customers;
2. relationship, emphasizing various aspects of buyer/seller relationships.
Figure 2.5 provides an overview and a comparison of customer portfolio approaches in
(business) marketing, including the identification of the main focus. The list is arranged
in a chronological order.
54
Source dimension 1 dimension 2 main focus
Fiocca (1982) - difficulty in managing - strategic importance 3. profitability

(two step) - customer’s - strength of buyer/seller 2. attractiveness vs.
attractiveness relationship position (GE-like)
Cunningham and Homse sales volume technical interaction 1. interaction (IMP)

(1982)
Campbell and - number of buyers - number of suppliers 1. relationship (power

Cunningham (1983) - growth rate of - relative share of balance)
(two step) customer’s market customer’s purchases 2. attractiveness vs.
position (BCG-like)
Dubinsky and Ingram present profit contribution potential profit 3. profitability

(1984) contribution
Turnbull and Valla (1986) Attractiveness complexity of interaction 1. interaction (IMP)

(major vs. other accounts)
Shapiro et al (1987) net price cost to serve 3. profitability
Krapfel et al (1991) relationship value interest communality 4. relationship (type)
Dick and Basu (1994) relative attitude repeat patronage 4. relationship (loyalty)
Storbacka (1994) relationship revenues relationship profitability 3. profitability
Strandvik and Liljander relationship commitment relationship loyalty 4. relationship (strength)

(1994)
Homburg and Daum price sensitivity cost to serve 3. profitability

(1997) customer’s attractiveness business position 2. attractiveness vs.
(two step) position (GE-like)
Schijns (1998) perception of relationship actual relationship loyalty 4. relationship (strength)

strength
FIGURE 2.5 Overview and comparison of customer portfolio approaches in marketing
2.3.4 Customer profitability models
Many companies claim to be customer driven. Foster et al. (1996) found that,
paradoxically, most management accounting systems focus not on the customer, but on
products, departments, or geographic regions. It is clear that some orders are more costly
than others. Differences in customer profitability are likely to occur. Moreover, these
differences have proven to be enormous. According to Cooper and Kaplan (1991)
customer profitability based on activity based costing has forced managers in a number
of industries to redefine the traditional Pareto-20-80 rule (20% of the customers account
for 80% of the profitability) with a 20-225 rule. Here 20% of the customers account for
225% of the profitability, which indicates that some customers are very unprofitable. The
focus therefore should not be on sales, but on profitability.
The basic idea of any customer profitability analysis is that the supplier should seek to
assign all revenues and all costs to individual accounts. Revenue differences may arise
from a variety of sources, such as differences in prices, in selling volume levels, or in
price discounts. Foster en Gupta (1994) distinguished four types of costs: transaction
specific costs, customer specific costs, customer group costs, and marketing support cost.
Many costs of servicing customers are shared amongst several customers. It is quite a
problem how to allocate these overhead and other indirect costs to individual customers.
Dubinsky and Ingram (1984) presented a customer portfolio based on present profit
contribution and potential profit contribution of customers. Their portfolio consists of
four quadrants in which customers can be classified. The customers with a low present
profit contribution are classified as:
- ‘undesirable accounts’ with low potential profitability, or as:
- ‘undeveloped accounts with a high potential profitability.
On the other hand, customers with a high present profit contribution are labelled as:
- ‘desirable accounts’ in case of high potential profitability, or as:
- ‘developed accounts’ in case of low potential profitability.
The positions in the matrix are not static. Sales personnel should take action, trying to
move customers to a desired position. This customer portfolio approach is presented as a
tool which allows sales managers to analyze their present customer base. Moreover, the
tool suggests actions to effect a desired customer base.
According to Shapiro et al. (1987) it is useful to think of customers in terms of two

dimensions:
- net price realized, and
- cost to serve.
Price differences are usually well-known. However, the cost of serving customers and
filling orders can vary significantly too. Four types of costs were used to define the cost to
serve: presales costs, production costs, distribution costs, and post-sale service costs.
Marketing managers often assume that price-sensitive customers will accept lower
quality and lower service, while demanding customers are likely to pay more for a better
total product. However, these assumptions are rarely fully met. Shapiro et al. (1987)
claimed that in a wide variety of situations, they consistently observed a lack of
correlation between price and the cost to serve. This claim is confirmed by the results of
the Turnbull and Zolkiewski (1997)-study. The classification matrix of Shapiro et al.
(1987) is much quoted in the literature concerning customer profitability.
56
The grouping of customers as such is static. An additional analysis is therefore required

in which customers are followed over time. Doing so, typical migration patterns will
become visible and predictable. Often a relationship begins with customers who need
extensive service and support, but do not worry much about prices. Later however, as
customers gain experience with the product, they need less support and are likely to
become more price sensitive. Buyers may pressure the supplier for price reductions, even
while service requirement stay high. Turnbull and Zolkiewski (1997) confirmed that
monitoring migration patterns of customers with respect to the grid positioning is
extremely important for marketing management. Rangan and Bowman (1995) described
the marketing dynamics that turn each new product inevitably into a commodity. This
process, evocatively called ‘the commodity magnet’, is shown in figure 2.6. The pattern is
based on the regular product life cycle concept in which new products follow a similar
profit pattern. Rangan and Bowman (1995) argued that in the end only four feasible
strategies are possible:
1. Value-added strategy: increasing the price as well as augmenting services.
2. Process innovation strategy: decreasing the price and some cost-to-serve as well.
3. Market focus strategy: focusing on customers who would pay the additional price for
augmented services.
4. Service innovation strategy: mainly by decreasing the cost-to-serve. Like the market
focus, this strategy will lead to a reorientation of the customer base.
Price High - Selected products for selected markets

- Product/Market - Problem solving emphasis
expanison - Customer needs knowledge
- Competitive activity
Low - Knowledgeable - Product/Marked proliferation
customers - Marked volatility
- Aggressive customers
Low High
Cost-to-serve
FIGURE 2.6 Market life cycle: alternative framework

Source: Rangan and Bowman (1995: 140)
Customer portfolio models have received a great deal of attention from management and
from the academic world. However, we found little empirical research either to support
or to reject customer portfolio models. There are some exceptions that describe the
problems and results of the actual use of customer portfolio models. Turnbull and Topcu
(1994) tested the Fiocca-approach, identifying problems with data calculation, problems
with the interpretation of concepts and problems with the demarcation between ‘low’
and ‘high’ values. Yorke and Droussiotis (1994) also undertook an empirical study to test
the Fiocca-matrix. They reported subjective elements in the model, which could cause
problems when actually undertaking the analysis. Turnbull and Zolkiewski (1997)
decided to test Shapiro et al.’s (1987) theories relating to gross margin dispersion. The
research seems to validate Shapiro et al.’s (1987) findings on the wide range of profit
dispersions and the applicability of their customer classification matrix. The observations
suggested that the migration of customers over time is more complex than has been
suggested by Shapiro et al. (1987). Zolkiewski and Turnbull (2001) conclude that the
concept of relationship portfolios provides both scope for academic investigations and
managerial prescriptions, especially as an aid for strategic decision making. However,
there are some common limitations to customer portfolio models:
- the difficulty in collecting the appropriate data an d the time it takes;
- the achievement of year-on-year consistency of data;
- the exact meaning of the axes (there is a reasonable amount of variability in the
suggested means of customer profitability and cost-to serve; moreover, there are different
interpretations of ‘difficulty in managing an account’);
- in business-to-business markets there is often a lack of accurate market data such as
percentage of market share by various firms, implying that some of the axes can not be
readily calculated.
- the demarcation problem: it is not defined what value represent ‘low’ and ‘high’ cost to
serve, which makes it arbitrary to position customers in a matrix.
2.3.5 Customer relationships models
Since the early eighties much academic attention has been paid to relationships in
marketing. Developments in information and communication technology have made it
possible to communicate and interact more directly with individual customers. This
‘relationship marketing’ focuses on attracting, developing and retaining long-term
customer relationships (Berry and Parasuraman, 1991). According to some marketing
scientists, like Grönroos (1990, 1994), Gummesson (1987) and Webster (1992), there is a
contrast to transactional marketing with a supposed focus on discrete transactions.
Although a case can be made against the difference between relationship marketing and
transactional marketing (see for instance Gelderman and Tuninga, 1998), it is clear that
the renewed attention on relationships in marketing has lead to the development of
customer portfolio matrices, based on various aspects of customer relationships. Some of
them are inspired by the body of thought of the IMP-group (International Marketing and
Purchasing). The objective of the IMP research group was a better understanding of the
nature of buyer/seller relationships in industrial markets. The well-known interaction
approach to marketing and purchasing emerged form this research, emphasizing the
active roll of both buyer and seller (Håkansson, 1982). As exponents of the IMP-group,
Cunningham and Homse (1983) presented a customer portfolio analysis which focuses
58
upon the nature of the interactive relationship between supplier and customer
companies. They combined two evaluation criteria: the sales volume and the (expected)
benefits from technical co-operation. The result is a classification into four categories. The
management of customer relationships involves the planning and handling of personal
contacts between staff in the supplier company and their counterparts in the customer
companies. Therefore, resource constraints must be taken into account. What kind of
interaction and attention should be rendered to the various customers?
Krapfel et al. (1991) constructed a relationship matrix based on two dimensions: interest
commonality and relationship value. When the goals of the buyer and the goals of the
seller are compatible, interest commonality is high, and vice versa. High interest
commonality is believed to lead to a more co-operative attitude, and vice versa. The
relationship value embodies the factors from which dependence on a specific customer
relationship flows. In their relationship type matrix four possibilities are described:
1. partner - a relationship having high economic value coupled with high goal
compatibility
2. friend - a relationship with low current economic value, but high interest commonality
3. rival - the current economic value of the relationship is high, but a strategic choice has
been made to pursue self gain, because interest commonality is low
4. acquaintance - an exchange partner with low current economic value and low interest
commonality.
The authors concluded that the relationship management style should be varied
according to the perceptions of power and interest commonality. In contrast with all
other customer portfolio models, the relationship matrix explicitly includes the interests
of the buying companies. This approach in marketing is obviously to be the exception to
the rule.
One of the main tasks in (consumer) marketing is often viewed in terms of customers’
loyalty toward the products or services of a company. Generally speaking, loyalty has a
behavior dimension (i.e. repeat purchases) and a perception dimension (i.e. attitude).
Taking this as the starting point, Dick and Basu (1994) distinguish four different forms of
customer loyalty, based on the relative attitude and the repeat patronage. Strandvik and
Liljander (1994) dealt with a comparable question, namely of measuring relationship
strength. They proposed a multidimensional measure consisting of relationship
commitment and relationship loyalty. The research of Schijns (1998) was also aimed at
the measurement of the strength of relationships between consumers and organizations.
After the example of Krapfel et al. (1991) customers were categorized as friends,
acquaintances, sympathizers, or functionalists.
The buying behavior and the attitude of customers are obviously significant aspects of
customer relationships, though especially from a marketing perspective, not from a
purchasing perspective. Relationship strength and customer loyalty are typical examples
of marketing performance measures that contrast with direct customer profitability
measures.
Most studies on relationship models are conceptual by nature. We found a small number
of empirical studies, illustrating the possibilities of relationship models in practice.
Strandvik and Liljander (1994) performed a small-scale study which resulted in a
classification of customers according to their loyalty and commitment to the bank.
Turnbull and Zolkiewski (1997) tested the Krapfel et al. (1991) matrix. As the studied
relationships were either repeat or follow-on purchases, all customers were positioned in
two of the four quadrants in the matrix: partner and friend. Therefore, the scope of the
study was rather limited as well. Schijns (1998) presented the results of his empirical
study which was aimed at measuring the strength of relationships between consumers
and organizations. Members of two organizations were approached, which resulted in
more than 1,200 respondents. It was found that customer relationships could be classified
and monitored by means of a relationship perception/relationship loyalty matrix.
Section 2.3.4 listed some common problems with customer portfolio models: data
problems, the meaning of dimensions, and demarcation problems. Obviously, these
problems apply to the relationship models as well. In addition, there is no substantial
empirical evidence on the adoption and effectiveness of relationship models. The
empirical studies merely confirmed the possibility to segment a database of customers
into the segments of a portfolio matrix. These studies illustrated some conceptual and
measurement problems as well.
2.3.6 Conclusions
In this section we have identified and described the most common portfolio models in
marketing management. These models were classified into three groups: models for new
products or R&D projects, product (line) portfolio models, and customer portfolio
models. The last group was subdivided in models with a profitability focus and models
with a relationship focus. We began this study of marketing portfolio models with the
proposition that they would be of special interest for purchasing, considering the fact that
a buyer and a supplier represent both sides of the same business relationship. Regrettably,
to a large extent this proposition has to be rejected. Product portfolio models are limited
to the issue of resource allocation: which projects should be invested in and which
products are expected to produce economic value? These models show much
resemblance to those for selecting a financial portfolio of investments, with no significant
role for trading partners. In addition, there are many marketing portfolio models
preoccupied with customer profitability, balancing costs and financial benefits (e.g.
Fiocca, 1982; Shapiro et al., 1987; and Homburg and Daum, 1997). The customer input is
mainly limited to cost related issues such as ‘cost to serve’, ‘price sensitivity’ and
60
‘difficulty in managing’. Surprisingly, the purchasing perspective was seldom found in

customer relationship models as well. As an exception should be mentioned the Krapfel
et al. (1991) relationship matrix which explicitly includes interest commonality between
trading partners: the level to which the goals of the seller and the buyer are compatible. A
positive relationship was assumed between interest commonality and the likelihood of
co-operative attitudes and behaviors.
We have found many different customer portfolio models in the field of marketing
management. By taking a closer look, they all seem to deal with the same question of how
to allocate different amounts resources to different types of customers. Examples of these
varying resources are attention, investment, service, technical support and interaction. Of
course, these kinds of allocations decisions are to be justified by an economic criterion, i.e.
profitability. Most models asses the relative importance of customers, one way or
another, in order to justify choices to be made by marketing management with respect to
the allocation of resources. Customer portfolio models are heavily concerned with
profitability of individual customers. From a purchasing perspective, it is striking that
none of the customer portfolio models explicitly include a dimension or factor with
respect to purchasing and supply strategies of buying organizations.
Marketing is supposed to be customer orientated. However, most management
accounting systems do not focus on customers, but on products, departments, or
geographic regions (Foster et al., 1996). Perhaps due to difficulties and costs of gathering
data on individual customers, many companies do not dispose of the necessary
information that is needed to use a profitability model. In contrast with the increasing
number of customer portfolio models, it seems that they are not widely adopted.
With a few exceptions, customer portfolio models are not investigated by means of
empirical study. Most authors introduce new models or versions of existing portfolio
models.
Customer relationship models do not always reveal their theoretical assumptions.
Customer portfolio models are rarely tested on their assumptions and effectiveness. This
applies especially for the customer relationship models. The scope of most studies is rather
limited, including a relatively small number of respondents. We have found a small
number of empirical studies to the use and effectiveness of customer profitability models as
well. The studies revealed the following problems and limitations:
- practical problems with data calculation and the categorization of customers (Turnbull
and Tupcu, 1994);
- subjective elements in the Fiocca-model (Yorke and Drioussiotis, 1994);
- the arbitrary definition of ‘high’ and ‘low’ cost to serve (Turnbull and Zolkiewski, 1997).
To conclude, there are common issues to the implementation of customer portfolio

models, including the time taken to collect the appropriate data, the interpretation of
dimensions and concepts, and demarcation problems.
2.4 Concluding remarks

The main purpose of the literature study to portfolio models in related disciplines, was to
learn from areas with a longer experience than purchasing management. In this
concluding section we will list the main ‘lessons’ that we have learned from other
disciplines.
Investment theory
- There are many differences in scope and perspective between investment portfolio
management and purchasing portfolio management (items, measurement,
recommendations, market dynamics and human behavior).
- Purchasing managers should be aware of the trade-off between risks and returns (an
increase in risk for instance should be matched by a higher reward for the company).
Strategic management
- Every portfolio model for strategic management is a tool, not a comprehensive theory.
- Recommendations derived from portfolio analysis should not be followed without
additional considerations and managerial judgment.
- Classifications are sensitive to the choice of dimensions, variables and weights
(specification) and the rule to divide a dimension into a ‘low’ and a ‘high’ category
(demarcation).
- The customization of a portfolio model leads to a better understanding of the strategic
issues at hand, it provides a structure for analysis and facilitates the communication and
sharing of judgments and assessments.
Marketing management
- Product portfolio models are limited to the issue of resource allocation with no
significant role for trading partners.
- Customer portfolio models are preoccupied with customer profitability, again with no
significant role for trading partners.
- When using a portfolio model, one has to deal with some common problems and issues:
difficulties of collecting appropriate data, the interpretation and operationalization of
dimensions, and the (arbitrary) demarcation between ‘high’ and ‘low’ categories.
With these insights in mind, we will start the next part of our literature study which is
devoted to the review of portfolio approaches in purchasing management.
3 Portfolio approaches in purchasing

management
In this chapter we will discuss the main portfolio approaches in purchasing and supply
management. In chapter 1 we have posited that a portfolio model is a tool that combines
two or more dimensions into a set of heterogeneous categories for which different
(strategic) recommendations are provided. Classification tools, like the ABC-analysis, fall
outside the scope of this definition.
According to the three basic elements of any portfolio model, the portfolio approaches in
purchasing will be discussed and evaluated on (1) dimensions, (2), categories, and (3)
strategic recommendations. Because we are interested in issues like the general purpose,
the reported acceptance and adoption, empirical support for effectiveness, and more
general possibilities of a portfolio approach in purchasing, we will also pay attention to
(4) use issues.
At the end of this chapter we will return to the problem statement of this study,
connecting the findings and conclusions of the literature study to the research questions.
This will result in the answers to some of the research questions and to the more
elaborated and articulated (sub)questions, suitable for further research in the case studies
and the survey.
3.1 The Kraljic portfolio approach

In the 1983 Purchasing Conference in Copenhagen, Kraljic, director in the Düsseldorf
office of McKinsey Company, presented a new and promising instrument for the
determination of a set of differentiated purchasing strategies and a policy for the more
fundamental restructuring of the portfolio as a whole. The Harvard Business Review
published his seminal paper “Purchasing must become supply management”, which
pointed at the need for companies to progress toward more effective supply
management, accompanied by a practical portfolio tool for ‘shaping the supply strategy’.
By now the HBR-1983 contribution has probably become the most cited and referred to
article in the field of purchasing and supply (chain) management. It is noted that Kraljic
(1977) presented exactly the same concepts and ideas much earlier, albeit in a German
business journal (‘Beschaffung aktuell’) that did not get much attention in the
international business and academic community. In a recent interview Kraljic explained
that the matrix was developed for Basf in the early seventies (approximately 1973, 1974),
within the context of a large cash management
64
project. Purchasing management was just one of the involved business functions. As a
McKinsey consultant he was asked to develop a new tool for purchasing, similar to the
recently introduced marketing matrices, e.g. BCG matrix (Gelderman and Van Haaster,
2002).
Kraljic (1983) proposed a four-stage approach as a framework for ‘shaping the supply
strategy’:
1. Classify all the purchased materials or components in terms of profit impact and
supply risk.
2. Analyze the supply market for these materials.
3. Determine the overall strategic supply position.
4. Develop materials strategies and actions plans.
Phase 1: classification
On the basis of two dimensions Kraljic (1983) classified all materials and components:
profit impact and supply risk. The profit impact of a given item can be defined in terms
of the volume purchased, percentage of total cost, or impact on product quality or
business growth. Supply risk is a more complex composite dimension. It is assessed in
terms of availability, number of suppliers, competitive demand, make-or-buy
opportunities, and storage risks and substitution possibilities.
Each dimension has two possible values: ‘low’ and ‘high’. The result is a 2x2 matrix and a
classification in four categories:
- strategic items (high on both dimensions)
- bottleneck items (low profit impact, high supply risk)
- leverage items (high profit impact, low supply risk)
- non-critical items (low on both dimensions).
Each of the four categories requires a distinctive approach, in proportion to the strategic
implications. Kraljic identified main tasks, the required information and the decision
level in organizations for each category, see figure 3.1.
Procurement Main tasks Required information Decision level

focus
Strategic - Accurate demand forecasting - Highly detailed market data Toplevel

items - Detailed market research - Long-term supply and (e.g. vice president
- Development of long-term supply demand trend information purchasing)
relationships - Good competitive intelligence
- Make-or-buy decisions - Industry cost curves
- Contract staggering.
- Risk analysis.
- Contingency planning
- Logistics, inventory and vendor
control
Bottleneck - Volume insurance (at costs - Medium-term supply demand Higher level
items premium if necessary) analysis (e.g. department
- Control of vendors - Very good market data heads).
- Security of inventories - Inventory costs
- Backup plans - Maintenance plans
Leverage - Exploitation of full - Good market data Medium level

items purchasing power - Short to medium term (e.g. chief buyer)
- Vendor selection demand planning
- Product substitution - Accurate vendor data
-Targeted pricing - Price/transport rate of
strategies/negotiation forecasts
- Contract/spot purchasing mix
- Order volume optimization
Noncritical - Product standardization - Good market overview Lower buyer

items - Order volume monitoring/ - Short-term demand forecast (e.g. buyers)
optimization - Economic quantity inventory
- Efficient processing levels
- Inventory optimization
FIGURE 3.1 Classifying purchasing materials requirements

Source: Kraljic (1983: 112)
Non-critical items require efficient processing, product standardization, order volume

and inventory optimization. Leverage items allow the buying company to exploit its full
purchasing power, for instance by tough negotiating, target pricing and product
substitution. Bottleneck items on the other hand cause a lot of problems and risks.
Volume assurance, vendor control, security of inventories and backup plans are
recommended. Finally, the strategic items for which a main task in terms of ‘developing
long-term supply relationships’ attracts attention. For the strategic items additional
analytic techniques, including market analysis, risk analysis, price forecasting, computer
simulation and optimization models, might be used.
66
However, in the following three phases the Kraljic approach concentrates on the items
identified as strategic in the classification-phase.
Phase 2: market analysis

In phase 2 the strengths of the suppliers are weighted against the company’s own
strength as a customer. The following criteria are just examples of possible factors, no list
will be equally applicable to every industry, said Kraljic (1993: 113).
Indicators for the ‘supplier strength’ Indicators for the ‘company strength’
- market size versus supplier capacity - purchasing volume versus capacity of main units
- market growth versus capacity growth - demand growth versus capacity growth
- capacity utilization or bottleneck risk - capacity utilization of main units
- competitive structure - market share vis-à-vis main competition
- ROI and/or ROC - profitability of main end products
- cost and price structure - cost and price structure
- break-even stability - cost of non-delivery
- uniqueness of product and technological - own production capability or integration depth
stability
- entry barrier (capital and know-how - entry cost for new sources versus cost of own
requirements) production
- logistics situation - logistics
Phase 3: strategic positioning

After analyzing the market, where the company weighs the bargaining power of its
suppliers against its own strength, the company positions the strategic items in a
purchasing portfolio matrix. The matrix shows the relative position of the company and
the advised overall purchasing strategy. It depends on the relative power position if a
company should choose for an aggressive strategy (‘exploit’), for a defensive strategy
(‘diversify’) or for a well-balanced strategy (‘balance’); see figure 3.2.
Company strength
High exploit exploit balance
medium exploit balance diversify
Low balance diversify diversify
low medium High

Supply market strength
FIGURE 3.2 The purchasing portfolio matrix

Source: Kraljic (1983: 114)
Phase 4: action plans

Each of the three strategic thrusts has distinctive implications for the individual elements
of the purchasing strategy, such as volume, price, supplier selection, material
substitution, inventory policy, and so on. Facing a dominant supplier the company must
go on the defence. It may have to increase spending on market research or supplier
relations, or even consider backward integration. In short, the company may want to
generate and use new supply options. The company could consolidate its supply position
by concentrating fragmented volumes in a single supplier, accept high prices, and cover
the full volume requirements through supply contracts. To reduce the long-term risk of
dependence on a single source, however, the company should also search for alternative
suppliers or materials, or even consider backward integration (‘diversify’). When
bargaining from weakness the company may have to offer long-term contracts and accept
higher prices in order to ensure an adequate supply. On the other hand, if the buying
company is stronger than the suppliers, it can bargain and act from a position of strength.
The company could press for preferential treatment. It can spread volume over several
suppliers, exploit price advantages, increase spot purchases, and reduce inventory levels.
To conclude, with no dominant party at hand, a well-balanced intermediate strategy is
advised.
In phase 4 the company should explore a range of supply scenarios. The end result
should be a set of documented strategies for critical purchasing materials that specify the
timing of and criteria for future action.
3.2 Discussion of the Kraljic approach

In this section the Kraljic approach will be discussed and evaluated, focussed on:
- dimensions,
- categories,
- strategic recommendations,
- and use issues.
3.2.1 Dimensions
(a) Theoretical foundation

A general accepted view on the purchasing function is that it should prevent disruptions
in production and other activities. Some authors describe the general objectives of
purchasing and supply management in terms of ‘the five rights’: professional purchasing
should achieve the acquisition of materials: of the right quality, from the right supplier,
in the right quantity, at the right time, and at the right price (Dobler and Burt, 1996). The
‘right supply’, the right time, and the right place refer to logistic aspects of supply
management. Products should be bought for the ‘right price’, which refers to commercial
and financial aspects of purchasing. An attractive feature of the Kraljic portfolio approach
68
is that it encompasses two key variables with respect to these key aspects: ‘profit impact’
is linked to commercial requirements, ‘supply risk’ is (among others) related to logistic
issues. However, it is not clear why these particular dimensions are selected for use in
Kraljic’s portfolio approach.
We have to conclude that Kraljic’s article does not provide any reference to a theoretical
foundation or comprehensive perspective. In his article Kraljic offers a basic tool for
purchasing management, albeit it without any reference to literature or documented
evidence. The tool is developed for practical use. The combination of the two dimensions
is intended to “minimize supply vulnerability and make the most of potential buying
power” (Kraljic, 1983: 112). Kraljic’s article does not provide any reference to literature,
other research or a theoretical foundation. In a recent interview Kraljic acknowledged
that the selection of dimensions was based on discussions with purchasing professionals,
in search of ‘things that really matter in purchasing’. Basically, a matter of common sense
(Gelderman and Van Haaster, 2002). This confirms the idea that there is no clear
reference to a comprehensive theory to make a reasonable case for using ‘profit impact’
and ‘supply risk’. This lack of background needs to be further investigated to provide a
theoretical foundation to the Kraljic matrix.
(b) Measurement issues

In general, decisions based on portfolio models are proven to be sensitive to the choice of
dimensions, factors, and weights. There is a demarcation problem with respect to the
measurement of key variables. What is the exact distinction between ‘a high’ and ‘a low’
supply risk? Homburg (1995) and Heege (1981) concluded that the classification of
products is therefore rather arbitrary. Olsen and Ellram (1997) emphasized that the
weighting of each factor is the most important part of the implementation process, but at
the same time very subjective. The decision-makers must come to an agreement on the
relative importance of each factor. De Boer (1997) suggested a fully customized approach:
organizations should determine their own criteria as well as their own specific threshold
values. It is as yet unclear how this determination process should or could take place.
To make things worse, what is exactly meant by ‘profit impact’ and ‘supply risk’? These
concepts are actually made up of a number of “nebulous concepts without operational
dimensions”, as formulated by Ramsay (1996a: 15). Nellore and Söderquist (2000)
pointed out that there is a risk that the variables used in portfolio analysis might not be
accurate proxies for the dimensions they are supposed to measure. Day (1986) concluded
that measurement is considered to be the Achilles’ heel for all portfolio models.
(c) Interdependencies
Olsen and Ellram (1997) observed that the literature on buyer-supplier relationships
tends to focus on a single relationship or type of relationship, ignoring or downplaying
the important interdependencies between relationships and the important task of
allocating scarce resources between relationships. The authors have interpreted this fact
as a need for the development and use of portfolio models for the management of
supplier relationships. Dubois and Pedersen (2002) agreed that the concept of
interdependence between relationships is seldom discussed, especially in portfolio
models. When the concept is brought up, it has to do with the costs of dealing with
different supplier relationships (e.g. Olsen and Ellram, 1997) or with the relationship
specific investments of both parties (e.g. Bensaou, 1999). Interdependencies are being
discussed in relation to the issue of the optimization of scarce resources (resource
allocation). From their IMP-network perspective Dubois and Pedersen (2002: 40) pointed
at “other kinds of interdependence among supplier relationships or in relation to other
parties in the network context of the buying and the supplying firm, of which the
relationship is a part”. Heege (1981) stated that the value of portfolio methods is limited
because these models are not appropriate for the development of strategies for a whole
set of products, but rather for individual products (aggregation issue). According to
Homburg (1995) the most important weakness of portfolio models is the disregard for
interdependencies, for connections between products in the matrix
(“Systemzusammenhangs”). For each category recommendations are provided in
isolation of the recommendations for other categories. For that matter, comparable
criticism has been registered with regard to marketing portfolio models which tend to
focus on product level decisions as well. Product portfolio models in marketing are said
not to recognize possible interdependencies on customer levels, and in particular the fact
that customers buy or could buy different products from the same company (cross
selling).
In general, the portfolio concept stresses the importance of the whole rather than the
parts, where it should focus on the interdependencies among management decisions
(Turnbull, 1990). However, the alleged lack of attention for interdependencies in
purchasing portfolio models can be connected with the ascribed rigor of the application
of the strategic recommendations. Some authors hold a rather deterministic view on these
recommendations, neglecting the necessity of a customized approach and neglecting the
necessity of a reflective and critical attitude of the portfolio user. In other words, no
portfolio model should be used or seen as a deterministic model from which rather
mathematical strategies or strategic recommendation can be deduced. The issues and
questions with respect to portfolio-based recommendations are posited and elaborated in
section 3.2.3. In advance, it is clear that portfolio use in purchasing should include
attention for the supplier’s side and for the nature of and the connection between
(differentiated) actions and strategies.
(d) Confusion over the first matrix-dimensions

Many authors refer to the Kraljic approach as being a single portfolio matrix, based on
the dimensions ‘strategic importance’ (or just ‘importance’) and ‘complexity of the supply
market’, see for instance Kamann (2000a), Olsen and Ellram (1997), and Lilliecreutz and
70
Ydreskog (1999). It should be noted that Kraljic himself is partly responsible for this
confusion over the names of the dimensions. Before introducing the portfolio approach,
Kraljic (1983: 111) presented a figure in a matrix format (“Exhibit 1”) that uses
‘importance of purchasing’ and ‘complexity of supply market’ as axes. However, this
figure is not a part of the portfolio approach, it does not classify product categories, nor
does it provide strategic recommendations. The name of the picture clarifies the purpose
of the matrix: ‘stages of purchasing sophistication’. There are four stages: (1) purchasing
management, (2) materials management, (3) sourcing management, and (4) supply
management. Kraljic (1983: 110) argued that the greater the uncertainty and the
vulnerability, the more important supply management becomes. Facing unimportant
purchases in not complex supply markets, a less sophisticated procurement focus will do.
3.2.2 Categories
(a) The focus on one category

Kraljic is concerned about disastrous supply disruptions of vital materials. From this
perspective it is logical that Kralic should focus on strategic items with a high profit
impact and a high supply risk. The second matrix only applies to the strategic items. For
the other categories a list is provided merely of main tasks, the required information and
the advised decision level. The conclusion is that these categories are rather disregarded.
However, in practice the picture will be completed by thinking through possible plans
and strategies for bottleneck, leverage and non-critical items. In the next section we will
present an elaboration of strategies for all categories, derived from Van Weele (1992,
1994). As will be concluded, other portfolio models in purchasing management make
intensive use of Kraljic basic ideas. They too, can be seen as elaborations of the original
Kraljic model.
(b) The role of power and dependence

It is generally agreed that the role of power and relative dependence is important in
understanding exchange relationships. However, the Kraljic approach does not explicitly
deal with issues of power and dependence. Some of the strategic recommendations
obviously refer to the prevailing power structure (‘exploit power’), others do not. In the
first matrix it is not clear in what way ‘profit impact’ and ‘supply risk’ are related to the
relative power position of the buying company. There is no unambiguous relationship
between the two dimensions and issues of power and dependence. Presumably, the
buyer is more powerful than the supplier in the case of leverage items, while the opposite
might be true for bottleneck items. In the second matrix, for the strategic item categories,
the role of power is more clearly: a buyer-supplier relationship can be balanced, buyer-
dominated or supplier-dominated. The Kraljic approach deals in a rather implicit way
with issues of power and dependence. The application implies dealing with two matrices
without being clear about the role of power and dependence.
3.2.3 Strategic recommendations
(a) The supplier’s side

Often the suppliers’ side of the buyer-seller relationship is considered as a disregarded
element in Kraljic’s model. The Kraljic approach does not explicitly take into account the
possible strategies and reactions of suppliers (e.g. Heege, 1981; Kamann, 2000b). In a
critical review of the Kraljic-approach Dubois and Pedersen (2002: 35) argued that
purchasing portfolio models using “given products” as a point of departure, in addition
to a dyadic perspective, may be counterproductive where purchasing efficiency is
concerned. Nellore and Söderquist (2000) confirmed that it is imperative for any portfolio
use to indicate the characteristics of the supplier with regard to the specification
generation, the required relationship and the required type of specification for a given
component. The design of a product entails issues that are not explicitly considered in
portfolio models. Obviously, whether the product is developed by the supplier, by the
customer, or developed jointly impacts on the relationships between parties (Araujo et
al., 1999). Mismatches between buyer and seller are likely to occur if one does not take
into account how a supplier (i.e. a marketing of sales manager) assesses the situation.
And vice versa, of course. A partnership is only possible if that is the strategic intent of
both parties. Organizations must match their intentions and strategies.
Unquestionably, the supplier’s side should be included in any strategic thinking in the
field of purchasing and supply management. Different solutions have been proposed for
this issue, although it should be said that the Kraljic approach does not imply the neglect
of the supplier’s side. The impression might be nourished by Kraljic’s focus on supply
vulnerability, threats of materials scarcity and the situations on supply markets (Kraljic,
1983: 109). Lilliecreutz and Ydreskog (1999) proposed an additional evaluation of
suppliers according to three dimensions: performance assessment, relation characteristics
and network position. Purchasing strategies should bridge the gaps between the actual
and the desired co-operation with different suppliers. The recommendation to include the
supplier’s perspective can be found in many other sources. A rather different approach
however, is the combination of a classical purchasing portfolio with a corresponding
customer portfolio. To our knowledge Carter (1995) introduced this new type of matrix,
combining a purchasing’s view as well as a supplier’s view in a single graph. The basic
idea is that the purchasing professional puts himself in the position of the suppliers,
assessing his own company from the perspective of the suppliers. In Carter’s approach
companies assess their customers on two attributes: the attractiveness of the customer’s
account and their own competitive position in that market-place. The categories are:
- nuisance (giving low attention to non-attractive customers from a weak position)
- exploit (charging premium prices to attractive customers from a strong position)
72
- develop (nurturing attractive customers from a weak position)

- core (defending rigoursly the relationships with attractive customers from a strong
position).
Purchaser’s view
Nuisance Exploit Develop Core
Strategic
Bottleneck
Leverage
Non-critical
Supplier’s views
FIGURE 3.3 The combination of the purchaser’s and the supplier’s view
Source: Carter (1995: 47)
The most appropriate approach to adopt with each supplier can be determined by
combining the assessment of the supplier’s view with the own purchasing Kraljic matrix.
For example, if a product is a strategic or bottleneck item from a purchasing perspective,
but the account is categorized by the supplier as nuisance or exploitable, any aggressive
or confrontational behavior could result in the supplier withdrawing from the situation,
leaving the buying company with a serious problem. The key skill would be the ability to
influence people and gain favored customers status (Carter, 1995), Alternatively, should a
product be strategic in the Kraljic matrix and also be core for the supplier, there will be
possibilities in forming a close relationship and a partnership. Carter’s idea can also be
found in a number of (Dutch) publications, namely Kempeners and Van Weele (1997),
Verheul and Santema (1997), Rietveld (1998) and Van Weele and Rozemeijer (1999).
(b) Influencing the power balance: dynamics in the matrix

Lilliecreutz and Ydreskog (1999) stated that “strategies that are based solely on Kraljic’s
matrix lack the dynamics of the power that the supplier can obtain”. They too stressed
the importance of taking the supplier’s situation into account. Kempeners and Van Weele
(1997) pointed at the natural conflict of interests in buyer/seller relationships. Both are
likely to prefer a dominant power position due to the attached benefits. As a result,
positions in the Kraljic matrices will always be amendable to the dynamics of buyer/seller
relationships. Parties are inclined to seek for possibilities of influencing their relative
power position. Cox (2001) explicitly posited that a sufficient condition of success would
be the ability to find ways to move from current positions of power to other more
favorable positions. The Kraljic framework however, does not provide guidelines for
movements within the matrix. Under which conditions is it advisable and feasible to
pursue movements in the (first) matrix? How should those movements be accomplished?
There is no explicit connection between the strategic recommendations in the (first)
Kraljic matrix, prevailing business conditions, purchasing goals and purchasing
strategies.
(c) The nature of the strategic recommendations

Some argue that portfolio models are not designed to deal with proactive thinking about
new opportunities (Day, 1986; Cox, 1997). Portfolio analysis is said to provide merely
guidance for operation action. The methodology describes what is, it does not provide us
with any proactive thinking about what can, or should be done to change the existing
reality of power (Cox, 1997). For this matter, we only partly subscribe to this firm
conclusion.
In the first matrix the recommendations are quite reactive by nature. They react and
adapt to the prevailing structure of power in buyer-supplier relationships. It is not clear if
and how other positions in the matrix are to be pursued through the implementation of
the recommended ‘main tasks’. However, the second matrix - for the strategic items - is
especially designed to develop counterstrategies vis-à-vis key suppliers (Kraljic, 1983). By
plotting the buying strengths against the strengths of the supply market, three basic
power positions are identified and associated with three different supplier strategies:
balance, exploit, and diversify. This ‘reverse marketing’-approach intends to identify
areas of opportunity or vulnerability, assess supply risks, and derive basic strategic
thrusts. It should be noted however, that the strategies are rather generic by nature,
providing only rough indications for the most appropriate supplier strategies.
Discussing Kraljic and purchasing portfolio models, most publications are limited to the
first matrix with the well-known categories: strategic, leverage, non-critical and
bottleneck. It might be assumed that all strategic products are to be managed by means of
(strategic) partnerships. Looking at the second matrix, we must conclude that this was
certainly not Kraljic’s intention.
3.2.4 Use issues
(a) General purpose

The introduction of the Kraljic portfolio approach can be considered as a major
breakthrough in the development of professional purchasing. The method is an
important step beyond the relative simple ABC-analysis. The Kraljic approach filled the
gap between the developing practice and the lagging theory in the field of purchasing.
The portfolio approach is a useful tool with diagnostic and prognostic powers. It
provided a comprehensive view of the large number of items that companies usually buy.
The general idea of the Kraljic approach is to “minimize supply vulnerability and make
74
the most of potential buying power” (Kraljic, 1983: 112). By categorizing these items,
sensible guidelines are derived for the management of supplier relationships. It can be
concluded that Kraljic’s approach represents the most important single diagnostic and
prescriptive tool available to purchasing and supply management (Syson, 1992).
However, others find the Kraljic approach counterproductive, providing
recommendations either to exploit power (if the customer is in power), or to avoid risk
associated with the supplier exercising power (Dubois and Pedersen, 2002). Some argue
that the complexity of business decisions does not allow for simple recommendations.
How could one deduce strategies from a portfolio analysis that is based on just two basic
dimensions? (Heege, 1981). However, a technique should not be criticized because it
simplifies, but rather because it focuses on unimportant factors. After all, the logic for
portfolio techniques is in the first place a tool for management so that complex problems
can be simplified and solved in an acceptable way. The strength of any portfolio model is
that it breaks down a complex situation to its basic dimensions.
(b) Possibilities for customization

An important benefit is that the actual using and customizing of a portfolio model will
lead to a better understanding of the strategic issues at hand. The process of customizing
and using a portfolio model is in itself a strategic thinking process (Wind and Mahajan,
1981). During this process the decision makers have to discuss inconsistencies among
themselves and agree on different elements of the portfolio (Olsen and Ellram, 1997).
Portfolio models provide a structure for analysis that facilitates the communication and
sharing of judgements and assumptions about strategic issues (Day, 1986). Kraljic’s
portfolio approach allows for sufficient customization. Kraljic introduced matrices and
dimensions that are described in general terms, allowing for customized use. The
dimensions ‘profit impact’ and ‘supply risk’ (in the first matrix) can be defined in
different ways. For the second matrix, exemplary lists of possible factors are provided, to
be adapted to individual requirements and wishes. Kraljic (1983: 113) clearly stated that
“no list of evaluation criteria is equally applicable to every industry”.
(c) Acceptance and adoption

Kraljic (1983) made a reasonable case for the usefulness of the portfolio approach by
describing the experiences of four large industrial companies: a welding materials
producer, a maker of electrical equipment, a chemical company and a heavy-equipment
maker. The adoption of the portfolio analysis has been a rather gradual process. In the
eighties the Kraljic matrix received hardly any attention, with some notable exceptions:
Heege (1981) in Germany, Elliott-Shircore and Steele (1985) in Great Britain, and Van
Weele (1988) in the Netherlands. It might be expected that the speed of adoption will be
rather low. Even in the early nineties Homburg (1995) suggested that purchasing
portfolio models are not very popular with purchasing (“Beschaffungsportfolios erfreuen
sich, unseres Wissens, noch keiner sehr großen Beliebtheit”). Not until the late nineties
many more publications refer to a portfolio approach in purchasing, indicating higher

levels of acceptance and adoption. Now, many years after its introduction, the
purchasing portfolio approach is being used by several other large companies, for
instance Shell, Alcatel, Philips en Siemens (Van Weele, 2000), DSM (Louwers et al., 1999)
and Océ van der Grinten (Wynstra, 1998). In a survey of Dutch companies Boodie (1997)
found that 44% of the responding purchasing managers said that they used Kraljic for
formulating purchasing strategies. A peak of 80% was found for industrial companies
that operate on a mass production basis. Kamann (2000a) estimated that 20% of the Dutch
companies uses a portfolio method for the management of their supplier base. Lamming
and Harrison (2001) stated that Kraljic’s matrix remains the foundation of purchasing
strategy for many organizations across sectors.
In the course of time the Kraljic approach has entered a lot of textbooks on purchasing
and supply management. Especially in Western Europe Kraljic has received large-scale
recognition and has gained an increasing degree of adoption. Gradually Kraljic has
gained acceptance in other countries, notably in the USA, Canada and Northern Europe.
3.3 Elaboration of purchasing strategies for all categories

Kraljic uses the classification (in strategic, bottleneck, leverage and non-critical items) for
the following purposes:
1. identifying main tasks, the required information and the decision level in
organizations per category;
2. identifying strategic items for which the next three phases (of market analysis, strategic
positioning and action plans) are completed.
Kraljic’s approach concentrates on strategic items without paying much attention to
strategic aspects of the other product categories. It is therefore remarkable that nowadays
the majority of practitioners and tutors are unfamiliar with Kraljic’s second matrix.
first matrix.
There are authors who have elaborated the ‘main tasks’ for bottleneck, non-critical and
leverage items into strategic recommendations: one overall purchasing strategy for each
cell/ category. These elaborations are usually limited to the construction of one matrix.
See for instance Van Weele (1992, 1994), Syson (1992), and Elliott-Shircore and Steele
(1985). It is recognized that these portfolio models are variations on and refinements of
the original first Kraljic matrix. We will address this issue later in this chapter.
In this section we present a typical elaboration of Kraljic’s first matrix, adapted from Van
Weele (1992, 1994). In line with ‘the main tasks’, Van Weele and other authors, identified
four general purchasing strategies for the corresponding categories:
- partnerships for strategic products
- assurance of supply for bottleneck products
76
- exploitation of power for leverage products (‘divide and rule’)

- systems contracting for non-critical products.
As a rule, strategic products together with the leverage products make up 80% of total
turnover. This means that minor changes in price levels or other developments in the
supplier market will have an immediate impact on the end product’s cost price.
Strategic products represent a considerable value to the organization in terms of a large

impact on profit and a high supply risk. Examples are engines and gearboxes for
automobile manufactures, turbines for the chemical industry and bottling equipment for
breweries. Sometimes strategic products can only be bought from one supplier (single
source). Since the organization experiences significant supply risk, it will strive for a
partnership relationship with suppliers of these products. Close and lasting cooperation
with suppliers must lead to improvements in product quality, delivery reliability, lead
times, product development, product design, and cost reduction. By intensifying the
relationship with suppliers the company tries to reduce its supply risk to a minimum,
while actively pursuing cost reduction at the same time.
first matrix with the well known categories: strategic, leverage, non-critical and
bottleneck. It might be assumed that all strategic products are to be managed by means of
(strategic) partnerships. Looking at the second matrix, we must conclude that this would
be in variance with Kraljic’s intention. In later work Van Weele (1997) included the issue
of power and dominance in buyer-supplier relationships, distinguishing three
possibilities in the strategic quadrant: a buyer dominated segment, a supplier dominated
segment and a balanced relationship.
Bottleneck products have less influence on the financial results, but they are vulnerable in
regard to their supply. Suppliers have a dominant power position. The purchasing
strategy is therefore primarily focussed on assurance of supply, if necessary even at
additional cost. Keeping extra stocks of the materials concerned or developing consigned
stock agreements with suppliers are examples of this strategy. To determine the most
important bottlenecks and their consequences a risk analysis should be employed.
Contingency planning might be a possibility for dealing with unexpected bad situations.
In general measures should be taken that will lead to a more balanced relationship with
the suppliers of these products. In terms of the Kraljic matrix this means moving from the
unfavourable ‘bottleneck product’-quadrant toward positions with less supply risk
and/or less financial impact. The purchasing strategy would imply the search for
alternative products and services. Supplier development would be a suitable purchasing
strategy in pursuit of a less dependent buying position.
In general leverage products can be obtained from various suppliers. These products
represent a relatively large share of the end product’s cost price in combination with a
relatively low supply risk. As a consequence, purchasing has interesting possibilities for
negotiating. Small percentages of cost savings involve large sums of money. At the same
time the supply risk is minimal. These characteristics justify an aggressive approach to
the supply market. A purchasing strategy based on the principles of competitive bidding
can be pursued. Since suppliers and products are interchangeable, there is no need for
long-term supply contracts. In general, a coordinated purchasing approach in the form of
a centrally negotiated umbrella agreement with preferred suppliers is appropriate here.
Using departments in the organization can place call-off orders as an administrative
formality. This strategy is also suited for more costly non-production areas. The buying
power is actively used as a means for getting better deals with interchangeable suppliers.
Non-critical products usually have a small value per unit and there are many alternative
suppliers. From a purchasing point of view, these items produce only few technical or
commercial problems. As a rule routine products require 80% of the purchasing
department’s time, while they often represent less than 20% of the purchasing turnover.
The administrative work should be organized in a very efficient way. The handling of
these products requires a purchasing strategy aimed at reducing the logistic and
administrative complexity. Systems contracting is generally advised as the way of doing
business with suppliers of routine products. Of course, on a short-term basis, there is no
need for long-term supply contracts. The main idea is to enhance purchasing power by
standardization and bundling of purchasing requirements.
Relevant for the purchasing strategy in case of routine products are:
- standardizing the product assortment to reduce product variety;
- systems contracts for groups of suppliers (for instance MRO items), instead of buying a
few products from one supplier;
- (as a consequence) reducing the number of suppliers;
- delegating the ordering function to user departments;
- reducing the number of invoices.
In addition to the systems-supported strategies, some companies have introduced the
purchasing card. For small single transactions purchasing cards are very cost-effective,
for they would relatively cost a fortune in paperwork and time when dealt with by the
purchasing department. Other possibilities for non-production are consortium buying
and the use of trading houses.
Syson (1992) discussed a very comparable elaboration of strategies, based on the original
Kraljic model. More interesting is the operationalization suggested with respect to the
measurement of the basic dimensions. The number of suppliers who are able and willing to
supply a specific item, is interpreted as a reflection of the supplier risk and the relative
78
strength of suppliers. The value of a purchased item has a relative significance. It is

relative to: (1) the total value of purchases (the percentage of buyers total purchase) and,
(2) the degree of importance attached to any particular sum by the seller (the percentage
of suppliers total sale). These key indices constitute a measure for the company’s strength
in any particular buyer/seller relationship.
Very similar to the first Kraljic matrix is the supply strategy square, developed by Hadeler
and Evans (1994). The basic dimensions for positioning the product categories are the
product’s complexity and the product’s value potential. Products that are not very complex
and have little value potential should be bought with simple contracts. The key is to find
an efficient and dependable supplier and negotiate a blanket order agreement. For
expensive non-technical products it is worth the effort of searching the globe for best
deals. By means of global trading companies can use their opportunities for corporate
leveraging. On the other end of the spectrum, complex technical items that do not make
up a big part of one’s spending, would require an entirely different approach. Here, close
relationships are more important than a worldwide search for best buys. It is critical to get
suppliers who are innovative and willing to work with the company, to identify new
applications, better solutions and substitute products. Strategic products require strategic
partnerships. The company needs close long-term working relationships with these
suppliers.
3.4 Supply positioning analysis

The portfolio approach of Elliott-Shircore and Steele (1985) starts with the listing of all the
goods and services which are purchased by an organization. These items are plotted in a
two-dimensional chart in which the X axis represents the profit/value potential of the
item (or category) and the Y axis represents supply exposure/vulnerability. The result is a
procurement positioning overview, as shown in figure 3.4. Recognizing the appropriate
category enables the determination of the resource and effort that should be allotted and
the basic approaches and methods that should be adopted. The procurement positioning
overview is now better known as the supply positioning analysis, a concept introduced by
Steele and Court (1996: 55).
FIGURE 3.4 Procurement positioning overview

Source: Elliott-Shircore and Steele (1985: 23)
The profit/value potential variable is an indicator for the relative importance of an item in
financial terms. Steele and Court (1996) proposed 0.5 percent as a threshold value: a
category which accounts for more than 0.5 percent of total expenditure would be located
in the middle of the X axis and would move further to the right as this proportion
increases. The operationalization and definition of the Y axis is less objective. Steele and
Court (1996) assumed that the variable will be a mixture of the following:
1. Supply availability
2. Quality requirements
3. Safety/environmental reliability.
The position of an item will move further upwards (‘more risk’) in case of (1) shortage or
limitations on the number of suppliers, (2) high quality requirements of tight tolerances,
and (3) overriding and important safety/environmental reliability considerations. Steele
and Court (1996) considered the lack of objective measurement as a beneficial
80
characteristic. The positioning of items should be the result of a discussion by individual

managers who need to come to an agreement with respect to all the purchased items.
The next step is to divide the chart into four quadrants as shown in figure 3.4. Each of the
four quadrants has a name which describes the nature of the category:
- tactical acquisition
- tactical profit
- strategic security
- strategic critical.
More interesting than the names of the categories are the purchasing goals and strategies
which are prescribed for each category.
1. The category tactical acquisition will account for a large proportion of the low value
items. The main purchasing objective here is to minimize attention by purchasing
professionals. They should spend as less time as possible on those low value, low risk
items. It is recommended to automate and/or delegate the purchasing processes as far as
possible.
It is preferable to have a combination of centralized purchasing and decentralized
purchasing: the purchasing professionals deal with the first phases of the buying process
(specification, supplier selection, negotiation and contracting) and the internal users deal
with the last phase, mainly the ordering part. Well known possibilities are:
- systems contracting
- blanket orders
- purchasing cards/credit cards.
By standardization and grouping of a large number of small items into a larger package it
is possible to contract just a few suppliers. As a result total cost will decrease, due to the
enhanced buying power, and the purchasing staff has less administrative workload.
The mental set for the purchasing professional would be ‘organize and let it go’.
2. The tactical profit items are of relatively high cost, but still have no major
complications. Professional purchasers are in a position to look for opportunities to cut
cost: drive profit. With an easy supply market they can afford to take risks, using
competition between potential suppliers.
The purchasing strategy is aimed at lowering cost by changing suppliers when necessary.
The needs of the company are mainly met by short term contracts in flexible supply
markets. The purchasing professional plays an active (sourcing) role, based on detailed
knowledge of markets and suppliers.
The mental set for the purchasing professional is that of ‘trade’.
3. For the strategic critical category the situation is quite different. The items are of high
cost in combination with high risks. As a consequence of the first characteristic cost
reduction can contribute heavily to the financial results of the organization. But on the
other hand it will be essential to ensure availability of supply. Professional purchasers

need to spend much time and energy in managing suppliers. This category requires the
highest level of buying skills.
Possible strategic actions would be: finding new suppliers (supplier development),
overcoming some of the special considerations which prevent the use of competition, and
ensuring that existing suppliers are capable of meeting future demands.
Buyers will need to obtain detailed information on individual suppliers and be seeking to
develop them to mutual advantage. In the strategic critical category we would expect to
find strategic partnerships, such as co-makership.
The mental set for the purchasing professional would be ‘checking and searching’.
4. The strategic security items are also critical to the success of the organization.
Therefore, the major goal would be: ensure supply. Cost considerations come in the second
place. We would expect to see a decrease in the sensitivity to price. Still, purchasing
professionals should be aware of increasing cost, i.e. moving to the right in the
positioning analysis and becoming ‘strategical critical’. The assurance of supply can be
provided by long-term contracts, for instance with the use of indices and formulae to fix
prices. Another way of ensuring supply would be to hold buffer stocks or to agree that
suppliers will do so.
The mental set for the purchasing professional is ‘reduce the problem’.
Discussion
It is quite obvious that there is a striking resemblance with the Kraljic approach and the
elaborations such as those by Van Weele (1994) and Syson (1992). Elliott-Shircore and
Steele (1985) used (practically) the same dimensions and provided (practically) identical
recommendations for the same four categories. A ‘difference’ would be that in the supply
positioning analysis the four categories have other names. In conclusion, there are no
principal differences with Kraljic’s approach.
There are some publications that refer to the supply positioning analysis, using labels
such as ‘tactical acquisition’ and ‘strategic critical’. The literature search resulted in just a
few documented examples of the application of portfolio techniques in purchasing and
supply management. Carter (1997) described the practical application of the supply
positioning analysis at SGX Corporation, a conglomerate with five major strategic
business units. SGX wanted to make better use of the likely advantages as a result of
combining purchasing and supply possibilities across SBU’s. For all products, the
positions in the matrix and the strategic actions were identified, emphasizing the special
possibilities of joint SBU-buying. Examples are: corporate blanket purchase orders with
large national distributors (tactical acquisition), joint safety stocks and assisting each
other with urgent requirements (strategic security), world wide commodity analysis and
competitive bidding (tactical profit), and to manage certain key suppliers on a group
82
basis (strategic critical). Carter (1997) concluded that this portfolio approach proved to be
a powerful tool for coordinating the sourcing patterns of the various strategic business
units. Another application of the supply positioning analysis is provided by Croom
(2000), illustrating how MRO items could benefit from the implementation of e-
procurement. As a direct consequence of the advantages of e-procurement, MRT items
may switch from ‘tactical acquisition’ to ‘tactical profit’ or even ‘strategic critical’.
Improved information transparency makes it possible to adopt a more ‘strategic’
procurement approach.
3.5 Cunningham’s interaction approach to portfolio analysis

In the context of the IMP-research Cunningham (1982) presented an ‘interaction approach
to purchasing strategy’. For the formulation of purchasing strategy Cunningham
advocated three basic phases: an analysis of the purchasing environment, the setting of
purchasing objectives and the choice of alternative purchasing strategies for their
achievement. An important part of the environmental analysis is the analysis of supplier
relationships which consists of a portfolio analysis, a supplier performance analysis, and a
purchasing power analysis.
Cunningham (1982) advocated a portfolio analysis of the suppliers currently used, and
also desired by a company. The main idea is that the supplier base should be a balanced
portfolio. Various criteria can be used to determine (un)balance in the portfolio of
supplier relationships. Referring to the concept of a balanced portfolio of financial
investments, Cunningham advanced the thesis that it is beneficial to have relationships
with different types of suppliers. Two examples are given to support his view. (1) The
spread of suppliers across different countries reduces the risks of movements in currency
exchange rates and inflation. (2) Having suppliers at different stages of development, or
of varying sizes, is beneficial because each category of suppliers requires a different level
of resource commitment for effective interaction. Each of them may bring different types
of benefits, for instance one supplier serves as a source of technical innovation, one for
non-standard items, and another for low cost supply.
According to Cunningham (1982) the achievement of a balanced portfolio of supplier
relationships is an important purchasing objective, in addition to the fundamental
purchasing objectives with respect to the assurance of (long term) supply in an efficient
way. Suppliers are or should be selected on the basis of the benefits which can be
obtained by the company. Therefore the portfolio analysis of supplier relationships must
be related to the key purchasing objectives and the key strategies to achieve them.
Cunningham identifies four key purchasing objectives:
1. Security of supply - to stabilize supply inputs and prices over time or to avoid over-
dependence on a single source of supply.
2. Matching with appropriate supplies - to ensure that suppliers have desired
characteristics for long-term relationships and will be able to meet specifications and
price requirements.
3. Controlling relationships - to achieve satisfactory results, such as developing increased
power over the supplier, closer collaboration, etc.
4. Cost saving and stimulating competition - to achieve greater cost efficiency in
purchasing.
A wide range of actions can be identified to achieve these broad strategic objectives.
Conclusion
The most remarkable point in Cunningham’s interaction approach to portfolio analysis is
that it is very deviating from the mainstream portfolio approaches. No clear differences
are made between supplier categories and the approach does not include a classification
of items or suppliers. As a consequence no strategic recommendations can be given with
respect to the issue of differentiated supplier strategies. Cunningham just made the point
that different suppliers can provide for different benefits and that a balanced portfolio of
suppliers is important. The conclusion is that Cunningham’s interaction approach does
not meet the basic requirements of a purchasing portfolio approach.
3.6 The Olsen and Ellram approach

Olsen and Ellram (1997) suggested a portfolio approach for managing supplier
relationships. Based on a literature review and the critique of portfolio models in general,
a three-step approach is recommended.
3.6.1 Description of the model
The first step in the portfolio analysis is to categorize the company’s purchases according
to the strategic importance of the purchase and the difficulty of managing the purchase
situation. A list of nine factors influencing the strategic importance of a purchase is
provided. A distinction is made in competence, economic and image factors. In total, six
illustrative factors are listed, influencing the difficulty of managing the purchase
situation. A distinction is made between product characteristics, supply market
characteristics and environmental characteristics. Due to the relative large number of
possible factors, the weighting of factors on their perceived importance is a vital, but very
complex part of the portfolio analysis.
The dimensions make up a portfolio model with four categories that have the same
names as the categories in the (first) Kraljic matrix: strategic, bottleneck, leverage and
non-critical items. Normative guidelines are provided on how to manage the
relationships associated with each of the four categories.
84
For strategic items it is recommended to ‘establish a close relationship’, for instance by

means of early supplier involvement and joint product development. Systems contracting
is advised to leverage volume across product lines and suppliers and to lower material
costs.
Non-critical items should be standardized and bought by a small, and if necessary
reduced, number of suppliers (standardization and consolidation). Finally, for the
bottleneck items the company should try to standardize the purchases or to find
substitutes if possible. These guidelines or strategies are viewed as idealizations. In the
second step the actual relationships with suppliers are to be analyzed by means of
another portfolio model.
Olsen and Ellram (1997) proposed to categorize supplier relationships, based on the
supplier attractiveness and the relationship strength. There are many factors influencing the
relative supplier attractiveness. A total of 24 factors are described, divided over five
general groups: economic factors, performance factors, technological factors,
organizational, cultural, and strategic factors, and other factors. The strength of the
relationship is described by factors that create bonds between two companies. Again, an
extensive list of 15 factors is provided, divided over four groups: economic factors,
character of the exchange relationship, cooperation between buyer and supplier and the
distance between the buyer and seller.
The final result of the second step is the construction of a portfolio matrix with respect to
current supplier relationships, as shown in figure 3.5 A three-by-three matrix
distinguishes nine different types of supplier relationships.
FIGURE 3.5 Analysis of supplier relationships

Source: Olsen and Ellram (1997: 107)
In the third step action plans are developed for moving from current to ideal supplier
relationships. The last step is meant to be the link between the previous steps, by
comparing the results of the corresponding analyses. The cells in figure 3.5 are described
in three groups. Based on the categorization of the associated purchase(s) in the first
matrix, examples of possible strategies for relationships are provided.
Cell 1, cell 2, and cell 4 include relationships with a high or moderate supplier
attractiveness and a low or average relationship strength. A number of different
strategies are recommended, based on the classifications of purchases in the first matrix.
In case of strategic items it is very important to strengthen the relationship and to keep a
loyal supplier. A relationship can be strengthened by enhancing the communication,
providing more volume, or involving the supplier in product development. If the item is
a non-critical or leverage purchase, the company could consider to strengthen the
relationship without allocating considerable resources to the relationship. For instance,
by giving the supplier more volume. Action plans for these purposes require long-term
resource allocation, because it takes time to build relationships.
Cell 3, cell 5, and cell 6 contain relationships where the supplier has a moderate or high
relative attractiveness and the relationship is relatively strong. For all types of items, a
strategy is recommended that would include the reallocation of resources in order to
86
maintain a strong relationship. However, in case of non-critical or leverage items, the

company should consider enhancing the supplier attractiveness by systems contracting or
reducing the resources spent to manage the relationship. The payoff from a strong
relationship is relatively low in this situation.
Cell 7, cell 8, and cell 9 include the relationships with a low supplier attractiveness. There
relationships warrant attention because a reasonable strategy would be to change the
supplier. However, if the purchase is strategically important or it is a bottleneck
purchase, it is crucial that the company develops an action plan on how to secure the
delivery or create substitutes. It may be more efficient to develop the current supplier
rather than to establish a relationship with a new supplier. The purpose of the action plan
would be to improve the supplier attractiveness or the performance of the relationship.
To summarize, three different relationship-strategies are suggested:
(1) strengthen the relationship,
(2) maintain the relationship,
(3) develop or replace the relationship.
For each relationship strategy action plans are developed, indicating the resources that
need to be allocated to the various supplier relationships.
3.6.2 Discussion of Olsen and Ellram’s model
(1) Dimensions
The Olsen and Ellram-approach is at variance with the more common practice of
purchasing portfolio approaches. It is a three-step approach including the development
of two portfolio matrices. The first matrix is a copy of Kraljic’s first matrix, although the
dimensions are labeled differently. The four dimensions of the second matrix in their turn
are exactly the same as the dimensions that Fiocca (1982) introduced for his (second)
matrix from a marketing perspective. It is remarkable that the four Olsen and Ellram-
dimensions completely match Fiocca’s four dimensions: the strategic importance of
customers, the difficulty of managing customers, the customer’s business attractiveness
and the strength of the buyer/seller relationship. The conclusion is that the Olsen and
Ellram-matrices are based on Fiocca (1982). There are no other theoretical foundations
mentioned. Nellore and Söderquist (2000: 247) assumed that the rationale behind the
selection of dimensions would be “experience-based”, because they would reflect the
way in which purchasing and engineer staff actually think about purchasing situations.
This conclusion might be questioned, as we shall elaborate under (4) use issues.
In general, the same measurement issues apply for the Olsen and Ellram-approach as for
the Kraljic approach, albeit that Olsen and Ellram put a lot of work into identifying all
kinds of possible factors that may be of interest for the management of supplier
relationships. Their
extensive literature search produced more than fifty factors, influencing one way or
another the different dimensions. As such, this would create additional measurement and
implementation problems.
(2) Categories
The first portfolio model in the Olsen and Ellram-approach uses exactly the same
category labels as the first Kraljic matrix: leverage, strategic, non-critical, and bottleneck.
It is not clear why the same labels should apply in case of different dimensions. The
categories in the second portfolio model are not named at all. Nine positions in the
second matrix are identified, based on supplier attractiveness and relationship strength.
The merit of this three-step approach is that it provides a more detailed analysis,
compared to more classical portfolio models. Their matrices include the positioning of
items (products) and the positioning of suppliers (relationships) as well.
(3) Strategic recommendations

It is clear that Kraljic’s recommendations are very similar to the Olsen and Ellram
recommendations in the first matrix. However, for the strategic items (in the second
matrix) Kraljic chose a power perspective: the company strength in comparison with the
supply market strength. Olsen and Ellram strongly prefer a relationship perspective,
combining relationship attractiveness and relationship strength. On the long run any
exploiting strategy is expected to have dangerous and averse effects. In their view, power
and the risk of opportunistic behavior are only two factors influencing the appropriate
supplier strategy.
The ‘action plans’ in the third step are developed for moving from the current to the ideal
supplier relationships. In contrast with the Kraljic approach, these recommendations
explicitly allow for improving prevailing conditions. It is a strong point that strategies
and action plans invite the purchasing professional to look for better positions in the
matrix, bridging the gap between ideal and actual positions (dynamics). A key element in
their approach is the allocation of scarce resources to the various relationships, which
addresses the issue of interdependencies in the matrix. A weak point however, is that the
strategies and recommendations are put in rather general terms. The strategic
recommendations suggest that relationships should be strengthened, maintained,
developed, or replaced, using more, fewer, or the same resources.
(4) Use issues

Olsen and Ellram (1997) emphasized that the weighting of each factor is the most
important part of the implementation process, but at the same time very subjective. The
decision-makers must come to an agreement on the relative importance of each factor. De
Boer (1997) suggested a fully customized approach: organizations should determine their
own criteria as well as their own specific threshold values. It’s not clear how this
determination process should or could take place. The complexity of the dimensions used in
88
the portfolio models ‘could be a problem’ in the implementation phase, as concluded by

the authors themselves. This would be an understatement, considering the vast number of
factors and the unclear methodology of weighting and scoring factors. This is perhaps why
Olsen and Ellram concluded that it is important that decision-makers in the company
discuss all the important factors and decide for themselves which factors should be used
in their specific case. They could decide that not every factor has to be included. As a
result, factors will vary from company to company. It is suggested that the process of
categorizing items and the process of achieving consensus might be more important than
the classification matrices. Møller and Momme (2000) found in a case study at Bang &
Olufsen that the Olsen and Ellram-approach was rejected because it gave room for
arguments and disagreement on factors and variables and it could potentially fuel an
internal political debate between parties with interest in supplier relationships. The
model did not provide sufficient guidance for practical use.
As all the other purchasing portfolio approaches, the Olsen and Ellram-approach lacks
substantial empirical support. Olsen and Ellram (1997) suggest that future research
should include an extensive empirical testing of the usefulness of the portfolio approach
and the normative strategic suggestions.
3.7 Bensaou’s specific investments approach

As part of a broader research project on supplier relationships Bensaou (1999)
administered a survey questionnaire to Japanese and U.S. automobile manufacturers. The
purpose of the study was to answer two questions: (1) which relational design should a
firm choose under different external conditions?, and (2) what is the appropriate way to
manage each different type of relationship?
3.7.1 Description of the model
In the design of a portfolio model, it is common practice to start with the determination of
the dimensions, based on theoretical and practical insights. Bensaou however, did not
operate in this manner. Searching for patterns within the data, it was found that the level
of specific investments made by either partner significantly correlated with practices
commonly associated with strategic partnerships, such as long-term relationships, mutual
trust, and cooperation. Specific investments are investments that are difficult or
expensive to transfer to another relationship. Buyer’s specific investments include
tangible investments in buildings, tooling and equipment dedicated to the supplier or in
products and processes customized to the components procured from the supplier.
Intangible investments refer to investments in people or in time and effort spent learning
the supplier’s business practices and routines. Supplier’s specific investments include
plant or warehouse location and
specialized facilities and dies. Intangible investments include sending guest engineers
and developing information systems compatible with the buyer’s proprietary database or
electronic data interchange protocols.
Bensaou (1999) classified buyer-supplier relationships into four generic categories, using
buyer’s and supplier’s specific investments as dimensions:
- strategic partnership (high mutual specific investments),
- market exchange (low mutual specific investments),
- captive buyer (high buyer’s specific investments, low supplier’s specific investments),
and
- captive supplier (low buyer’s specific investments, high supplier’s specific investments).
On the basis of these classification, no strategic recommendations were formulated.
Bensaou (1999) followed another, an empirical approach to this issue. In search for
strategic recommendations three dimensions of ‘relationship performance’ were
considered (Bensaou and Venkatraman, 1995):
(1) supplier rating index assessed by the manufacturer’s team of engineers of ten
dimensions, e.g., development time, delivery performance, quality performance, price
competitiveness, contribution to lowering cost, engineering capabilities,
(2) perceived satisfaction with the relationship along seven criteria, e.g., the quality,
amount and accuracy of the information exchanged, and
(3) level of buffers between the two firms, i.e., average level of inventory kept by the
assembler, by the supplier, shipment increments and average quality levels.
Then each cell was split into high-performing and low-performing relationships.
Variables were identified that displayed a difference between the two performance
subgroups. Three types of variables were found:
- information-sharing practices
- characteristics of the boundary spanner’s job
- the social climate and process characteristics.
The management profiles describe per category the richness of information exchanged,
the amount of time spent with supplier’s staff, the frequency and character of (mutual)
visits, the relationship climate, the reputation of the supplier, and the involvement in
design and cooperation. Figure 3.6 shows the advised ‘management profiles’ for each
type of buyer-supplier relationship.
90
Captive buyer Strategic partnership
Information-sharing mechanisms Information-sharing mechanisms

- ‘Broadband’ and important exchange of detailed - ‘Broadband’, frequent and ‘rich media’ exchange
information on a continuous basis - Regular mutual visits and practice of guest engineers
- Frequent and regular mutual visits
Boundary spanners’ task characteristics
Boundary spanners’ task characteristics - Highly ill defined, ill structured
- Structured task, highly predictable - Nonroutine, frequent unexpected events
- Large amount of time spent by buyers’ purchasing - Large amount of time spent with suppliers’ staff,
agents and engineers with supplier mostly on coordinating issues
Climate and process characteristics Climate and process characteristics

- Tense climate, lack of mutual trust - High mutual trust and commitment to relationship
- No early supplier involvement in design - Strong sense of buyer fairness
- Strong effort by buyer toward cooperation - Early supplier involvement in design
- Supplier does not necessarily have a good - Extensive joint action and cooperation
reputation - Supplier has excellent reputation
Market exchange Captive Supplier
Information-sharing mechanisms Information-sharing mechanisms

- ‘Narrow-band’ and limited information exchange, - Little exchange of information
heavy at time of contract negotiation - Few mutual visits, mostly from supplier to buyer
- Operational coordination and monitoring along
structured routines Boundary spanners’ task characteristics
- Limited time allocated by buyer’s staff to the supplier
Boundary spanners’ task characteristics - Mostly complex, coordinating tasks
- Limited time spent directly with supplier staff
- Highly routine and structured task with little Climate and process characteristics
interdependence with supplier’s staff - High mutual trust, but limited direct joint action and
cooperation
Climate and process characteristics - Greater burden put on the supplier
- Positive social climate
- No systematic joint effort and cooperation
- No early supplier involvement in design
- Supplier fairly treated by the buyer
- Supplier has a good reputation and track record
FIGURE 3.6 Management profiles for buyer-supplier relationships

Source: Bensaou (1999: 39)
Bensaou (1999) searched for any performance differences among the four types of buyer-
supplier relationship. Remarkably, no one type of relationship seemed to be superior to
the others. In other words, each type of relationship can be well or poorly managed. This
in support of the hypothesis that successful supply chain management requires the
effective and efficient management of a portfolio of relationships. Bensaou (1999)

concluded that good practice means, first, properly balancing a portfolio of relationships
adapted to product and market conditions, second, managing each type of relationship
effectively.
3.7.2 Discussion of Bensaou’s model
(1) Dimensions
A very strong point of Bensaou’s model is that it is based on empirical research, albeit in
just one industry (automotive). The two dimensions - buyer’s and supplier’s specific
investments - were uncovered by an exploratory analysis of the data. The rationale for
using these criteria was their identification as discriminating factors for clustering the
categories of supplier relationships in a sample of almost 450 purchasing situations. This
clearly is a solid rationale in terms of empirical validity (Nellore and Söderquist, 2000). In
addition, there is a theoretical foundation as well. The concept of ‘specific investments’ is
derived from the transaction cost theory (Williamson, 1975). Transaction-specific
investments have an important role in the transactions cost theory in determinating the
level of transaction costs. The basic assumption of the transactions cost theory is that the
choice of an optimal coordinating mechanism will depend on the level of transaction
costs relative to internal coordination costs (Williamson, 1985). It is posited that
transaction-specific investments determine largely the choice of the optimal coordinating
mechanism: markets or hierarchies. When asset specific investments are made, they must
be safeguarded against opportunism (Heide and John, 1988). Under conditions of high
asset specificity, the transaction cost theory predicts and prescribes moving away from
market transactions (markets) toward vertically integrated relationships (hierarchies).
(2) Categories
The four types of buyer-supplier relationships are clearly connected with relative power
positions. The ‘captive’ buyer depends heavily on the supplier and is obviously less
powerful. The reverse holds true for the ‘captive’ supplier. The ‘strategic partnership’
and the ‘market exchange’ categories reflect a balanced situation. Power implicitly plays
an important role in Bensaou’s model. However, the dimensions do not correspond
unambiguously with the balance of power. Indeed, the buyer’s and the supplier’s specific
investments are important aspects, but there are other, neglected determinants to any
relative power position. The balance of power is by definition not just determined by the
ratio of buyer’s specific investments and supplier’s specific investments.
If we take a close look at these four categories, it seems that Bensaou’s classification is in
accordance with Kraljic’s first matrix.
92
- strategic partnerships (balanced power, high mutual specific investments) refers to

‘strategic’
- market exchange (balanced power, low mutual specific investments) refers to ‘non-
critical’
- captive buyer (supplier dominated) refers to ‘bottleneck’
- captive supplier (buyer dominated) refers to ‘leverage’.
(3) Strategic recommendations

Bensaou’s portfolio model is richer in content than the other portfolio models because it
includes relationships performance and management profiles for the various supplier
relationships. High performing and low performing relationships were considered for the
discovery of appropriate management profiles for each relationship type in the portfolio
(‘contextual profile’). Again, empirical findings were used as a foundation for the content
of the relationship portfolio matrix.
Less satisfactory is the nature of the managerial recommendations in Bensaou’s
approach. Bensaou’s approach results in a description of the current state between buyer
and supplier. A prominent conclusion is that no one type of relationship is superior to the
others: each type can be well or poorly managed. The implication is that the buying firm
should adapt to prevailing circumstances, of course in the best possible manner. For
instance, if a relationship is supplier-dominated, then the management profile concerned
does not provide guidelines for achieving a better position. The managerial
recommendations do not allow for any strategic switch from one type of relationship to
another. In other words, the model is reactive by nature. Purchasing management is
implicitly supposed to reconcile itself to prevailing conditions.
Bensaou introduced and used measures of ‘relationship performance’ in terms of an

average of overall supplier ratings and perceived satisfaction mainly with the
information exchanged. There are questions with respect to the measurement of
‘relationship performance’. How should we evaluate the construct validity of the
‘relationship performance’ concept? The construct validity refers to the correctness of the
operational measures for the concept being studied (Yin, 1994). Does the construct
measure what it is supposed to measure? In addition, there is a question of internal
validity (Yin, 1994): is a causal relationship established? Do certain conditions with
respect to the ‘managerial profiles’ actually lead to different ‘relationship performances’?
Or is it merely a matter of statistical correlation? The recommended management profiles
seem to reflect behavior that is statistically connected with ‘high performing’
relationships. No causal model is provided to explain and improve ‘relationship
performance’. Apparently, an explorative and descriptive study is stretched to causal and
normative propositions.
(4) Use issues

Bensaou’s study categorized and listed a large number of relationships in the automotive
industry. The model is based on empirical research. However, it is not actually used as a
portfolio model for purchasing and supply management. The identified categories refer
to four basic types of buyer/supplier relationships. The strategic recommendations
however, would have to be adjusted into a more pro-active, dynamic use. As far as we
know, there are no publications that report on applications of Bensaou’s model. The
possibilities for practical applications however are promising.
3.8 Conclusions of the literature study: addressing the

research questions
In this last section the results of the literature study will be connected with the research
questions of this study. Based on the conclusions of the literature study, subsection 3.8.1
will answer the first research question that refers to differences and similarities between
the various purchasing portfolio models. In the subsections 3.8.2 and 3.8.3 the second and
third research questions will be addressed. The analysis of the literature will provide the
basis for the elaboration of the research questions into a number of more articulated sub-
questions.
3.8.1 Answers to the first research question: differences and similarities
The first research question of this study refers to the comparison of purchasing portfolio
models:
“What are differences and similarities between the various purchasing portfolio models?” (1)
We agree with Nellore and Sörderquist (2000) who contended that all portfolio
approaches basically use the same three steps: the analysis of products and their
classification in a matrix, the analysis of required supplier relationships to deliver the
products (objectives), and the development of action plans in order to bridge the gap
between current and required supplier relationships (strategies). Our literature suggests
that no substantial differences are to be found in the general approach of the various
portfolio models in purchasing. In this section we will substantiate this conclusion which
constitutes the answer to the first research question.
In chapter 1 we have indicated that the comparison will concentrate on three core
elements of portfolio models: (1) dimensions, (2), categories, and (3) strategic
recommendations. In addition, a comparison has been made on (4) use issues (acceptance
and adoption). Beginning with acceptance and adoption, it is clear that Kraljic (1977,
1983) introduced the first comprehensive portfolio approach for purchasing and supply
management. Many other authors have used Kraljic’s basic ideas for the development of
94
similar models. For instance Elliott-Shircore and Steel (1985) merely changed the labels in
their supply positioning analysis. Van Weele (1992, 1997) and Syson (1992) elaborated
Kraljic’s work, especially with respect to the strategic recommendations for the leverage,
the non-critical and the bottleneck category. These contributions are a useful supplement
to the 1983-article, because the original Kraljic approach is very much focussed on the
strategic category. By now, probably a minority of practitioners is familiar with the
second matrix: company strength versus supply market strength for the positioning of
strategic items. It is fair to conclude that in common parlance the ‘Kraljic approach’ is
fully equated with the original first Kraljic matrix, which was not presented as a ‘matrix’
anyway, but as a table (‘exhibit II’): ‘classifying purchasing materials requirements’. The
categories were labeled as the ‘procurement focus’, for which the main tasks, the required
information and the decision level were specified (Kraljic, 1983: 112).
Kraljic (1983) provided the first comprehensive portfolio approach for purchasing and
supply management. Many years after the introduction there is a reasonable amount of
evidence that Kraljic’s basic ideas and concepts represent the dominant approach in the
profession. On a large scale purchasing professionals and academics have adopted
Kraljic’s basic ideas and methodology. The (first) Kraljic-matrix has become the standard
in the field of purchasing portfolio models. Kraljic’s terminology is generally accepted
and has become the standard for scientists and practitioners. This however does not
apply for the other portfolio models, which are hardly known and used in practice.
Therefore, Kraljic’s approach will be used as point of departure in the following
discussion and evaluation of other approaches. The Kraljic-model will serve as a natural
reference point.
Kraljic Elliott-Shircore Van Weele Olsen and Bensaou

(1983) and Steel (1992, 1994) Ellram (1997) (1999)
(1985)
Dimensions of the first or only matrix
1. Profit impact Profit/value Profit impact Strategic Supplier's specific
potential importance investments
2. Supply risk Supply Supply risk Difficulty of Buyer's specific
vulnerability managing the investments
purchase
situation
Dimensions of the second matrix

1. Company n/a n/a Supplier n/a
strength attractiveness
2. Supply market Strength of
strength relationship
Categories/cells
1. strategic strategic strategic strategic strategic
critical partnership
2. leverage tactical profit leverage leverage captive supplier
3. bottleneck strategic bottleneck bottleneck captive buyer
security
4. non-critical tactical non-critical non-critical market exchange
acquisition
Recommendations for:
- strategic items exploit, diversify manage partnership close management
or balance * suppliers relationships profiles,
- leverage items exploit power drive profit exploit power leverage volume in terms of
- bottleneck items volume ensure supply assurance of standardize and information
insurance supply find substitutes sharing, tasks
- non-critical efficient minimize systems standardize and and climate
items processing attention contracting consolidate for each category
* depending on the balance of power
FIGURE 3.7 Overview and comparison of purchasing portfolio approaches
Various authors have discussed and introduced similar models (see figure 3.7 for an
overview and comparison of the portfolio models we have discussed in this chapter). The
representation is focussed on dimensions, categories, and strategic recommendations.
The comparison of the different portfolio models suggests that there are more similarities
than differences. This is especially true for the Kraljic (1983), Elliott-Shircore and Steel
(1985), Van Weele (1992, 1994) and Olsen and Ellram (1997). They use practically the
same dimensions (for the first or only matrix), the same categories and the same
recommendations. In other words, there are hardly any differences with the original
Kraljic matrix. Purchasing portfolio models are all aimed at identifying ‘(un)important’
items and at developing different strategies for different categories, using practically the
96
same dimensions, categories and strategic recommendations. The figure does not show
the additional strategic recommendations that Olsen and Ellram (1997) have provided for
their second and final matrix, although they might be simply classified as ‘manage the
intensity of the supplier relationship by allocating scarce resources’. The Olsen and
Ellram-approach could be seen as a relative outsider. Nevertheless it is to a large extent
based on Kraljic’s work. Being more complex and less transparent for purchasing
managers, the Olsen and Ellram-approach will not be widely applied. Bensaou (1999) is
the real outsider, with more differences than similarities in comparison with the Kraljic
approach, see figure 3.8. Bensaou’s recommendations are put in the form of advised
management profiles, matching different buyer-supplier relationships. The Kraljic
strategies are formulated in terms of general guidelines, referring to the possibilities at
hand. Because of differences in scope, the strategic recommendations of the approaches
are to a large extent incompatible and incomparable.
The role of power in Kraljic is quite explicit for the strategic items, judged from the
recommended purchasing strategies: exploit, diversify and balance. For the other
categories it is less clear how issues of power are handled. Here is a glaring contrast with
Bensaou’s approach that is quite explicit about the role of power and dependence,
distinguishing the categories in the matrix. Consider for instance the ‘captive buyer’ and
the ‘captive supplier’ cells. However, the balance of power is not explicitly accounted for
in the managerial recommendations, which merely adapt to prevailing conditions. In
addition, the dimensions do not correspond unambiguously to power in buyer-supplier
relationships. In conclusion, both approaches have their own strong points. However,
these issues need elaboration and study in more detail, which will be done in the next
chapter.
Kraljic (1983) Bensaou (1999)
General idea minimize supply risk and make match appropriate management
the most of buying power profiles to different relationships
Dimensions
- theoretical foundation unclear transaction cost theory
- empirical support professional experience uncovered by data analysis
- supplier's side not explicitly included explicitly included
Categories
- unit of analysis product categories buyer-supplier relationships
- role of power explicitly for strategic items, explicitly for all categories
unclear for other categories
Strategic recommendations
- type generic supply strategies management profiles
- dynamics unclear about switchting to static, reactive
other positions in the matrix
- changing conditions yes, however only explicitly for no, adaptive to external
the strategic items conditions
Adoption and use large scale single case study
FIGURE 3.8 Kraljic versus Bensaou
3.8.2 Elaboration of the second research question: explaining the use of a

portfolio approach
The literature study made clear that the Kraljic matrix is the dominant model in
purchasing portfolio approaches. The study yielded a number of references that report on
the actual use of a portfolio approaches in practice. Some are of an anecdotical nature
(with respect to the experience of a single company), others contain statements about use
percentages. In addition we have found literature that contributes to the understanding
of reasons for using a portfolio approach and literature that lists possible factors and
variables, explaining the use of a portfolio approach in purchasing. Based on the findings
of the analysis of literature, our second research question can be further elaborated:
“Which factors would explain the utilization of the purchasing portfolio analysis?”(2)
Before we explain the use of the portfolio tool, we should first establish the use
percentage and the use intensity. There are differing statements and estimates of the use
percentage. Most publications report a rather ‘high’ and increasing use. For instance,
Lamming and Harrison (2001: 596) stated that “Kraljic’s matrix remains the foundation of
purchasing strategy for many organizations across sectors”. But there are opposite views
too: Olsen and Ellram (1997: 110) stated that “the use of portfolio models (…) in
98
purchasing has been very limited” and Christopher and Jüttner (2000: 120) found “few
companies following the advice to apply portfolio techniques to classify relationships, but
instead classifying relationships in the form of a simple hierarchy (of preferred suppliers
– 2nd tier suppliers – 3rd tier suppliers – potential newcomers)”. Most publications do
not report concrete numbers on portfolio models, although there are a few exceptions. In
a survey of 126 Dutch companies Boodie (1997) found that 44% of the responding
purchasing managers said that they used Kraljic for formulating purchasing strategies. A
peak of 80% was found for industrial companies that operate on a mass production basis.
In contrast, Kamann (2000a) estimated that 20% of the Dutch companies uses a portfolio
method for the management of their supplier base. Besides ‘use percentage’, the ‘use
intensity’ is an interesting variable too. Reasons enough for the following research
question:
Research question (2a)

How many firms actually use a purchasing portfolio approach?
In addition, in more detail the study will investigate specific reasons for using the
portfolio in purchasing and the extent to which these reasons can be found in practice:
Research question (2b)

Why is the purchasing portfolio approach being used in practice?
Research question (2c)

Why are other professionals not using a portfolio approach in purchasing?
The use of the portfolio could be associated with other use-related factors, such as
satisfaction with portfolio use and the applied measurement method for the positioning
of items in a matrix.
Obviously, the decision to use the portfolio tool in practice will be motivated by the
expected benefits and results. In general, in the case of a management tool, the question
of effectiveness naturally arises. On the bottom-line managers want to know: ‘does it
work?’ and ‘what are the results?’ and perhaps ‘what are the cost reductions?’
Performance measurement in general is a difficult problem in management studies. Many
attempts have been made to capture performance with mixed results. This led
Venkatraman and Ramanujam (1986: 801) to suggest that the treatment of performance in
research settings is perhaps “one of the thorniest issues confronting the academic
community”. Over the past years many performance management systems and key
performance indicators have become popular. However, reporting purchasing cost
savings and supplier performance measures is in most cases not sufficient to get the buy-
in from the top (Van Weele, 2002). Traditional management accounting focuses on prices
and costs, which seems to fit in well with a transactional approach to purchasing
(Axelsson, et al., 2002). In a more relational approach to purchasing other measures have
to be used, capturing inter-organizational processes and other accountable contributions
of purchasing to the competitive position of the firm. This means that the selection of
performance measures for the purchasing function is a critical but not an unambiguous
task. A major problem in relating strategy to performance is that many factors intervene
the process (Venkatraman and Ramanujam, 1986), causing the independent variable
(strategy) to have a less significant or non-causal relation with the dependent variable
(performance). For intermediary activities it is therefore better to look for intermediary
measures of success or performance (Mol, 2002). This points at the issue of accountability.
Van Weele (1984) emphasized that purchasing performance measurement should only be
established in those areas in which a buyer can be held responsible. The same line of
reasoning applies for the performance measurement of the portfolio approach: only those
areas should be included for which the tool can be held responsible.
Discussing the effectiveness of portfolio analysis, we should return to the source. Kraljic
(1983) proposed a four-stage approach as a framework for ‘shaping the supply strategy’:
classification (first matrix), market analysis, strategic positioning (second matrix for
strategic items) and action plans. For this matter one should remember that this approach
is focussed on the items in the strategic category. Other authors have elaborated the
strategic recommendations for the other categories. Kraljic (1983: 112) posited that in
order “to minimize their supply vulnerability and make the most of their potential buying
power”, a number of companies have ‘successfully used the four-stage approach’.
However, the issues of how to determine these effects (measurability) and how to
attribute the outcomes of the portfolio analysis to these effects (accountability), are not
addressed and therefore not solved. Supply vulnerability and buying power are at best
overall performance measures for purchasing.
The literature study suggested that the portfolio analysis can be used for diagnostic and
prescriptive purposes, although the strategic recommendations are rather generic and non-
specific by nature. The diagnostic powers are related to the additional insights that are
gained, using the portfolio tool. The prognostic powers however are limited and the
portfolio should therefore not be used in a prescriptive way. Portfolio analysis enables
purchasing professionals to develop differentiated purchasing and supplier strategies.
Keeping in mind the necessary condition of accountability, it is concluded that the
effectiveness of a portfolio approach should be assessed in terms of direct accountable
impact, operationalized to the extent that users:
- experience additional understanding of current purchasing problems and possibilities
(1), and
- develop differentiated strategies for their purchasing and supplier management (2).
100
These effects can be considered as intermediary measures for overall purchasing

performance measures (cf. Mol, 2002). This leads to the formulation of the following
hypotheses, to be tested with the data from the survey:
Hypothesis A1 The use of the purchasing portfolio analysis is positively related to

the level of understanding of the problems and possibilities of the
purchasing function.
Hypothesis A2 The use of the purchasing portfolio analysis is positively related to

the extent to which differentiated purchasing and supplier strategies
are developed.
Ultimately, the acquirement of additional understanding and the development of

differentiated strategies should result in higher levels of overall purchasing performance.
Obviously, the overall purchasing performance of a firm will be shaped by a large
number of (partly) unforeseen influences. In general, the implementation of
(differentiated) strategies will be of great importance to purchasing performance. This
line of reasoning is followed in figure 3.9, illustrating a chain of ‘cause-and-effect’
relations between the portfolio analysis and (overall) purchasing performance.
use of Differentiation Purchasing Overall

portfolio better better decisions purchasing
in purchasing
analysis strategies and better performance
strategies
Understanding better
more of problems and
possibilities
results in
FIGURE 3.9 Effectiveness of purchasing portfolio analysis: a chain of cause-and-
effect relations
The second research question aims at explaining the use of the portfolio tool. From
another perspective some additional insights might be acquired, differentiating between
users and non-users. What are the main differences that would explain why some firms
use the portfolio analysis and other firms do not? These variables play an important role
in a conceptual use-model that intends to explain the actual use of the portfolio analysis.
This leads to the formulation of:
Research question (2d):

What are the differences between users and non-users of a purchasing portfolio
approach?
For the construction of the use-model, literature has been studied in search for variables
that might explain the use of a portfolio approach in purchasing. The literature study did
find some factors and enablers that can be used as explanatory variables in the use-
model. Five (groups of) variables were found (see figure 3.10). Based on the literature
study, the company size, the share of purchasing, the position of the purchasing
department, purchaser’s professionalism, and the orientation of purchasing have been
identified as explanatory variables.
In his survey Boodie (1997) found a positive relationship between company size and
portfolio use. Less than 10% of the smallest firms used the purchasing portfolio, while the
largest companies (with more than 5,000 employees) show a use percentage of 85%. In
addition a positive relation was suggested between the use of portfolio models and the
percentage of total purchase cost: purchasing share. Clauwaert (1993) posited as well that
the portfolio analysis is being used especially by companies with relatively high
purchasing shares, since these companies rely substantially on their supply markets. Van
Weele et al. (1998) pointed at the maturity of the purchasing function within companies
for a possible explanation of the adoption of a portfolio approach. This is in line with
another finding of the Boodie-study, that organizations with high scores on World Class
Purchasing-variables use a portfolio approach for their purchasing policies (Boodie,
1997). Therefore, the maturity of the purchasing function might be a key variable for
gaining insights in the use and effectiveness of a portfolio approach in purchasing. For
the operationalization of ‘purchasing maturity’ the results of Rozememeijer’ study will be
used. Rozemeijer (2000: 170, 232-233) identified ‘position of purchasing’, ‘purchasing’s
professionalism’, and ‘purchasing orientation’ as the main factors, constituting the
construct variable ‘purchasing maturity’. Obviously, there are large differences between
companies, considering the orientation of the purchasing function. In addition, different
aspects of ‘orientation’ can be distinguished. We could look at the extent to which
purchasing is oriented towards collaboration with suppliers, towards tough negotiations
with suppliers in pursuit of the lowest prices, or at the extent to which the work of
purchasing professionals is related to clerical duties and operational activities.
102
company size
share of purchasing
position of purchasing portfolio use
purchasing’s professionalism
orientation of purchasing
FIGURE 3.10 Conceptual model, explaining the use of a purchasing portfolio approach
The conceptual model allows for the formulation of the following hypotheses:
Hypothesis A3 The company size of users will be larger than the company size of
non-users of the purchasing portfolio analysis .
Hypothesis A4 The purchasing share of users will be higher than the purchasing
share of non-users of the purchasing portfolio analysis.
Hypothesis A5 The position of purchasing is better at companies where the

purchasing portfolio analysis is used, than at non-using companies.
Hypothesis A6 The professionalism of purchasing is higher at companies where the

purchasing portfolio analysis is used, than at non-using companies.
Hypothesis A7a The users of the purchasing portfolio analysis will be more
orientated towards collaboration with suppliers than non-users.
Hypothesis A7b The users of the purchasing portfolio analysis will be less orientated
towards tough negotiations with suppliers than non-users, in
pursuit of the lowest prices.
Hypothesis A7c The users of the purchasing portfolio analysis will be less involved
with clerical and operational activities than non-users.
These hypotheses too will be tested with the data that will be gathered in the survey.
3.8.3 Elaboration of the third research question: the practice of experienced

professionals
In contrast with a growing use and adoption of purchasing portfolio approaches, we have
established that there is a lack of empirical research to provide insights into practical
problems and solutions when using a purchasing portfolio approach. In general, little is
known about the actual use of portfolio models in purchasing management. Publications
do not reveal how purchasing professionals handle measurement issues, which strategic
recommendations are considered, adapted or neglected, which circumstances are of
particular interest for the development and selection of differentiated strategies, what are
specific purposes, goals and results of the use of any purchasing portfolio model. These
issues relate to the third research question that is formulated as:
“How are portfolio models employed by experienced purchasing professionals?” (3)
The literature study has provided the insights necessary to articulate and elaborate this
research question. Obviously, we did not find answers to the question of how portfolio
models are actually used in practice. We did find a number of unanswered questions
with respect to measurement issues and the way strategic issues are handled in practice.
Wind et al. (1983) proved that decisions based on portfolio models are quite sensitive to
operational definitions, the rules to divide dimensions into low and high categories, and
the weighting of the variables constituting composite dimensions. Portfolio models in
general provide a structure for analysis and stimulate strategic thinking. Our analysis of
literature did not reveal how decisions are actually made if they are based on portfolio
analysis. Which strategies are recommended and under which conditions? This means
that in addition to measurement issues, there are unsolved strategic issues, referring to the
recommendations that are provided for the different categories and to the conditions
under which these recommendations are valid. The field research of this study will
address these measurements and strategic issues.
The two dimensions of the Kraljic matrix can be considered as composite dimensions,
consisting of various factors. Kraljic suggested a number of factors that could be relevant
for the ‘measurement’ of these dimensions, but he provided no guidelines or
measurement rules for the combination of these underlying factors, assessing positions in
the matrix. Olsen and Ellram (1997) emphasized that the weighting of each factor forms
the most important part of the implementation process, but that it is at the same time
very subjective. This leads to the following sub-question:
104
Research question (3a):

Considering the unclear guidelines and the unanswered questions with respect to the
measurement of (composite) dimensions and the weighting of factors in the use of a
purchasing portfolio approach, how are these issues handled to the satisfaction of
experienced purchasing professionals?
Kraljic’s strategic recommendations for the categories are usually summarized into four
simple concepts: ‘efficient processing’, ‘exploit power’, ‘strategic partnership’ and
‘volume assurance’. At first sight these are quite logical and sound recommendations.
However, if we take a closer look at the nature of these strategic recommendations, we
must conclude that these strategies are rather generic by nature, providing only rough
indications for the most appropriate supplier strategies.
They merely react and adapt to prevailing circumstances, taking the current power and
dependence structure for granted. Most likely, purchasing professionals will always look
for possibilities to move to different and better positions in the matrix. It is not clear if
and how other positions in the matrix are to be pursued through the implementation of a
recommended purchasing strategy. This leads to the following sub-questions:
Research question (3b):

What kind of specific strategies of purchasing and supply are based on Kraljic’s portfolio
matrix?
Research question (3c):

What kind of movements are considered in the Kraljic matrix in terms of current
positions, future positions (goals) and means (strategies)?
Literature does not reveal how a purchasing portfolio approach could or should be used
in practice. The case studies are aimed at shedding light on these matters.
3.8.4 Elaboration of the fifth research question: power and dependence in the
matrix
We have argued that power and dependence are very important in understanding
buyer/supplier relationships. Earlier we have concluded that the Kraljic matrix has
become the dominant approach in the discipline. In combination with the fact that the
Kraljic approach is not explicit on issues of power and dependence, we have formulated
the fifth and last research question:
“ What is the role of power and dependence in the Kraljic approach?” (5)
In his seminal article Kraljic (1983) introduced a very useful tool for purchasing
management, albeit without any reference to literature or comprehensive theory. Why are
‘profit impact’ and ‘supply risk’ the most appropriate dimensions? To what extent are the
strategic recommendations aimed at using or influencing the relative power position of
the buying company? This lack of theoretical background calls for further investigation.
Kraljic (1983: 112) posited that the general idea of the portfolio approach is to “minimize
supply vulnerability”(dependence) and “make the most of potential buying power”
(power). Obviously, there are clues that point at the role of power and dependence,
which are basic constructs in Resource Dependence Theory (Pfeffer and Salancik, 1978). This
leads us to the following sub-question:
Research question (5a):

Are the foundations of the Kraljic approach to be found in Resource Dependence Theory?
In the next chapter, we will have a closer look at Kraljic’s matrix, answering this sub-
question. Also we will elaborate the overall research question with respect to power and
dependence, which will produce additional (sub)research questions
107
4 A power-dependence perspective
We have found that the Kraljic matrix has become the dominant approach in purchasing
portfolio analysis. The choice of dimensions seems plausible, the framework has proved
to be useful, the strategic recommendations make sense, which has resulted in a large
scale acceptance and adoption. However, we have concluded that there are unanswered
questions with respect to the theoretical foundations that support the selected dimensions
and the provided recommendations.
In this chapter we will make a reasonable case for a connection between the Kraljic
matrix and the resource dependence theory. In addition, we will elaborate on new
perspectives and insights to be gained from a power-dependence perspective on the basic
portfolio approach, which includes the formulation of additional research questions.
4.1 Rotating the Kraljic-matrix: new perspectives

The starting point for the next analysis is that we need more understanding of the
foundations of the Kraljic approach. We will take a closer look at the two basic
dimensions: profit impact and supply risk. The leverage quadrant is usually considered
to be the most favorable in the matrix, because of recommended ‘exploitation of full
purchasing power’. In other words, the combination of a high profit impact and a low
supply risk enhances the relative power position of the buyer. This means that the upper-
left area in the Kraljic model points at a buyer-dominated segment. In contrast, a lower
profit impact and a higher supply risk worsens the buyer’s position, but improves the
relative power of the supplier. In that case a supplier-dominated area is at hand. We can
bring these insights into the Kraljic matrix, by drawing a diagonal that runs from
southeast to northwest, representing a power-axis. Figure 4.1a shows the Kraljic model
with an emphasis on the power-factor. Kempeners and Van Weele (1997) introduced a
comparable figure, splitting the matrix in a ‘buyer dominated’ area and a ‘supplier
dominated’ area. In between there is a neutral territory where parties are expected to
have a power balance. If one party is dominated by the other, than the strategy could be
directed at a movement in the matrix towards a more favorable position.
108
profit impact
high
buyer dominance
power axis
supplier dominance
low high
supply risk
FIGURE 4.1a A power-axis in the Kraljic model
We can enrich the matrix even more by bringing another axis into the representation, that
is at right angle to the power-axis. The combination of a high profit impact and a high
supply risk implies that we are dealing with strategic items and critical supplier
relationships. A low profit impact and a low supply risk are indications for not-important
items. The new axis seems to have a logical connection with ‘importance’, one of the key
concepts of the resource dependence theory (RDT). In the RDT the ‘importance of a
resource’ is a crucial factor, determining organizational dependence. Subsequently, there
are two factors to the importance of any resource exchange: the relative magnitude of the
exchange and the criticality of the resource (Pfeffer and Salancik, 1978). These factors
correspond with the two dimensions of the Kraljic matrix. ‘Profit impact’ can be seen as
equivalent to ‘magnitude’ while ‘supply risk’ comes close to ‘criticality’. Therefore, we
could say that the combination of profit impact and supply risk determines the importance
of an item in the matrix. The importance-variable is hereby positively correlated with
profit impact and with supply risk. We can illustrate this relationship in the Kraljic
matrix by drawing a diagonal that runs from southwest to northeast, representing an
importance-axis (see figure 4.1b). The axis has a compensatory property: a low profit
impact can be compensated by a high supply risk, so that the item is still classified as
‘important’ (and vice versa). What we have actually done, is to rotate the original matrix,
using other derived axes, connected with the resource dependence theory. By rotating the
matrix we have combined the original dimensions (profit impact and supply risk) into
two new axes (power and importance). By doing so, we have dealt with the problem of
the uncertain consequences of product categories being supplier or buyer dominated.
profit impact
high
importance axis
important,
buyer dominated
not important, important,

buyer dominated supplier
dominated
not important,
supplier dominated
power axis
low high
supply risk
FIGURE 4.1b The rotated Kraljic matrix
The main advantage of the discussion so far is that we have provided some new
perspectives, bringing key issues such as power and importance directly into the Kraljic
matrix. We have connected the choice of the dimensions to key elements of the resource
dependence theory. As we have stated earlier (section 3.3), most commonly, four general
strategies are recommended for the corresponding categories. As such these
recommendations accommodate the development of differentiated strategies between the
categories. By drawing the two diagonal axes in the Kraljic matrix, we have included and
created more possibilities: there is differentiation within categories as well. Our ‘rotated’
Kraljic matrix illustrates that:
- the strategic items are important, the relationship is either buyer or supplier dominated;
- the leverage quadrant is buyer dominated, items are either important or not important;
- the bottleneck quadrant is supplier dominated, the items are either important or not
important;
- the non-critical quadrant contains not-important items, the relationship is either buyer
or supplier dominated.
110
By recognizing differentiated purchasing strategies within the categories, we avoid over-

simplistic and deterministic views on purchasing portfolio management. The
interpretation of the four quadrants is improved in the rotated Kraljic matrix. The
quadrants in this matrix differ with respect to the newly constructed dimensions ‘power’
(or dominance) and ‘importance’. The power-perspective plays a significant role in the
process of selecting purchasing and supplier strategies. For instance, non-powerful
companies might prefer relationships with multiple sources for reasons of vulnerability,
while larger, more powerful companies are able to make an optimal use of single sources
for important products. A reasonable instance for this proposition is given by Baatz
(1999), in a case study comparing the procurement of a small and a large manufacturer in
the American ‘white goods’-sector. The large manufacturer followed a strategy of
developing long-term relationships with a small number of main suppliers, offering more
stability and a share of the growth. These relationships require high levels of trust. For
the small company sole sourcing was not an option. The company preferred to maintain
relationships with multiple sources. In this study the difference in supplier strategy was
directly related to and largely explained by differences in relative power.
Although the rotated matrix provides additional insights and possibilities, there are
limitations and drawbacks. The rotated matrix contains no area of balanced power in a
buyer-supplier relationship. This is a problem, since many believe that the most stable
relationships are balanced with respect to the power issue. This means that a new
segment should be recognized in the matrix, broadening the importance-axis into a
power balanced area. A more serious problem is the rather limited conception of ‘power’
and ‘dependence’ being determined by supply risk and profit impact. Therefore, in the
next section we will elaborate on the issues of ‘power’ and ‘dependence’ in buyer-
supplier relationships.
4.2 Power and dependence in buyer-supplier relationships

Power and dependence are generally considered as important concepts for the
understanding of buyer-supplier relationships. Traditionally they are used in marketing
and organizational studies and, though to a lesser extent, in purchasing studies.
Especially in the nineties, the attention was obviously shifted towards partnering and
collaborative relationships, which implied the refraining from power use. More recently
there has been a renewed interest in a power-dependence perspective.
4.2.1 Dependent: to be or not to be?
Organizations are by nature dependent on their environment for the supply of needed
resources of various kinds. Purchasing and supply management are concerned with
resources from suppliers at the input side of the production processes. Just by relying on
the goods and services of a supplier, a certain degree of dependency is a fact. One could
wonder whether ‘dependence’ is good or bad. Steele and Court (1996: 144,145)
introduced the term ‘dependency dilemma’, expressing that we deal with a very basic
issue for purchasing management. There are driving forces pointing at opposite
dependence-relationships. Market trends might suggest placing business with many
suppliers, but corporate strategy might dictate a greater dependence on fewer suppliers.
One of the main strategic questions for any firm is the issue of determining the core
activities. These activities constitute the competitive power of the organization, for the
present and for the future. The control over the core activities will not be passed into the
hands of suppliers, even though it is an actual possibility. Car manufacturers may
outsource all of the component parts of a car, but they are not likely to outsource the
design. Design is considered to be the essence of the organization and the principal factor
differentiates the product (Steele and Court, 1996: 136). There are strong arguments
stressing the positive sides of being dependent on specialist, high performing suppliers.
Companies need complementary cognitive competence from partners to appreciate
opportunities and threats they could not have appreciated themselves. In addition, under
conditions of complexity and variability of technology and markets, there is a greater
need for firms to seek relations with outside sources to compensate for their own
restraints (Nooteboom et al., 2000: 117). Dyer (1996) demonstrated the advantages for
companies in the auto industry of using and being dependent on specialized supplier
networks. Frohlich and Westbrook (2001: 186) reported a growing consensus concerning
the strategic importance of integrating suppliers, manufacturers and customers into
value/supply chains. They recognize different types of integration and coordination: the
physical flow of deliveries, the information technologies and the flow of data within the
supply chain. In general, supply chain integration aims at various forms of performance
improvement, such as productivity gains, inventory reductions, quality improvements,
lead-time and delivery-time reduction, customer service and customer satisfaction
improvements. The price for these advantages might be considered as the acceptance of
higher levels of dependence between the ‘chained’ companies.
In contrast, it is common sense for companies to avoid excessive and dangerous
dependence on any one trading partner. Dependency increases the organization’s
vulnerability by creating problems or uncertainty or unpredictability, it reduces the
organization’s autonomy and degree of strategic freedom, and allows the direct transfer
of benefits and profits from the dependent on the dominant organization (Bourantas,
1989: 140-145). Lusch and Brown (1996: 33) found empirical evidence for a significant
relationship between the dependency structure and the performance of buying
organizations: “Thus, as we expect, when a wholesaler aligns itself with a weaker
supplier, the wholesaler’s performance rises.” Miles et al. (1999) examined the use of
strategic alliances by small technology-based firms. They found that dependence on
alliance relationships showed a negative association with overall performance. Heide and
John (1988: 20) concluded that the financial performance of agencies improved when
112
dependence was reduced, providing that levels of specific investments were high.
Moreover, they recommended that the contingent effects of dependence on performance
should be examined more closely.
The choice of becoming dependent on a supplier involves dependence on the supplier’s
technology and competences. Laseter (1998: 98-100) discussed four broad areas for
assessing supplier competency: capabilities, cost structure, risk factors, and relationship
potential. The capabilities refer to the supplier’s technical and business know-how and its
processes. Does the supplier have sufficient capacity? How flexible are production
processes? Will the supplier be able to meet future needs in terms of volume and product
technology? The supplier’s cost structure determines whether its pricing is sustainable
over the long term. Is a supplier using expensive process technology? How is capital
equipment utilized? The risk factors deal with the supplier’s financial strength and
stability. Is the supplier able to meet its commitments and to invest in the relationship?
Finally, it makes a significant difference whether a supplier is willing to invest in
building a relationship, or not. What are the main interests of a supplier? How does the
supplier fit into the long-term purchasing and supply strategy? Laseter (1998: 100) gave
some serious warnings against simplistic supplier assessments. An overly simple scoring
system is to compare different suppliers along the four dimensions by awarding scores,
multiplying with weights, and adding up the results. However, a low score on a factor
can not always be compensated by a high score on another factor. For instance, the
supplier’s capabilities are either acceptable or unacceptable. Furthermore, it should be
recognized that some shortfalls are more easily corrected than others. Obviously, it
would be easier to improve a supplier statistical process control, than to raise
technological competences.
Resource dependence theory is concerned with the organization’s vulnerability to extra
organizational influence. Confronted by powerful external organizations, it is
recommended that companies should develop strategies to avoid and manage their
dependence on other organizations (Pfeffer and Salancik (1978: 106-111). From another
perspective transaction cost theory too is concerned with the issue of dependence. Under
conditions of high asset specificity, the transaction cost theory predicts and prescribes a
moving away from market transactions (markets) towards vertically integrated
relationships (hierarchies). When asset specific investments are made, they must be
safeguarded against opportunism (Heide and John, 1988: 21). Heide (1994: 73) concludes
that transactions cost theory parallels resource dependence theory in that it views
nonmarket governance as a response to environmental uncertainty and dependence.
This first exploration of power and dependence leads us to the conclusion that
‘dependency’ as such is not a question of being good or being bad. On the one hand,
there are good reasons for avoiding (too much) dependency, but on the other hand there
are
equally good reasons for being dependent on suppliers. As Young and Wilkonson (1997:
56) observed, within buyer-supplier relationships there is a tension caused by the desire
to remain independent and at the same time to depend on others to achieve common
ends. In the next section additional insights will be gained from two streams of ‘resource’
theories.
4.2.2 A resource perspective: RBV and RDT
Within the field of strategic management there is a school that provides for a resource-
based view (RBV) of the firm. Porter (1980) saw the linking of the firm to its environment
as the root of competitive advantage. Industry structure was to be analyzed by the
famous five forces-model to determine the extent of competition and the profit potential.
Companies should adapt to their environment and they should position themselves
according to these forces. By means of this ‘outside-in’ approach Porter paid homage to a
capabilities analysis within the firm (McKiernan, 1997: 793). Throughout the 1980s a
steady flow of criticism kept on going, especially with respect to the restrictive mutual
exclusivity of its generic strategies. This meant that the attention switched to the analysis
of the firms’ resource base, searching for what could lead to a competitive advantage.
This approach is commonly known as ‘inside-out’. The resource-based view has a long
history. Conventional economics focused on traditional resources, like land, labor and
capital (Marshall, 1890). Their quantitative models were based on the principles of
diminishing returns to scale. These models however have serious problems with the
qualitative interactions of human effort and experience, which are believed to provide
unique advantages. In this line of thinking Prahalad and Hamel (1990) argued that the
real sources of competitive advantage lie in the management’s ability to consolidate
corporate-wide skills into competences that empower business units to adapt quickly to
changing opportunities. Their concept of core competences underlined the notion and the
importance of internal resources. Competitive advantage is built on a unique bundle of
skills and competences that are very difficult to imitate. The competitive advantages are
‘sustainable’ in the sense that they depend on a continuous development of key
resources. It is noted that the resource-based view of the firm restricts itself mainly to
internal resources.
A modern variant of the resource-based view of the firm is provided by Hunt and
Morgan (1996: 108-109), presenting their resource-advantage theory of competition. In this
theory logic relationships are posited between ‘resources’ and ‘market position’, and
between ‘market position’ and ‘financial performance’. A superior financial performance
is believed to result from occupying marketplace positions of competitive advantage. So,
competition is a constant struggle for comparative advantage in resources that will yield
favorable marketplace positions. Relevant resources are considered to be: the societal
114
resources on which firms draw, the societal institutions that frame the ‘rules of the game’,
the actions of competitors, the behavior of consumers, and public policy decisions.
Proponents of the resource-based view have argued that firms who are able to accumulate
resources and capabilities that are valuable and scarce will achieve an advantage over
competing firms (Barney, 1991; Dierickx and Cool, 1989). In order for a firm’s resources
to provide competitive advantage, four criteria must be attributable to the resources
(Barney, 1991):
(1) value – the resource must be valuable to the firm;
(2) rareness – the resource must be unique or rare among a firm’s current and potential
competitors;
(3) imperfect imitability – the resource must be imperfectly imitable;
(4) non-substitutability – the resource cannot be substituted with another resource by
competing firms.
Some authors in the RBV literature share the view that purchasing activities have little or
no significant role to play in strategic issues. Firms can not ‘purchase’ a sustainable
competitive advantage on open markets, because many inputs are freely tradable
(Barney, 1991: 117 and Dierickx and Cool, 1989: 1505). Ramsay (2001: 260,261) concluded
that it is extremely difficult, and will tend to be the exception rather than the rule, for
Purchasing to generate and protect competitive advantage. From his perspective,
Purchasing could (only) identify and develop unknown suppliers, enclose known
suppliers, and buy in a hard-to-imitate manner. Other views on purchasing and supply
management emphasize the possibilities of relationships with suppliers for the
development of competitive advantages. Carter and Narasimhan (1996: 24-25) performed
an empirical study to purchasing’s role in and impact on corporate performance. They
concluded that the single most important finding of this study was that purchasing can
have a significant impact on competitive position, profitability, and market share.
Moreover, the purchasing function should be viewed as a key component of firm
competitiveness. Mol (2002: 262) found that firms indeed obtained competitive advantage
by managing supplier relations. The study confirmed a positive effect of managing co-
operative relations on the economic and the strategic performance of firms. Specialized
supplier networks are more generally considered as an important source of competitive
advantage; see for instance Dyer (1996) who supported this proposition with empirical
evidence from the auto industry. It is agreed that the improvement of the product
development process and the access to innovative technology are of paramount
importance (see for instance Håkansson and Erikkson, 1993; Morgan and Garnsey, 1994;
Calabrese, 2000). In some cases, the buyer and supplier form a strategic alliance. This
alliance can be a source of competitive advantage, through idiosyncratic complementary
resource combinations between firms (Knudsen, 2002). The embeddedness of the firm’s
relational assets make it difficult for competitors to imitate (Lorenzoni and Lipparini,
1999). Dyer and Singh (1998: 661) stated that idiosyncratic interfirm linkages with
suppliers and other outside partners can be a source of relational rents, superior
performance and competitive advantage. To conclude, the supplier base has become
more and more important as a resource base for organizations (Lilliecreutz and
Ydreskog, 1999: 64). By nature purchasing and supply management have a focus on the
acquisition of external resources. The roots of the resource based view are to be found in
early organizational theory, namely the Resource Dependence Theory that focuses on the
external environment.
The main principle of the resource dependency theory is that it considers the ability to
acquire and maintain resources as the key to organizational survival. Organizations
require personnel, money, social legitimacy, customers, and a variety of technological
and material inputs. In a very broad sense every organization must transact with
elements in the environment to acquire the many resources that it depends on. However,
in accordance with Pfeffer and Salancik (1978: 2-3), dependency as such is not
problematic. Problems arise when dealing with less dependable suppliers. These
problems are more serious when the buyer is more dependent on a supplier than when
the reverse is the case. The Resource Dependence Theory has proved to be a useful
perspective for understanding and studying inter-organizational relationships. Oliver
and Ebers (1998: 565) conducted a study of concepts, theories and research perspectives
that have been employed within inter-organizational network research. They found that
resource dependence, political power and network approaches “clearly and consistently
emerge as the most frequently employed theories”. From a political-economy perspective,
every buyer/seller relationship is by nature subjected to political processes, in the sense
that parties have partly common goals and partly conflicting interests. One aspect of the
political-economy theory focuses on inter-organizational relationships and their
environmental contexts. There is an extensive research in distribution channels, based on
the principles of the resource dependence theory. Power and dependence have been
investigated extensively in the channels literature. Many researchers have studied topics
such as structure, control, conflict and power with respect to channels of distribution.
Surprisingly, given this extensive research, there is a limited, though growing, use of the
Resource Dependence Theory for studying buyer-seller relationships in business markets.
4.2.3 Power: definitions and associations
Traditionally, writings with respect to negotiating in purchasing have always spent

significant attention to ‘power’ issues. What are the sources of power? What is our
relative power position when negotiating with a specific supplier? Power is believed to
be a very important factor for the course and the results of any negotiation. However, in
the nineties the ‘power’-concept apparently vanished from books and articles with
respect to purchasing and supply management. Concepts such as ‘cooperation’,
‘partnerships’ and ‘integrated Supply Chain Management’ seem to have placed ‘power’
116
in the background. More recently there is a renewed interest in the power perspective in
procurement and supply management (e.g. Laseter, 1998; Gelderman and Van Weele,
2000; Cox, 2001). The special Spring 2001-issue of The Journal of Supply Chain Management
was dedicated to ‘the power perspective in procurement and supply management’. In
that issue Cox (2001: 9) posited: “(…) it is surprising that the intuitive understanding (…)
that all buyer and supplier relationships operate in an environment of relative buyer and
supplier power, appears to have been lost by many practitioners and their advisors.” This
statement is in accordance with Gelderman and Van Weele (2000) who put forward
similar statements. Handfield and Bechtel (2002) concluded that the perception of
dependence is generally considered as an important dimension of any buyer-seller
relationship. Keep et al. (1998) examined the history of business-to-business
relationships. They found that in each investigated case dependence asymmetry was an
important force that influenced relationship development. Faria and Wensley (2002: 607-
609) discovered (much to their surprise) in an empirical study to buyer-supplier
relationships “the substantial expressions of power use and conflict by interactive
partners in every narrative”. The researchers contrasted the central importance of power,
politics and negotiation with “the fetishing and alienating representation of (…)
managerial practices by much of the SCM literature.” We must conclude that ‘power’ and
‘dependence’ are important concepts for the understanding of buyer-seller relationships.
Power has an ideological tinge and produces negative associations, according to Pfeffer
(1981) who recognized that power is a topic that makes people uncomfortable. Using
power is often seen as unethical and counterproductive. The Kraljic approach is criticized
for its strategic recommendations that entail the exploitation of a power advantage
position. The exploitation of power is viewed as dangerous because market condition
change rapidly (Olsen and Ellram, 1997), or because it endangers and obscures the
potential for enhancing productivity and innovativeness in industrial networks (Dubois
and Pedersen, 2002). However, this should not mean that power and dependence are
unimportant. Our study aims to describe and explain certain aspects of purchasing and
supply management. It can not be denied that power and dependence do exist in
business relationships. Ignoring these issues will not make them less important for
understanding buyer-seller relationships. In fact, it would be counterproductive to take a
normative perspective, denouncing and condemning factors that exist and are important
in the everyday reality of purchasing professionals.
It should be noted that an unbalanced relationship does not automatically involve the
actual use or misuse of power. There is an important difference between the possession of
power and the use of power. The distribution of power determines the potential
influence, the actual influence depends on the extent to which the power is used (Arndt,
1983). As Pfeffer (1981: 7) put it: “Power is a property of the system at rest.” The
difference is important because a dominant organization has the choice of not making use
of a power position. Heide and Minor (1992: 275) suggested that, among other variables,
the “restraint in the use of power” is an important element of cooperation between firms.
It might be argued that a high level of power in an exchange relationship will lead to the
exploitation by the dominant party. The rationale is that the possession of power will
encourage a firm to act opportunistically to take advantage of the other party (Frazier
and Rody, 1991). In contrast, it might as well be argued that the role of power provides
for effective coordination of exchange relationships. Power can be used to enhance the
nature of relational exchange between trading partners (Frazier and Antia, 1995). The
distribution of power can become legitimated over time, so that both social actors expect
and value a certain pattern of influence. The exercise of power which has become
legitimated, is expected and even desired in the social context (Pfeffer, 1981). Provan and
Gassenheimer (1994:) pointed out that, while all power arises from dependence, it is not
necessarily enacted or exercised. Power is obviously not always being used. Still, it can
influence decisions and strategies, just because it is recognized by both trading partners.
Brown and Frazier (1978) found in an empirical study that the more manufacturer power
is perceived by dealers, the less those power sources need to be used. There is always a
threat of (mis)use of power to which parties respond in advance (opportunistic behavior).
Studying networks, Thorelli (1986) argued that power is the central concept in networks
analysis, because its mere existence can condition others.
Theoretical foundations of writings with respect to ‘power’ and ‘dependence’ can be

found in the resource dependence theory. The main principle of the resource dependence
theory is that it considers the ability to acquire and maintain resources as the key to
organizational survival. Organizations require personnel, money, social legitimacy,
customers, and a variety of technological and material inputs. In a very broad sense every
organization must transact with elements in the environment to acquire the many
resources that it depends on (Pfeffer and Salancik, 1978). Organizations are by nature
dependant on their environment. This does not mean that ‘dependence’ is either good or
bad. As Pfeffer and Salancik (1978: 3) put it, “dependency as such is not problematic.
Problems arise when dealing with less dependable suppliers”. These problems are more
serious if the buyer is more dependent on a supplier, than vice versa.
In sociology and in organizational studies it is generally agreed that power characterizes

relationships among social actors. Mintzberg (1983: 4) defined power as “the capacity to
effect or affect organizational outcomes”. However, this description reflects too broad a
view on power, since many actions and circumstances may effect organizational
outcomes. An often quoted definition of power is given by Dahl (1957: 202-203): “A has
power over B to the extent that he can get B to do something that B would otherwise not
do.” Blau (1964) conceptualizes power as the ability of persons or groups to impose their
118
wills on others. A similar definition is provided by Emerson (1962: 32): “The power of
actor A over actor B is the amount of resistance on the part of B which can be potentially
overcome by A.” According to El-Ansary and Stern (1972) the power of a channel
member is his ability to control the decision variables in the marketing strategy of
another member. Power refers to the capability of one social actor to overcome resistance
in achieving a desired result or objective (Pfeffer, 1981). Clearly, there is agreement
among authors of frequently cited definitions that power is essentially the ability to cause
someone to do something he/she would not have done otherwise (Gaski, 1984). This
might explain the negative associations provoked by the power-concept.
Another point of consensus is that power is not seen as an absolute quantity. Power
always relates to another social actor. We conclude that a buying organization is not
‘powerful’ in general, but only with respect to a particular supplier in a specific
buyer/seller relationship.
There is a close relationship between power and dependence. Not being dependent, a state
of independence, refers to the concept of autonomy. Dependence poses constraints in the
freedom of choice of actions. A company becomes vulnerable when it looses control over
resources to its exchange partners and finds itself dependent on its partner (Spekman and
Strauss, 1986). With increased dependence also comes strategic vulnerability (Van de Ven,
1976). Dependence implies vulnerability, whether or not on a voluntary basis. Frazier et
al. (1989) define dependence as the degree to which a party needs to maintain its
relationship with another party in order to achieve the desired goals. Dependence on an
exchange partner is often connected to the costs associated with terminating the
relationship and switching to an alternative exchange partner (Joshi and Arnold, 1997;
Heide and John, 1988).
Most treatments of power emphasize the critical role of dependence. In organizational
studies dependencies have traditionally been used to determine the existence of power
relationships (Provan and Gassenheimer, 1994). The role of power in social exchange was
developed by Emerson (1962). In his formulation, the relative dependence between two
actors in an exchange relationship determines their relative power. Power derives from
having resources that the other needs and from controlling the alternative sources of
those resources. Emerson’s power model was elaborated and generalized to the
organizational level in the resource dependence theory as developed by Pfeffer and
Salancik (1978).
4.2.4 Relative power and total interdependence
Organizations always depend, to varying extents, on different suppliers. Obviously, this

dependence is mutual. From the perspective of a supplier, selling organizations depend
on organizations buying their products. Early research on channel dependence focussed
on the effects of a firm’s absolute dependence on its partner, without reference to its
partner’s dependence (e.g. El-Ansary and Stern, 1972). More recent studies however have
incorporated both firms’ dependence (Buchanan, 1992; Kumar et al. 1995; Geyskens et al.,
1996). An appealing and well-known definition states that the relative power of an
organization over another is the result of the net dependence of the one on the other. If A
depends on B more than B depends on A, then B has power over A (Pfeffer, 1981). The
relative power can be defined as “the dependence of one party compared to the
dependence of the other party” (Bacharach and Lawler, 1981: 65). Dickson (1983)
acknowledged that the power of one party over another is a function of relative
dependence. Anderson and Narus (1990) used the term relative dependence, referring to the
difference between a firm’s dependence on its partner and its partner firm’s dependence
on the working partnership. The primary consequence of relative dependence was
indicated as power. In asymmetric relationships, the less dependent partner dominates the
exchange. Buchanan (1992) conceptualized these power-dependence imbalances as the
differences in the value buyers and sellers place on their relationships. Kumar et al.
(1995) defined interdependence asymmetry as the difference between the level of
dependence of the two partners on one another. Symmetrical interdependence exists
when parties are equally dependent on each other. Tuten and Urban (2001) recognized
that the balance of power can be aligned through the mutual dependence of supplier and
buyer. All these contributions and conceptualizations can be traced to Pfeffer’s (1981: 99)
viewpoint that:
the relative power of one social actor over another is the result of the net dependence of the one on
the other.
In this study we will subscribe to the definition that the relative power in a buyer-
supplier relationships is the difference between the dependence of the two parties on one
another. From the buyer’s perspective, his relative power is measured as the difference
between the supplier’s dependence and the buyer’s dependence.
Various researchers have argued that a comprehensive view of the interdependence of a
dyadic relationship should include both interdependence asymmetry (in our definition:
relative power or net dependence) and total interdependence (or total power), e.g.
Bacharach and Lawler, 1981; Gundlach and Cadotte, 1994; Kumar et al., 1995, Frazier and
Antia, 1995; Geyskens et al., 1996. The total interdependence refers to the intensity of a
relationship. A high level of total interdependence is an indicator for a strong,
cooperative long-term relationship in which both parties have invested. Mutual trust and
mutual commitment will characterize those relationships. If both sides possess large
amounts of potential power, there is less likelihood of either trying to convert their
potential power, due to the risk of retaliation (Ramsay, 1996b). Geyskens et al. 1996)
120
found strong evidence that total interdependence enhances both affective and calculative
commitment. In this study we subscribe to the definition of Geyskens et al. (1996: 306)
who posited that:
the value of total interdependence of a relationship refers to the sum of each firm’s dependence on
its partner.
The Dwyer et al. (1987) model of relationship development highlights the process of
gradual expansion of interdependence between buyer and seller. To arrive at a high level
of total interdependence, partners must go through a number of relationship
development phases in which they both continually invest in the relationship (Anderson
and Weitz, 1992).
To conclude, the interdependence of a relationship is reflected by the magnitude (total

interdependence) and the symmetry of dependence (net dependence) between parties.
Based on these insights, an interesting line of approach would be a portfolio matrix based
on the mutual dependence in buyer-supplier relationships. We will elaborate this point in
the next section.
4.3 A mutual dependence-based purchasing portfolio model

In this section we will discuss a number of classification and portfolio models that have a
focus on mutual dependence in buyer-supplier relationships. We will demonstrate that
there is a perfect match with the four positions in the Kraljic matrix, referring to different
kinds of mutual dependence. It is concluded that the Kraljic matrix is a special case of a
more general mutual dependence-based portfolio model.
In the eighties there were some early writings, referring to the concept of mutual
dependence to capture different possibilities of buyer-supplier relationships. To examine
the interdependency between manufacturers and distributors, Dickson (1983) developed
a channel dependence matrix. Each cell in the matrix represents the market share of a
particular manufacturer-distributor combination. The matrix shows the subdivision of a
manufacturer’s market share over distributors. Vice versa, the subdivision of the market
share of a distributor by manufacturers is shown as well. The magnitude and the
dispersion of the different shares is a strong indication for the positional power of
subsequent channel members. Scott and Westbrook (1991) proposed a comparable model,
which they labeled as the customer/supplier dependence grid. As dimensions they identified:
- the percentage of a customer’s purchases which come from a supplier, and
- the percentage of a supplier’s sales which go to a customer.
Framed in the exchange theory (Thibaut and Kelley, 1959), Anderson and Narus (1984)
developed a model based on the buyer’s and the supplier’s motivational investment in a
relationship. By highlighting the mutual motivational investments of parties, a map of
exchange possibilities was drawn, identifying four different kinds of relationships: a

buyer’s market, a supplier’s market, the bilateral maintained relationship, and the
discrete exchange/spot contracts (Dwyer et al., 1987). In essence, mutual dependence is
the key for classifying relationships in these early models. Campbell and Cunningham
(1983) too proposed a typology of buyer-supplier relationships which is based on mutual
dependence: (a) independent relationships, (b) interdependent relationships, and (c)
dependent relationships. Campbell (1985) attributed these relationships to the interplay
of interaction strategies (competitive, cooperative, and command), and to a variety of
other factors, such as the number of buyers and the number of suppliers in a market, the
share the buyer takes of the supplier’s output, the buyer’s need for the supplier’s skills,
the level of specialized can be found in Bensaou (1999: buyer’s and supplier’s specific
investments) and Cox (2001: attributes of buyer and supplier power).
Instead of using concepts such as motivational investments or market share, some
authors proposed to use buyer’s dependence and supplier’s dependence for the
classification of relationships. Blenkhorn and MacKenzie (1994) and Kumar (1996) made
a distinction between ‘low’ dependence and ‘high’ dependence with respect to both the
buyer’s dependence and the supplier’s dependence. Frazier and Antia (1995) proposed a
typology for classifying relationships on the basis of interdependence and environmental
uncertainty. They made a distinction between three levels of interdependence: low
interdependence (balanced), unbalanced exchange, and high interdependence (balanced).
In the case of a mutual dependence, power is in balance. In conclusion, there are four
possible combinations of dependence:
- high mutual dependence (balanced power)
- low mutual dependence (balanced power)
- high supplier’s dependence, low buyer’s dependence (buyer dominated)
- low supplier’s dependence, high buyer’s dependence (supplier dominated).
This classification is in accordance with the empirical findings of Bensaou (1999) who
segmented buyer-supplier relationships into four generic cells: strategic partnerships
(balanced power, high mutual specific investments), market exchange (balanced power,
low mutual specific investments), captive buyer (supplier dominated), and captive
supplier (buyer dominated).
If we take a closer look at these four basic power-positions, we must conclude that the
Kraljic matrix in essence refers to the same four positions. Implicitly, the Kraljic
categories match the four power-dependence combinations. Using our theoretical starting
point that the relative power of the buyer is the difference between the supplier’s
dependence and the buyer’s dependence, we can introduce a mutual dependence matrix.
This model uses ‘supplier dependence’ and ‘buyer dependence’ as the basic dimensions.
Figure 4.2 visualizes this matrix, which is made by adding Kraljic’s item categories to the
mutual dependence matrix (Gelderman, 2000). The generic categories correspond in a
logical way to the four power-positions in the matrix:
122
- strategic items: balanced power based on a high level of interdependence (mutual

dependence)
- non-critical items: balanced power based on a low level of interdependence (mutual
dependence)
- leverage items: buyer dominated based on the net dependence (a positive difference
between supplier’s and buyer’s dependence)
- bottleneck items: supplier dominated based on the net dependence (a negative
difference between supplier’s and buyer’s dependence).
In conclusion, we might say that Kraljic’s matrix is a special case of a more general mutual
dependence-based model, where both models are expressions of a power-dependence
perspective on purchasing portfolio models. To summarize, we have demonstrated that
the categories in the Kraljic matrix correspond to four basic mutual dependence positions
or power positions. It can be assumed that this has contributed to the generic
applicability of the Kraljic matrix and the adoption of the Kraljic approach. Since we have
established a close link with ‘dependence’, it is important to know what determines
organizational dependence. In other words: what are determinants of dependence?
supplier’s dependence
high
leverage: strategic:
buyer dominance interdependent
non-critical: bottleneck:
independent supplier dominance
low high
buyer’s dependence
FIGURE 4.2 Mutual dependence-based purchasing portfolio model
4.4 Determinants of dependence

In this section we will discuss possible determinants and antecedents of organizational
dependence, especially the determinants that constitute buyer-supplier dependence. The
discussion includes conceptual and empirical studies to the determinants of dependence.
The results of this analysis will be used in the mutual dependence-based portfolio model.
4.4.1 Conceptual studies to the determinants of dependence
Literature study resulted in many conceivable factors and variables that contribute to the
level of organizational dependence. In this section we will limit ourselves to the main
conceptual studies in this area, as summarized in figure 4.3.
source perspective: determinants of dependence

dependence of a(n)
Emerson (1962) social actor 1. motivational investment in goals mediated by the other
2. availability of those goals outside the relation with
this social actor
El-Ansary and channel 1. percentage of a channel member’s business

Stern (1972) member 2. commitment to another member (relative importance)
3. difficulty in replacing another member (cost and effort)
Thompson (1967) organization 1. need for resource/performance

2. ability of others to provide the same resource/performance
Jacobs (1974) organization 1. essentiality of the item

2. availability from other sources
Pfeffer and social actor 1. importance of the resource: magnitude and criticality
Salancik (1978) 2. discretion over the resource
3. number of alternative sources
Campbell and buyer 1. number of suppliers

Cunningham 2. share of buyer’s requirements purchased from the supplier
(1983) 3. need for supplier’s skills
supplier 1. number of buyers
2. share of supplier’s output taken by the buyer
3. customized product that requires a special investment
Mintzberg (1983) organization 1. essentiality

2. substitutability
3. concentration
Bourantas (1989) organization 1. importance of the resource

a) relative magnitude
b) functional criticality
c) strategic cruciality
2. substitutability of the source
a) existence of other sources
b) cost of substitution
3. discretion over the resource
FIGURE 4.3 Conceptual studies to the determinants of organizational dependence
A much quoted general definition of (social) dependence is provided by Emerson (1962: 32)
who wrote: “The dependence of actor A upon actor B is (1) directly proportional to A’s
motivational investment in goals mediated by B, and (2) inversely proportional to the
124
availability of those goals outside of the A-B relationship.” At a closer look it states that
dependence is determined in essence by two factors: the need for a resource and the
availability of alternative sources. Other conceptions of interorganizational dependence
identified very comparable determinants. Most closely to Emerson’s definition is
Thompson (1967) who focussed on an organization’s needs for resources and the
presence of other resource providers. Jacobs (1974) introduced the concepts of essentiality
and substitutability from economic theory. He pointed at the question whether A can do
without B (essentiality of a resource) or whether other sources are available
(substitutability of the resource). Dependence of an organization is thus directly
proportional to essentiality of resources controlled by the other organization and
inversely proportional to the availability of these resources from alternate sources.
Departing from Jacobs (1974), Mintzberg (1983) argued that the dispersion and
collaboration of buyers (or suppliers) could influence the possible impact on the behavior
of the supplying organization (or buying organization). Therefore he posited that there
are three key factors that lead to dependency (or power relationships): essentiality,
substitutability and concentration.
According to Pfeffer and Salancik (1978) three factors are critical in determining the
dependence of one organization on another:
1. the importance of the resource;
2. the extent to which the interest group has discretion over the resource allocation and
use;
3. the concentration of resource control, in other words, whether the organization has
access to the resource from additional sources.
These three factors all have a positive correlation with organizational dependence. The
dependence on a resource will increase, when the importance of the resource grows,
when the discretion enhances, and/or when the concentration of resource control
increases. In addition, the importance of a resource is determined by two variables: the
relative magnitude of the resource and the criticality of the resource. The second
determinant of dependence is the extent of discretion over the allocation and the use of a
resource. For an organization to be dependent on a supplier, the resource should be
controlled by this supplier. The most actual forms of discretion are: ownership of the
resource, control of access to the resource, control of the resource’s use, and the ability to
establish rules regulating the possession, allocation, and use of the resource (Bourantas,
1989). Patents of suppliers are a well-known factor to the determination of buyer’s
dependence. Suppliers might possess a concession or license which provides them special
rights (for instance drilling for oil in the Middle East). The third determinant is
concentration of resource control. The dependence on another organization also derives
from the concentration of the resource control, or the extent to which input or output
transactions are made by relatively few, or only one, significant organizations. However,
it is not the number of trading partners that is the critical variable, but whether the
organization has access to the resource from additional sources (Pfeffer and Salancik, 1978).
The availability of alternative sources or resources is generally recognized as an
important factor to organizational dependence. Next to the existence of other sources, the
cost incurred by substitution (switching cost) is another factor to determine the
substitutability of source (Bourantas, 1989).
El-Ansary and Stern (1972) are well known pioneers in the field of measuring power and
dependence in a channel context. El-Ansary and Stern (1972) viewed dependency as a
function of:
1. the percentage of a channel member’s business which he contracts with another member
and the size of the contribution which that business makes to his profits;
2. the commitment of a channel member to another member in terms of the relative
importance of the latter’s marketing policies of him;
3. the difficulty in effort and cost faced by a channel member in attempting to replace
another member as a source of supply or as a customer.
Compared to the conceptualizations of Emerson (1962) and Pfeffer and Salancik (1978),
we must conclude that principally no additional issues are raised, even within a channel
context. The only specific term is the reference to ‘marketing policies’, which can be
connected to their focus on channel leader and control. A channel leader can impose his
marketing policy on other organizations within the distribution channel. In line with this,
their conception of power in a given channel is “the ability to control the decisions
variables in the marketing strategy of another member” (El-Ansary and Stern, 1972: 400,
422).
The same conclusion can be drawn assessing the determinants of buyer’s and supplier’s
dependence that were identified by Campbell and Cunningham (1983). They identified
the number of buyers/suppliers, the share in each other’s business and a third
determinant that refers to forms of criticality (buyer: need for the supplier’s skills;
supplier: customized product that requires a specific investment).
In conclusion, the common findings of these conceptual studies is that dependence is a
function of:
- the importance of the resource and
- the substitutability of the source.
It is noted that these factors are very much in line with Emerson’s (1962) original
conceptualization of (organizational) dependence. The availability of alternative sources
however, is replaced by the concept of substitutability which covers the availability issue
as well as the cost incurred when replacing a trading partner (switching cost). In
addition, the importance of the resource is determined by relative magnitude and
criticality. To gain a more solid basis for a mutual dependence model, we will add an
analysis of empirical studies to the determinants of dependence.
126
4.4.2 Empirical studies to the determinants of dependence
The main question to be answered in this section is: what variables have proved to have a
statistically significant influence on the (level of) organizational dependence?
Traditionally, the dependence-construct has had a prominent role in channels research.
However, our literature review shows that there is only a very limited number of
empirical studies, devoted to the explanation of organizational dependence. We are
inclined to conclude that insignificant attention has been paid to the actual gathering of
empirical evidence on the determinants of dependence. In contrast, most empirical
studies that involve ‘dependence’-issues select organizational dependence as an
explanatory variable. In the last 30 years a variety of phenomenons has been explained by
organizational dependence, including:
- power and control (El-Ansary and Stern, 1972; Etgar, 1976; Brown et al., 1983; Frazier et
al., 1989; Anderson and Narus, 1990; Buvik and Halskau, 2001),
- affective and calculative commitment (Provan and Gassenheimer, 1994; Geyskens, 1996; De
Jong and Nooteboom, 2000; Kim, 2001),
- performance and satisfaction (Heide and John, 1988; Buchanan, 1992; Gassenheimer and
Ramsay, 1994; Lusch and Brown, 1996; Miles et al. 1999; Buvik and Reve, 2001)
- cooperation and competition (Sriram et al., 1992; Young and Wilkinson, 1997)
- governance and contracting (Heide, 1994; Frazier and Anita, 1995; Lusch and Brown,
1996)
- opportunistic behavior (Provan and Skinner, 1989; Nooteboom et al., 1997; Joshi and
Arnold, 1997; Nooteboom et al., 2000)
- relationship development (Keep et al., 1998)
- transaction costs (Sriram et al., 1992)
- innovative activities (Kamath and Liker, 1990)
- integration and adaptation (Hallén et al., 1991; Johnson, 1999)
- trust and supply chain responsiveness (Handfield and Bechtel, 2002).
This not-exhaustive list of phenomenons, explained by organizational dependence,
confirms the earlier notion that dependence is a key construct for understanding buyer-
supplier relationships.
source perspective: determinants of dependence empirical results:

dependence of a significance at
p < .05
Sriram, Krapfel and buyer 1. transaction importance positive
Spekman (1992) 2. specific investments made by negative
suppliers
Ganesan (1994) retailer 1. environmental uncertainty

a) volatility n.s.*
b) diversity negative
2. buyer’s transaction specific positive
investments
3. supplier’s transaction specific positive
investments
Berger, supplier 1. asset specificity positive

Noorderhaven and 2. sales to buyer as % of total sales positive
Nooteboom (1995) 3. knowledge exchange negative
4. goodwill trust negative
5. network embeddedness n.s.*

6. legal safeguarding n.s.*
7. competence trust n.s.*
8. relationship duration n.s.*
9. growth of sales to buyer n.s.*
10. buyer dependency n.s.*
Dant and Gundlach franchisee 1. environmental uncertainty negative

(1998) 2. relationship performance positive
3. relationship duration n.s.*
4. specific investments positive
(multi-unit ownership)
Nooteboom, De buyer 1. supplier’s value to the buyer positive

Jong, Vossen, Helper 2. alternative suppliers negative
and Sako (2000) 3. habituation between partners positive
supplier 1. customer’s value to the supplier positive

2. alternative buyers n.s.*
3. habituation between partners positive
4. dedicated investments positive
* n.s. = not significant
FIGURE 4.4 Empirical studies to the determinants of organizational dependence
Figure 4.4 summarizes the results of the empirical studies that have been found on the
determinants of organizational dependence. Next we will compare and analyse the
determinants with a statistically significant impact on dependence, although there are
128
differences in scope, perspective and design of the studies. In line with the main
conclusion regarding the conceptual studies, it is also concluded that in all empirical
studies two common elements arise:
1. substitutability, and
2. importance.
In all empirical studies the first component has a positive impact on dependence and the
second component has a negative impact, in line with prior expectations. The
substitutability can be subdivided in the level of specific investments and the availability of
alternative sources. The importance of a resource is operationalized in various ways, such
as ‘transaction importance’, ‘share of business’, ‘relationship performance’ and ‘value to
the other’. These operationalizations allow for the use of ‘importance’ as collective noun,
without losing critical information. The remaining variables that are selected in the
empirical studies can not be clustered in a similar, unambiguous way. They include
single-used relationship characteristics, such as ‘goodwill trust’, ‘knowledge exchange’
and ‘habituation’. However, due to the limited number of empirical studies, no decisive
answers were found concerning the statistically significance of the determinants of
dependence. We did find some tentative empirical evidence that ‘importance’ and
‘substitutability’ have a significant impact on dependence, confirming the main finding
of our analysis of the conceptual studies. These basic components will be elaborated into
a tentative model of organizational dependence, in a buyer-supplier context.
4.4.3 Towards a conceptual model
In our tentative conceptual model of the determinants of organizational dependence,

importance is conceptually composed of:
- the financial magnitude of the exchanged resources, and
- the criticality of the resources.
The substitutability encloses two elements too:
- the availability of alternative sources
- the switching cost, incurred when replacing a trading partner.
financial magnitude
+
resource criticality
+ organizational
dependence
availability of alternative
sources
-
switching costs
+
FIGURE 4.5 Tentative model of the determinants of organizational dependence
(1) Financial magnitude

The relative magnitude of a resource has a positive impact with the importance of that
resource (Pfeffer and Salancik, 1978). In a buyer-supplier context the magnitude of a
resource refers to the financial magnitude of the transaction. The financial magnitude of a
resource is measurable by assessing the proportion of total inputs (buyer’s dependence)
or the proportion of total outputs (supplier’s dependence) accounted for by the exchange.
In general: the larger the share, the greater the importance of the resource and the higher
the dependence on the other organization. The money value of purchases made from a
specific supplier can be divided by the organization’s total purchases. Another possibility
is dividing the purchases from a supplier by total spending in the product category. This
means that the relative magnitude of a resource obtained from a supplier could be
measured, assessing the proportion of total purchases or the proportion of a product
category (Bourantas, 1989). The value of these indices range between 0 and 1.
(2) Criticality
The criticality of a resource refers to the functioning of an organization. Criticality
captures the ability of the organization to continue functioning in the absence of the
resource. A resource may be critical to the organization even though it represents a small
proportion of the total input. In terms of the Kraljic matrix, this reflects the basic notion of
bottleneck items. Next to this (functional) criticality Bourantas (1989) introduced the
factor strategic cruciality that corresponds with a broader idea of resource importance:
the resource’s contribution to the organizational’s critical success factors, distinctive
competences or competitive advantages.
130
Dependence may be produced by an organization’s capabilities in performing vital

functions within a supply chain. Influence can be based on critical expertise and
specialized knowledge. In general, the greater the degree in which a buyer relies on the
critical expertise of a supplier, the higher the buyer’s dependence will be (Wilemon, 1971)
The reverse is also true. When a component is critical to the performance and competitive
advantage of an end product, the performance of the supplier is critical as well for the
functioning of the organization (Bello et al., 1999). The relationship with suppliers of
critical components is preferably based on close, integrated partnerships, not on
traditional arm’s-length agreements (Slade, 1993). Increased dependency can be the effect
of increased collaboration and supplier involvement in product development (Wynstra,
1998). Just-In-Time purchasing requires reliable suppliers that are willing and able to
comply to stringent delivery and quality requirements (Handfield, 1993).
Interdependence is the likely result of increased coordination and synchronization of
product and supply in a JIT regime.
We conclude that criticality can mean different things, although a common characteristic
is the indispensability of a resource, in view of the ability of the organization to continue
functioning in the absence of that resource. From a buyer’s perspective criticality refers to
the need for the resources of a supplier. An interrupted supply of a highly critical input
would produce significant problems to the functioning of the buying organization. From
a supplier’s perspective critical outputs are more profitable, embody the supplier’s core
technical and/or market competences, and strategically position the supplier in key
markets (cf Krapfel et al., 1991). Buyers that consume more critical outputs of a supplier
are more highly valued, resulting in a higher level of supplier’s dependence.
(3) Availability of alternatives

Access to the resources from additional sources is always key to organizational
dependence. Organizations will be more dependent on their trading partners, whenever
it is more difficult to acquire resources from others. Buyers and suppliers should be
aware of the number of alternative parties that are available at a given moment.
In economic theory the number of buyers and suppliers is key to the assessment of
market structures. If an organization faces a monopolist that controls a certain resource,
then the dependency involved is hard to master. Where there is just one available
supplier, that supplier will have great power. In contrast, commodity markets will not
produce much dependence on the side of the buyers, but will be problematic for the large
numbers of interchangeable suppliers. Where there are many suppliers, so that each
controls a miniscule fraction of the market, single suppliers will probably be powerless.
Finally, where there are a small number of suppliers, so that each controls a large
proportion of the market, then each supplier can be expected to have oligopolistic power
(Jacobs, 1974).
In the eighties and nineties it is observed that purchasing practice has been changing
from the traditional arm’s length relationships with suppliers to closer, more cooperative
relationships (e.g. Dwyer et al., 1987 and Swift, 1995). Management concepts such as Just-
In-Time management and Supply Chain Management have promoted a general tendency
to work more closely with a limited number of suppliers. These new approaches to
purchasing management have logically resulted in reduced supplier bases, more single
sourcing and higher levels of dependence on smaller numbers of suppliers. In many
cases first-tier suppliers have no liberty in selecting suppliers or products, because they
are pointed out by the customer, the OEM that dominates the supply chain. Once a
component is used, the OEM is reluctant to change the design or select different
components or suppliers. Many large companies therefore specify which suppliers are to
be used by their first-tier suppliers, mainly because particular critical components have to
fit with other critical components (Johnsen et al., 2000).
The dependence on a supplier obviously decreases when new suppliers enter the market.
The most common advice to avoid dependence on a single supplier would be to contract
two or three suppliers (dual or multiple sourcing). Despite early supplier involvement,
large manufactures in Japan still maintain relations with other suppliers, enabling a
double check and upholding a threat of potential shift (Harryson, 1995). Nicholson (1993)
advised purchasing managers to act like risk managers when it came to longer-term
supply strategies. Regarding the issue of ‘alternatives’, purchasing managers should ask
themselves questions such as ‘Are alternatives available?’ and ‘If not, could they be made
available, e.g. by encouraging or assisting other suppliers to develop alternatives, by the
buying firm developing its own alternatives or by changing the current product?’ With
respect to the substitutability Bourantas (1989) suggested that, besides the number of
alternative suppliers, the cost that would arise from substituting one source for another
should be included. This brings us to the fourth determinant of dependence: the
switching cost.
(4) Switching cost

The concept of switching cost refers to the difficulties or costs connected with changing a
firm’s current trading partner. Benito et al. (1999) consider two connected types of
switching costs:
- break-off costs that form a barrier to ending old business relationships, and
- set-up costs that form a barrier to engage in new business relationships.
Replacing an existing supplier with another can produce huge non-recurring expenses. In
the transaction cost theory these costs are ascribed to the transaction-specific investments
(Williamson, 1975). Transaction-specific investments can not be used for alternative
purposes without additional costs: switching costs. According to the transaction cost
theory, switching costs largely determine the choice of the optimal coordinating
132
mechanism: markets or hierarchies. Asset specificity refers to “the degree to which an

asset can be redeployed to alternative uses and by alternative users without the sacrifice
of productive value” (Williamson, 1989: 142). Many transactions require the use of
specific assets. In general, the more specific the asset, the harder it is to deploy for
alternatives uses, if the relationship should be terminated. Jackson (1985) discussed
switching costs as a key to understanding organizational buying behavior. Industrial
organizations invest in their relationship with a supplier through some combination of
money, people, lasting assets, and procedures. Money is directly invested to pay for the
purchases, especially capital goods (equipment and systems). Sometimes other lasting
assets are bought to use the new product. In addition, people are hired or trained to
work. From the seller’s point of view, it might be tempting to make it as difficult and
expensive as possible for buyers to switch to another supplier. Examples of such
(marketing) tactics are to lock-in the buyer to a particular system, to create strong links
(physical distribution, electronic links), to stimulate personal contacts, to provide ‘free’
supply of software (Nicholson, 1993).
Both buying and supplying organizations invest in the relationship with their trading
partner. The development and production of dedicated equipment assigned exclusively
to one customer, increases the switching costs and the supplier’s dependence. In the
automotive industry suppliers even build dedicated plants literally next door to their
customer’s assembly plant to enable just-in-time delivery. On the other hand, buying
organizations also face relation specific investments, making significant investments in
suppliers. For instance, 40 to 50 Honda purchasing professionals work full-time on-site
with suppliers as part of their unique supplier development program, presenting a major
investment by Honda (Laseter, 1998).
The issue of switching costs will be particularly important in collaborative, long-term

buyer-supplier relationships. In order to take full advantage of an ongoing relationship,
companies gradually adapt their resources and routines to the specific needs of that
relationship (Benito et al., 1999). This is likely to result in (mutual) exit barriers, high
switching costs and therefore a high level of interdependence.
To summarize, in this study we have selected four determinants of organizational

dependence on grounds of compelling logic and relevancy. Analysis of conceptual and
empirical studies have resulted in the following basic predictor variables: financial
magnitude, criticality, availability of alternative partners and switching cost. In the case
studies this conceptualization will be compared to the practical insights of purchasing
practitioners, allowing for adaptation in our tentative model of dependence.
4.5 Managing dependence

In this section we will discuss two main theoretical paradigms to clarify and explain the
role of ‘dependence’ in buyer-supplier relations: resource dependence theory and
transaction cost theory. Both theories describe and prescribe ways and strategies for
‘managing organizational dependence’. In addition, in line with the scope of this study,
connections will be made with the purchasing portfolio approach.
4.5.1 Resource dependence theory
In general, resource dependence theory characterizes the links among organizations as a

set of power relations based on the exchange of resources (Ulrich and Barny, 1984).
Organizations are assumed to change their dependence relationships by minimizing their
own dependence and by maximizing the dependence of other organizations on
themselves (Jacobs, 1974; Pfeffer, 1981). To obtain favorable exchange conditions,
companies should be engaged in avoiding, managing and balancing their dependence on
other organizations. Emerson (1962) presented three generic strategies to be applicable
for a company A which strives to improve its dependence position with company B:
(1) cultivate other sources than B,
(2) increase the importance of the exchange for B,
(3) decrease the importance of the exchange for A.
Pfeffer and Salancik (1978) discussed a number of business strategies to reduce or avoid
dependence, which are usually not within reach of purchasing management and
purchasing departments. They include: initiate mergers and vertical integration, alter the
purpose and structure of the organization, develop substitutable resources, diversify into
different lines of business, eliminate the concentration of control through antitrust suits
and the acquisition of countervailing control. In line with these overall business
strategies, many Japanese firms have taken an equity-approach to manage the threat of
suppliers. They have purchased a substantial equity position in their critical suppliers
who on their turn have purchased an equity position in their critical customers (Barney,
1996). These cross-equity relations tend to be very stable and form the basis of cooperate
relations (Ouchi, 1984; Harryson, 1995).
In line with resource dependence theory there are a number of more operational ways in
which organizations could manage their dependence on suppliers. A distinction can be
made between:
1. methods that reduce or even eliminate dependence and
2. methods that help limit the negative effects of dependence.
Strategies in the first category can be linked to the various determinants of dependence.
By changing the ‘value’ of such a determinant, the level of dependence will be directly
changed too. Examples would be to find or to develop other suppliers for the same input
(availability of alternative suppliers), to implement standardization programs (reduce
switching cost), or to spread business over multiple sources (financial magnitude).
134
Methods in the second category do not reduce the dependence on the exchange
relationship. They do not eliminate the organization’s vulnerability, because they do not
remove the basic sources of dependence. Examples are keeping excessive stocks to
survive periods of uncertainty, instability and scarcity, using contract negotiations to
acquire countervailing legal power, and the socialization of executives or the
development of norms and values which will restrict the exercise of interorganizational
influence (Pfeffer and Salancik, 1978; Bourantas, 1989).
To conclude, resource dependence theory is very concerned with the issue of avoiding
and reducing dependence on other organizations. Various ways and strategies are
proposed for reducing and avoiding external dependence.
However, it should be added that resource dependence theory places dependence

problems in perspective. Pfeffer and Salancik (1978: 3) concluded that “the fact that
organizations are dependent on their environment does not, in itself, make their existence
problematic. Problems arise not merely because organizations are dependent, but because
the environment is not dependable.” Buyers therefore will be challenged with problems of
dependence, especially when dealing with suppliers who are perceived as less
dependable. These problems are more serious if the buyer is more dependent on a
supplier.
4.5.2 Transaction cost theory
Transaction cost theory focuses on how an organization should organize its boundary-
spanning activities so as to minimize the sum of its production and transaction costs.
From an economic perspective, it is assumed that the most efficient relationship will
prevail for any given transaction configuration. In his early writings Williamson (1975,
1985) identified markets and hierarchies as the two modes of governance. Under
conditions where transactions are characterized by non-specific investments and the
availability of alternative trading partners market governance (e.g. spot contracts) will
prevail. Under market governance both buyer and supplier can move with relative ease
among exchange partners. One could argue that in terms of the Kraljic matrix, market
governance refers to the non-critical quadrant. As transactions are more characterized by
relationship specific investments and exchange partners become more interdependent,
the cost of strict market contracting becomes inexpensive and inefficient (Spekman and
Strauss, 1986). Moreover, the more powerful partner is likely to be motivated to behave in
a self-serving opportunistic manner. From this perspective, transaction cost theory
explains why an organization would choose to internalize the production of components,
even though its production costs might be higher than those offered by a specialized firm
(Barringer and Harrison, 2000). The basic choice is a make-or-buy issue, resulting in
either ‘make’ (hierarchical governance) or ‘buy’ (market governance). In later writings,
Williamson (1991) acknowledged additional interorganizational forms, such as a joint
venture or a network structure. The make-or-buy decision is expanded to make, buy, or

partner. In general, such bilateral governance is characterized by information sharing
and joint long-term planning. It aims at reducing costs of market governance, while
allowing the partners to remain flexible which is often lost in vertical integration
(Buzzell, 1983).
The fundamental prediction of transaction cost theory is that when transaction specific
investments are made, they must be safeguarded against opportunism. Protection is
achieved by moving away from an arm’s length market relationship toward a vertically,
integrated relationship (bureaucratic control). This traditional safeguard is obviously not
always feasible or desirable. A theoretical extension of the transaction cost theory is
offered by Stinchcombe (1985), who argued that safeguards can be obtained also by
explicit contractual agreements. Administrative control can also be inserted in such
agreements. Buvik and Reve (2001) for instance, found empirical evidence for the
relationship between unilaterally deployed specific assets and the level of formalized
contracting. In contrast, Achrol and Gundlach (1999) found that legal contracts are
largely ineffective in mitigating opportunism. In addition, contractual safeguards will not
be possible if the party with specific assets at risk is much smaller than the other party
and has less bargaining power (Heide and John, 1988). Despite the opportunities of even
the most elaborate contracts as safeguards, contracts are constrained in their ability to
‘presentiate’ or foresee and account for future contingencies (Macneil, 1980). There is
evidence and appealing logic for the conclusion that the use of explicit contracts is
associated with perceptions of distrust and low commitment (Jap and Ganesan, 2000).
The absence of mutual commitment is obviously not a condition that encourages and
stimulates joint development programs. This leads us to the conclusion that the classical
solutions of transaction cost theory are generally insufficient to solve the safeguarding
problem of relation-specific investments.
4.5.3 The Kraljic approach
In this section we will look at the Kraljic portfolio approach from a dependence
perspective. Figure 4.6 summarizes the generic recommendations that are provided for
the quadrants of the matrix. In addition, columns are added for the main objectives to be
pursued by the strategic recommendations and for their intended and expected impact on
the power-dependence relationships with suppliers.
From a dependence perspective we might conclude that the recommendations for the
bottleneck items and for the non-critical items have no significant impact on the buyers’
and supplier’s dependence. ‘Volume insurance’ and ‘efficient processing’ are adaptive
methods, aimed at other objectives than changing the prevailing power-dependence
relationships with suppliers. The recommendations handle problems that are a
consequence of the matrix position: the negative effects of a shortage of supplies and the
negative effects on the operational purchasing costs.
136
Quite a different picture is found in the leverage and strategic quadrant, where either the
buyer’s dependence or the supplier’s dependence is increased, or both. In other words,
the provided recommendations are aimed at changing the relative power position
towards suppliers. Buyers are advised to pro-actively use possibilities, especially in light
of the existing relationships with suppliers, attributed in terms of relative ‘dominance’.
The exploitation of purchasing power will expand the buyer’s dominance in the
relationship even more. In cases of supplier’s dominance more restraint actions are
recommended, such as ‘find material substitutes’ and ‘accept higher prices or long-term
obligations to prevent shortages of supply’. Finally, in case of power balance, a rather
adaptive strategy is recommended to match and to develop the existing (long-term)
relationship with the supplier in the strategic quadrant.
We conclude that there are definitely elements of ‘managing organizational dependence’
to be found in Kraljic’s basic recommendations. Obviously, power-influencing strategies
are pursued in practice, based on a portfolio approach to the development of purchasing
strategies. In the empirical part of this study we will elaborate on the various portfolio-
based strategies that are recognized in practice and the conditions that allow or promote
the selection of these strategies. In the next section however, we will first reflect on other
views on the issues of power and dependence, offering critical perspectives on the
selection and use of ‘dependence reducing strategies’ in buyer-supplier relationships.
item strategic objectives impact on dependence

category recommendations
bottleneck volume insurance prevent shortage of critical reduces the negative effects of resource
supplies criticality and dependence,
does not remove the sources for the high
level of buyer’s dependence
non-critical efficient processing reduce cost of ordering and does not affect the low level of buyer’s
materials handling dependence
leverage exploit purchasing reduce direct purchasing increases the supplier’s dependence
power cost
strategic (1): exploit power increase overall supplier increases the supplier’s dependence
buyer’s performance, incl.
dominance favorable pricing and
reduced inventories
strategic (2): diversify (a) find material substitutes/ in case of (a) reduces the high level of
supplier’s new suppliers or (b) prevent buyer’s dependence or
dominance shortage of critical supplies, in case of (b) does not change the high
e.g. accept higher prices or level of buyer’s dependence
longer-term obligations
strategic (3): balance develop long-term supply increases the high levels of buyer’s and
balanced relationships supplier’s dependence
relationship
FIGURE 4.6 Kraljic’s recommendations, objectives and their impact on dependence
4.5.4 Critical views on power-dependence strategies
In this section we will present some other, critical views on the issues of dependence and
reducing dependence in buyer-supplier relationships. The pursuit of power strategies
aimed at influencing dependence conditions has raised some significant issues in
literature. The main objections will be summarized that take a stand against strategies
that are unilaterally aimed at changing the relative power balance in one’s own favor. In
addition, counter-arguments will be provided resulting in a more complex picture and
additional insights on the issues at hand.
Benefits of being dependent

A focus on the negative aspects of dependence can be seen as just one side of the picture,
because there will be benefits as well. Increased dependence on a supplier is the logical
consequence of purchasing and business strategies that include early supplier
involvement (Wynstra, 1998). Johansson (1997) found empirical evidence that (even)
more dependent firms did not persist in dependence balancing strategies because that
138
would jeopardize the privileges that are not attainable in a less asymmetrical structure.
Nooteboom et al. (2000) concluded that under greater uncertainty (technology and
markets), there is a greater need for firms to seek relations with sources to compensate
their own cognitive and technological constraints. At the same time, they acknowledged
the fact that this consequence does not solve hold-up problems. However, by engaging in
specific investments one may develop a unique competence value for the partner, which
makes him dependent. If this is the case, the hold-up risk is effectively neutralized. The
authors found empirical evidence for this causal loop, which was labeled as the loop of
self-interested commitment: specific investments increase dependence on the other party,
but can also serve to increase one’s value to the partner, which makes him dependent and
reduces the incentive towards opportunism.
Comment: the fact that there are benefits and returns attached to dependence raised the
expectation that there will be risks as well. Moreover, there should be a positive relation
between risks and (expected) returns, like in any investment decision. The notion of
‘balance’ can also be found in the loop of self-interested commitment: an increased
dependence is not problematic under the condition that the other party becomes more
dependent as well.
Relational norms
We have concluded that the classical recommendations of transaction theory offer no
final solutions to the issues of unilateral dependence and the connected safeguard
problems. In addition to transaction theory, there is a more sociological perspective that
believes that a richer understanding of interfirm organization is available through
studying the embeddedness of economic actors (Granovetter, 1985). The development of
relational norms is generally considered as an alternative way for safeguarding specific
investments. Developing closer ties with exchange partners will reduce opportunistic
behavior. Heide and John (1990) found that suppliers who have invested specific assets in
a manufacturer, established close ties with that manufacturer by means of joint action
and expectations of continuity. In a related study, Heide and John (1992) demonstrated
that relational norms (i.e. flexibility, information exchange, and solidarity) were present
in buyer-supplier relationships and enabled buyers to protect their investments by
gaining control over supplier decision making, thus reducing the hazards of
opportunism.
Comment: in contrast to the proposition that buyers could consider the existence of
relational norms as safeguards, it can be argued that buyers have not removed or
handled the sources of their dependence on suppliers, which means that they remain
vulnerable to the (opportunistic) behavior of suppliers.
In addition, it can be argued that opportunistic, self-interest-seeking behavior might not
be perceived to be a problem but rather part of normal practice of doing business (cf. Cox,
1996). Companies assume that also their partners are and should be loyal to their own
interests. Trading partners are able to develop a set of relational norms in which
‘opportunism’ is not perceived to be inappropriate, but rather sound business practice
(Young and Wilkinson, 1997). In other words, certain kinds of opportunistic competitive
behavior are in fact part of the relational norms between companies.
A related issue is the exact meaning of opportunism which might have given rise to
misunderstandings. In transaction cost theory opportunism is defined as “self interest
seeking with guile” (Williamson, 1975: 6). Opportunism is often conceptualized in
behavioral terms such as ‘deceitful withholding of information’ and ‘failing to keep
promises’ (John, 1984). These behaviors should not be confused with normal business
practice based on implicitly shared (relational) norms, such as hard bargaining,
demonstrating competitive behavior, entering into necessary confrontations and
constructive conflicts (Young and Wilkinson, 1997). In general, companies should be
engaged in self-interest-seeking behavior, however without guile.
Counterproductive power strategies

The use of power-dependence influencing strategies is often viewed as
counterproductive. Johansson (1997) pointed at several more subtle strategies to be used
as alternatives to the more overt and explicit power strategies. The exploitation of power
is considered dangerous because market conditions change rapidly (Olsen and Ellram,
1997) or because it endangers and obscures the potential for enhancing productivity and
innovativeness in industrial networks (Dubois and Pedersen, 2002). Stannack (1996: 56)
concluded that purchasing needs to use a new range of power strategies in order to create
improvements in supply chain performance. He added that the traditional tool (money) is
not appropriate for the intended behavioral change of suppliers: “money can buy
compliance, it cannot buy commitment and the flexibility, innovation and responsiveness
that commitment can bring.”
Comment: there is a large number of publications that express a believe in the superior
advantages of ‘trust’ in buyer-seller relationships (for instance Morgan and Hunt, 1994;
Kumar, 1996; Doney and Cannon, 1997; Smith and Barclay, 1997 and Garbarino and
Johnson, 1999). Laseter (1998: 93) sharply criticized ‘popular literature’ that suggests that
‘trust’ is the key to effective relationships. He stated: “But simply preaching trust as a
new gospel is unworkable at best and definitely naïve. To build relationships on a more
solid footing, a translation of trust is suggested into actions requiring: mutual
dependence, goal congruence, and knowledge of competency.” The key issue here is not
the presence of power or the exercise of power as such. Earlier we have emphasized that
power can be used in a positive way, providing for effective coordination of exchange
relationships. Power can be used to enhance the nature of relational exchange between
trading partners (Frazier and Antia, 1995). The distribution of power can become
legitimated over time, so that both social actors expect and value a certain pattern of
influence. The exercise of power which has become legitimated, is expected and even
140
desired in the social context (Pfeffer, 1981). As we have seen earlier, various kinds of
power-influencing strategies are likely to be within the accepted boundaries of the
relational norms in many buyer-supplier relationships
Network approach
Stannack (1996) expressed his concern about the fact that exchange theories of power rest
upon individual transactions and fail to take into account transactions which are
embedded in networks. Anderson et al. (1994), for instance, have demonstrated that
networks have a significant effect upon dyadic interactions. The industrial network
approach strongly suggest that firms will have an interest in preserving an overall
network stability. Håkansson (1989) emphasized technology development within
networks, driven the firm’s pursuit of their own self-interests. If relationships are not
stable enough, the firms will not take the risk of making further commitments. This
implies that regardless of the power and dependence distribution, there is a limit to how
far a firm will be prepared to go in pursuing its self-interest and strengthen its own
position at the expense of others. Jarillo (1988) argued that networks are more efficient
than markets or hierarchies, because they tend to minimize transactions costs for
participating firms. Moreover, opportunism is minimized through mutual trust and the
desire to remain in the network.
Comment: it is recognized that the industrial network approach demonstrates an
optimistic view on collaboration within networks, where its main focus is on cooperative
aspects buyer-supplier relationships (Johansson, 1997). After many years of double-digit
growth, in 2001 a number of industries suffered from severe downturns (e.g. semi-
conductor, telecom). Several companies tried to pass their problems down the supply
chain. Some failed, others succeeded. Not until then, companies had to discover in a
painful way the real value of all their ‘partnerships’. Obviously, many firms suffered
from the seamy side of supply chain integration and asymmetrical dependence
structures. From a power-dependence perspective this will not come as a surprise.
Moreover, resource dependence theory would predict that companies who suffered most
are the less powerful and more dependent links of the supply chain.
4.5.5 Conclusions
We have reviewed resource dependence theory, transaction cost theory and Kraljic’s
portfolio model for their approaches to and ways of ‘managing dependence’. A common
characteristic of these perspectives is their focus on avoiding and reducing the
organization’s dependence. A useful distinction can be made between: (1) methods that
aim to reduce the dependence on another organization and (2) methods that limit the
negative effects of dependence but do not remove or reduce the sources of dependence. In
addition, counter-arguments were provided too, indicating that there are risks attached
to ‘being dependent’, but that it also brings in returns. No simple answers were found to
the complex issue of dependency in buyer-supplier relationships. Like in all investment

decisions, risks and returns should be fully considered and balanced.
Undoubtedly, there is no one best way of managing supplier relationships, and there is
no best way of ‘managing dependence’. Purchasing managers are challenged by the task
of developing differentiated purchasing and supplier strategies as a response to varying
circumstances and situations. Obviously, the Kraljic approach has much to offer. In line
with the scope of this study, after extensive literature study, we will next investigate the
possibilities of the portfolio approach in practice.
4.6 Answers to and further elaboration of the fifth research

question
The review of literature has confirmed that power and dependence should be recognized
as significant concepts for the understanding of any buyer-supplier relationship. Only
recently, there has been a renewed interest in the issues of power and dependence in
purchasing and supply management (see for instance Cox, 2001).
In this chapter we have elaborated on issues of power and dependence in buyer-supplier
relationships, with a special focus on relevance for the purchasing portfolio approach.
Analysis of literature has brought the answer to research question 5a which states:
“Are the foundations of the Kraljic approach to be found in Resource Dependence Theory?”
We have analyzed the dimensions of the Kraljic matrix, connecting them with the resource
dependence theory where a theoretical framework was found for the choice of the
dimensions (Pfeffer and Salancik, 1978). In addition, we have argued that the categories in
the Kraljic matrix correspond to four basic power-and-dependence positions:
- strategic: balanced power, interdependent
- non-critical: balanced power, independent
- leverage: buyer dominated
- bottleneck: supplier dominated.
Finally, we have reviewed Kraljic’s strategic recommendations from a power-dependence
perspective. An overview was provided of the main objectives to be pursued by the
strategic recommendations and their impact on the power-dependence relationships with
suppliers. The findings suggest that the element of ‘managing dependence’ is definitely
to be found in Kraljic’s recommendations. Dubois and Pedersen (2002) have
acknowledged that Kraljic’s portfolio approach is based on power-dependence balancing
issues, as concluded earlier by Gelderman and Van Weele (2000). In conclusion, we have
made a reasonable case that the resource dependence theory should indeed be considered
as the (implicitly applied) theoretical foundation for the Kraljic portfolio approach.
142
In addition, we have concluded that a comprehensive view of the dyadic nature of buyer-
supplier relationships should include the assessment of:
(1) the difference between buyer’s and supplier’s dependence (net dependence) which
corresponds with the relative power between parties;
(2) the sum of buyer’s and supplier’s dependence (total interdependence) which indicates
the mutual dependence and the intensity of the relationship between parties.
These concepts can be applied to the Kraljic matrix, which leads to the following research
question (5b):
“What are the levels of power and (total) interdependence in the categories of the Kraljic matrix?”
In other words, what will we find if we describe the categories in terms of their total
interdependence (low, medium, high) and their net dependence (positive or negative)?
Based on our analysis of the Kraljic matrix and the four categories, we will posit a
number of prior expectations.
If
BD = buyer’s dependence, and
SD = supplier’s dependence
then we are expecting
- supplier dominance and a moderate level of interdependence in the bottleneck quadrant:

BD > SD
(BD + SD) is medium
- a power balance and a low level of interdependence in the non-critical quadrant:

BD = SD
(BD + SD) is low
- buyer dominance and a moderate level of interdependence in the leverage quadrant:

BD < SD
(BD + SD) is medium
- a power balance and a high level of interdependence in the strategic quadrant:

BD = SD
(BD + SD) is high
We will compare these theoretical propositions with the survey data. The results of this
analysis will contribute to our understanding of power and dependence in
buyer/supplier relationships in general and in the Kraljic matrix in particular.
The further study of power and dependence in this chapter has lead us to an explorative
study of the determinants of organizational dependence. The results provide the basis for
research question 5c:
“What are the determinants of buyers’ dependence in the categories of the Kraljic matrix? Idem
for supplier’s dependence.”
Literature suggested that organizational dependence is a function of the importance of the

exchanged resources and the substitutability of the source. Analysis of various conceptual
and empirical studies to the determinants of dependence has resulted in a tentative
model where dependence is explained by four variables:
- financial magnitude,
- resource criticality,
- availability of alternative sources and
- switching cost.
In the case studies this conceptualization will be propounded to purchasing practitioners,
which allows for adaptation of the model and for the formulation of hypotheses. The
assumed impact of the variables on dependence will be statistically tested by means of
the survey data. The results of this analysis should be of significant relevance to
purchasing practitioners: only if one is clear about the determinants of and their impact
on a certain dependence position in a Kraljic category, will it be possible to reflect on
purchasing and supplier strategies aimed at changing that position
145
5 The cases studies
In contrast with a growing use and adoption, there is a lack of insight-providing

empirical research. Publications do not reveal how purchasing professionals handle
dimensions and measurement issues, which strategic recommendations are considered,
adapted or neglected, which circumstances are of particular interest for the development
and selection of differentiated strategies. They do not reveal what the specific purposes,
goals and results are of the use of any purchasing portfolio model. The case studies
address this gap. Additionally, reactions are gathered on the conceptual mutual-
dependence model that was introduced in the former chapter.
5.1 Methodology
In general, little is known about the actual use of portfolio models in purchasing
management. The main objective of the case studies is to identify and to describe
advanced current practices with respect to purchasing portfolio models. The case studies
are aimed at answering the third major research question:
“How are portfolio models employed by experienced purchasing professionals?”
In chapter 4 we have elaborated this question into a set of articulated research questions,
to be answered in the case studies:
3a) Considering the unclear guidelines and the unanswered questions with
respect to the measurement of (composite) dimensions and the weighting of
factors in the use of a purchasing portfolio approach, how are these issues
handled to the satisfaction of experienced purchasing professionals?
3b) What kind of specific strategies of purchasing and supply are based on
Kraljic’s portfolio matrix?
3c) What kind of movements are considered in the Kraljic matrix, in terms
of current positions, future positions (goals) and means (strategies)?
These research questions address issues with respect to the measurement of dimensions
and with respect to the strategic recommendations (purchasing strategies).
146
The case study method was chosen for a number of reasons (Yin, 1994; Miles and
Huberman, 1994). First, there is limited research on the actual use and possibilities of a
portfolio approach in purchasing and supply management. Publications are conceptual
or anecdotical by nature. Since we wanted to identify and describe advanced practices,
the case study method was a logical choice. In general, the case study method is advised
when the phenomenon of interest is not well understood. Remenyi et al. (1998) pointed at
a more focused view, i.e. that most case studies in business and management try to
illuminate a decision, or a set of decisions: why they were taken, how they were
implemented, and with what result. Second, case study research is preferable when the
research focuses mainly on ‘how’ and ‘why’ questions. We wanted to gain insights in the
use and the possibilities of a portfolio approach, exploring and identifying the advanced
practices of an experienced company. The questions in our research deal with exploratory
issues, rather than frequency or incidence. Third, a case study research has a distinct
advantage when these questions are being asked about a contemporary set of events over
which the investigator has little or no control.
Three in-depth case studies were conducted, involving three industrial firms in the
Netherlands. The case companies were selected and asked to participate in the research,
based on their experience with the use of a purchasing portfolio approach. For our
research we only included firms who have had more than four years of experience in
purchasing portfolio analysis and who are at present making intensive and significant
use of the portfolio approach. This selective, non-random sample is in line with the
exploratory nature of the research questions at hand. Purposive sampling enables the
researcher to use judgement to select cases that are most suited for answering the
research questions (Saunders et al., 2000). This form of sample is often used when
working with small samples and when one wishes to select cases that are particularly
informative (Neuman, 1997). Obviously, the composition of the sample is not made with
the intention to be statistically representative of a population. The cases were studied
sequentially, one after another.
Because we wanted to explore different possibilities of the portfolio approaches, different
units of analysis were included. There are important differences in the companies with
respect to the level of analysis. The first case study dealt with the use of a portfolio
approach on the corporate level of the company. The portfolio approach then is aimed at
gaining synergy and leverage across business units. The second case study is positioned
at the level of a large multinational business unit with many plants all over the world.
The third case study focuses on a business unit of a fairly small industrial company. The
variety in levels should reveal different kinds of practices, according to specific
circumstances and objectives. We are aware of the limitations of this approach and it is
not intended to give an exhaustive treatment of portfolio methods in purchasing and
supply management. Although case studies may deal with rather unique situations, their
results and conclusions can be compared. Comparison of cases may lead to the
formulation of theoretical conjectures (Remenyi, 1998). We will conclude this chapter
with a comparative analysis of the three case studies. Ultimately, the comparative
analysis has resulted in:
- a modification of our conceptual model of buyer’s and supplier’s dependence
(determinants),
- a description and overview of solutions to the measurement issues, and
- the development of a conceptual model of strategic directions in the matrix (strategic
issues).
Respondents were interviewed on the basis of a semi-structured questionnaire, allowing

for elucidation, elaboration and clarification. The use of a semi-structured interview is in
line with the nature of the exploratory research objectives. The interview guide can be
found in Appendix A.
Data were collected primarily through interviews and secondary resources, such as
Internet web sites, annual reports, internal reports, and purchasing plans. Altogether 26
interviews were conducted with 15 informants. The case studies entailed the use of a key-
informant method, interviewing a limited selected number of participants. The informants
were chosen not on a random basis, but because they were considered to have specialized
knowledge of and experience with the use of purchasing portfolio models (judgement
sample). This approach is appropriate for exploratory studies. In all case studies the first
key-informants were high-placed purchasing officers. In the DSM case study, the director
purchasing services was used as the first key-informant, in the Akzo Nobel Coatings case
study it was the purchasing vice president of a business unit, and in the TE STRAKE case
study we started with the strategic buyer of the company. This approach can be justified
by the fact that we needed a clear overview of the entire purchasing operation and that
we needed an entry to other respondents within the companies. The other informants
were chosen through a snowballing technique whereby the first informant nominated
other key-informants. These informants were all chosen for their specialized knowledge
of and experience with the use of portfolio models in real-life purchasing, notably
business unit managers, purchasing managers and senior buyers (see appendix B,
Sources of the case studies). We wanted to include and account for possible differences in
experience and views, related to different perspectives and positions within the
companies.
In-depth semi-structured interviews were conducted by means of face-to-face contact.
Several rounds of interviews were conducted with the respondents and as we reported
back the tentative analysis and conclusions from earlier rounds, we provided them with
the opportunity:
148
- to check and recheck interim reports,

- to improve the match with the intended information, and
- to explore issues in more detail.
The quality of the methodology for an exploratory case study should be judged on the
basis of construct validity, reliability and external validity (Yin, 1994). The construct
validity refers to the measures for the concepts being studied. To avoid the problems
related to the subjectivity of data, multiple sources of evidence were being used
(triangulation purposes). Additional and contextual information was found in written
documentary material, such as operational manuals, purchasing plans and websites.
Different types of informants were interviewed from a central purchasing perspective
and from a decentral purchasing perspective. The interviews were conducted by two
researchers to enhance interpretation and understanding of the gathered information.
The reliability refers to the possibility of repeating the study with the same results. To
enhance the reliability of the case study, the reader is referred to the interview guide in
Appendix A. The external validity refers to the domain to which a study’s findings can be
generalized. Obviously, this case study does not allow for any statistical generalization.
The case study aims to generate a particular set of results to some broader theory
(theoretical generalization). This theory concerns the relationship between conditions,
goals and purchasing strategies within the context of a portfolio approach. An analytical
generalization, however should be based on replications of the findings (cf Yin, 1994).
5.2 DSM
DSM is an integrated international group of companies that is active worldwide in the
field of chemicals, biotechnical products and plastics. In addition, DSM is engaged in the
exploration and the extraction of oil and natural gas. The group has annual sales of
approximately NLG 14 billion, is divided into 16 business groups that are subsequently
subdivided into business units, and employs about 23,000 people (half of them based in
the Netherlands) at more than 200 sites worldwide. DSM has a large number of
companies in Europe, the Americas, Asia and Australia. DSM’s head office is in Heerlen,
the Netherlands.
DSM’s activities are organized in business groups corresponding to the product/market
combinations. The company’s principal products are intermediates and ingredients for
the pharmaceutical and food industries, performance materials (like engineering plastics,
resins and synthetic rubbers) for the automotive and electronic industries and polymers
as well as industrial chemicals for a wide range of manufacturing industries. DSM is the
global market leader for several products, including anti-infectives, caprolactam,
melamine and EPDM rubber.
The company’s strategic focus is on those businesses in which it already has leading
positions internationally or has the capability to secure such positions. Its top priority
growth area is the supply of the life science industries. DSM is ‘a sizeable customer’ with
a EUR 4.95 billion purchase spend which corresponds with 70% of total turnover. The
main purchasing categories are:
- chemicals and raw materials (EUR 2.41 billion),
- facility good and services (EUR 0.36 billion),
- information and communication technology (EUR 0.23 billion),
- physical distribution (EUR 0.82 billion), and
- technical goods and services (EUR 1.13 billion).
5.2.1 Organization of purchasing
DSM is a diversified company with a strong decentralized structure. To a large extent,

the ‘empowered’ business groups/units are autonomous, although they obviously have to
give account to the Managing Board of Directors. Business groups and business units
have their own, decentralized purchase departments and employ their purchasing
professionals. However, some years ago it was felt that DSM as a company did not fully
exploit the possibilities of coordinated sourcing (‘waste of diversity’). Centralized
purchasing was not an option, because that would be in contradiction with the principles
of empowered groups/units. The solution was found in a service unit: DMS Purchasing
Services, which was set up in 1994. An important objective was the achievement of
purchasing synergy and leverage, across business groups/units. Purchasing Services
employs about 130 people, 100 of which are purchasers.
DSM Purchasing Services operates as a business-owned cooperation with the task of
generating optimum purchasing values for its owners, the business groups and the
business units. This is accomplished in close concert and cooperation with the BG’s/BU’s.
DSM Purchasing Services is supervised by the BG’s Policy Board, a cooperation of the 16
business groups. The BG’s Policy Board are able to enforce synergy. For instance, if there
are differences of opinion with respect to the purchasing strategy of a raw material, the
policy board can arrange settlements for the parties involved. There might be a conflict of
interests with respect to the joint sourcing strategy of a raw material, for example. For
business group A it might be a non-critical product, while the item is of strategic
importance to business groups B and C. If the joint purchasing strategy is at the loss of
group A, then this business group is compensated, through the agency of the policy
board.
The main objectives of DSM Purchasing Services are:
- to function as a center of purchasing excellence,
- to structure company-wide purchase coordination systems,
- to carry out purchasing and related activities for the BG/BU’s and
- to act as functional leader in the sourcing field (development of tools, competences and
methodologies) on behalf of the BG/BU’s.
150
Decentralized purchasing departments work closely together with DSM Purchasing

Services. There is an on-line system available for the purpose of purchasing planning that
provides a DSM-case bank of ‘best practices’ for similar situations. Purchasing Services
attaches great importance to building and sharing of purchasing know-how, experience
and expertise.
Obviously there is an area of tension between DSM Purchasing Services and the Business
groups/units. The purchase marketing plans are prepared by DSM Purchasing Services.
However, all plans have to be approved by the business groups/units involved. DSM
Purchasing Services has no kind of formal power. Purchasing Services is not a central
purchasing department. The cooperation with the Business units/groups is on a voluntary
basis. Plans and proposals of Purchasing Services can always be overruled by business
group/unit managers in case of disagreement. The department has to rely on expert and
convincing power.
Purchasing has changed over time from a production oriented scope to a process and
market oriented scope. Since 1996 DSM Purchasing Services has actively promoted the
professional development of the purchasing function. An important tool is ‘management
development’. High potential sales and marketing specialists were selected for
purchasing positions. In 2000 Purchasing Services had about 100 professional employees
(management developers, academics and BG Hirees). The role of purchasing has
developed from a traditionally supporting role to production planning into an important
business function interrelated and contributory to other key business processes. In terms
of the maturity of the purchase function, DSM Purchasing Services claims a strong
organizational development from a relatively ‘laggard’ role in the early nineties to that of
an ‘innovator’ in the new millennium. A top priority for Purchasing Services is the large
scale implementation of e-procurement as a strategic response to the new purchasing
opportunities of the Internet.
The “Concern Strategy Dialogue” of 1998 has identified manufacturing, selling and
sourcing as the three primary processes. This indicates that purchasing and supply
management are considered to be very important for DSM. As a general rule, it is
compulsory that every business plan has to deal with purchasing and supply
management. The quarterly purchasing reports on behalf of the Board of Directors are
another illustration of the agreed importance of purchasing.
5.2.2 Purchasing and supply strategies
DSM Purchasing Services promotes a company-wide purchasing practice, aimed more at

longer-term ‘value’ than at short-term oriented ‘savings’. Optimal purchasing value is
harvested in a sequential fashion leading through the areas of leverage, system and design.
A basic principle in DSM’s purchasing practice holds that price is always important. Not
all purchases are eligible for value optimization. In such cases a strategy of maximum
leverage and pricing pressure is preferred. However, the greater part of DSM’s purchases
is believed to be eligible for the value option. In this option optimal leverage is exerted to
the supply base and a selection of suppliers takes place. As pricing pressure is exerted,
these suppliers will look for other opportunities to compete beyond the mere price i.e. by
efforts to enter the value chain of DSM, moving from price competition to value added
competition. The idea is to provoke a change of perspective for suppliers by means of
pricing pressure as a necessary point of departure.
Selection criteria are suppliers’ capabilities to transgress from the traditional suppliers
role to that of partners in system- and design cooperation. Obviously there is and there
should be a strong connection between the purchasing strategy and the business strategy.
For a good understanding of DMS’s purchasing practice a reference should be made to
the main product/market combinations. The company’s principal products are
intermediates and ingredients for the pharmaceutical and food industries, performance
materials for the automotive and electronic industries and polymers as well as industrial
chemicals for a wide range of manufacturing industries. DSM operates in a number of
global markets where price and cost are always key success factors. A very basic principle
in DSM’s purchasing practice holds that price performance is always important, also in
strategic buyer/supplier relationships.
In the yearly Concern Strategy Dialogue decisions are made on the agreed strategic options
for the various business groups. For the chosen business strategies accompanying
decisions are made on, what are considered, the key success factors for each business
group/unit. Any purchase marketing plan is aimed at a fit between the supply market
conditions and these key success factors. Availability is critical for a large number of
materials and commodities. In addition, DSM faces strong price fluctuations for these
products in world supply markets. To make things more complicated, increases in prices
do not necessarily correspond to the prices DSM is able to charge for end products. These
are important factors, affecting profitability and continuity. In essence, for these materials
and commodities, DMS’s purchasing focuses on the matching of purchase prices and
selling prices, resulting in a reasonable profit margin.
To a manufacturer in the process industry, disruptions in production are disastrous.
Safety and environment are important too. Other key factors are the assurance of supply
and the buffing of price fluctuations. Every purchase marketing plan aims at identifying
and exploiting the link between supply markets and business group-specific key success
factors. These plans are all developed on the basis of the purchasing portfolio approach.
Connecting purchasing strategy to business strategy implies an understanding of:

- the supply industry
- the strategic intent of suppliers
- DMS’s position versus the supply industry.
152
The development of any purchasing (marketing) plan requires supply industry analysis,
supplier cost (price) analysis, internal analysis, product classification (Kraljic),
assessment, objectives and strategies, and a purchase action plan. The Porter framework
is used to describe and analyze the way an industry behaves. The supply industry
analysis is performed from a purchasing perspective. For the classification of products
DSM uses a Kraljic purchasing approach. Next, the performance and attractiveness of
(potential) suppliers are assessed. How do they meet DSM’s needs? An important
question is: what is the position of DSM as a customer? How attractive is DSM (as a
customer) within a particular segment? Answers are given from the point of view of the
marketing manager of one of the suppliers. This approach is labeled as ‘mirror image
thinking’. In accordance with Kraljic’s approach (second matrix) the supplier strength
and the DSM strength are weighted to assess the balance of (bargaining) power.
On the basis of a thorough assessment it is decided whether a supplier is qualified for a
partnership. Such assessment implies the identification of:
- the performance criteria (qualifiers),
- the ranking of potential suppliers on these criteria, and
- the suppliers with discriminating scores on important criteria (differentiators).
A sample list of performance criteria is: cost (and price) competitiveness, technical
support and developments, security of supply, commitment, capability to produce,
safety, and information exchange. Purchasing strategy and purchasing objectives have to
refer to the identified key buying-factors, for example, security of supply, low system cost,
and access to technological developments. A real strategic partnership should always
involve design optimization. Technology based partnerships are very valuable to DSM.
However, DSM Purchasing Services is critical of so-called ‘strategic partnerships’:
“Successful partnerships are rare”. Partnerships involve much time, commitment and
investments. In some cases there are locked-in situations where DSM is forced into a
‘partnership’. These ‘partnerships’ do not involve much mutual commitment. The
capabilities and performance of the supplier do not match those that are expected by
DSM. Obviously, there is much incentive to change these unfavorable conditions.
Finally, based on objectives, strategic options and choices are identified and evaluated.
The procedures and basic steps for developing purchasing strategies and purchasing
(action) plans, are described in the internal Guide to Purchasing Marketing Planning. From
1996 through 1999 approximately 50 purchase marketing plans were developed, 35 of
which exceeded the boundaries of individual business units. Half of the plans were made
for the category ‘chemicals and raw materials’ which corresponds with the share of this
product category in the total purchase spend. The portfolio approach is believed to be
applicable to all product categories that present a substantial value potential (earning
potential).
5.2.3 Purchasing portfolio analysis: dimensions, measurement and use
For the purchasing portfolio analysis a Kraljic-like approach is being used with
comparable dimensions for the classification of products:
1. the strategic importance of purchase
2. the supply risk.
The importance of a purchase is assessed in terms of:
- value added by product line,
- the percentage of purchased products as part of total cost and
- the impact on the company’s profitability.
The ‘importance’-dimension is therefore quite similar to Kraljic’s original ‘profit impact’.
Supply risk is measured by supply scarcity factors, such as
- state of the art technology,
- complexity of the supply market,
- materials substitution,
- barriers to entry,
- logistics,
- monopoly/oligopoly conditions.
The supply risk-dimension is adapted to the specific circumstances of DSM, dealing with
technical, complex products and supply markets.
Examples of the different product items purchased by DSM are:
- bottleneck items: peroxides, catalysts, additives, enzymes and chartering;
- non-critical items: process aids, voice and data, containers and office supplies;
- leverage items: IT, travel, temporary labor and polypropylene;
- strategic items: complex IT projects, glass fiber, gas/naphtha and aromatics.
The purchasing portfolio analysis is viewed as an important tool for developing

purchasing strategies. Purchasing planning is a continuous activity, taking results and
changing circumstances into account. Immediate causes for portfolio analysis are to be
found in acute changing market conditions, particularly in changing conditions in
marketing and sales (market driven).
Purchasing portfolio analysis is not considered as the absolute truth in purchasing
strategy. It is a tool that aims at starting and guiding strategic discussions with business
group management. There are no calculating rules to decide whether the importance of a
purchase is ‘high’ or ‘low’. The same holds true for the complexity of supply markets.
The philosophy of DSM Purchasing Services is that the Kraljic matrix should serve as a
framework for an in-depth discussion with representatives from the business units
involved. Preferably cross functional teams should decide and substantiate their points of
view with respect to the position of purchases in the matrix. The drawback of this method
is that the validation of measures is not possible. However, there is no belief in a
quantitative approach for measuring values of the dimensions: “It is better to be roughly
154
right, than exactly wrong.” This does not imply that the use of the matrix is a complete
subjective matter. Points of view have to be substantiated by facts and figures with
respect to underlying factors and variables.
Actual purchasing practices and strategies are compared with the strategic
recommendations: strategic items require partnerships, leverage items are
interchangeable, non-critical items require efficient buying and, in case of bottleneck,
security of supply is recommended. Differences between actual and recommended
practices are discussed. Views and plans are ‘challenged’, which functions as an internal
warranty of quality. The main purpose of the portfolio analysis is that it forces
management into a critical evaluation of supply markets, suppliers, purchase practices,
and the relationship between purchasing and business strategy. Long-term savings are
expected from the use of a portfolio approach, albeit that the ‘real’ savings should be
attributed to development and the implementation of purchasing and supply strategies.
In the last four years an estimated sum of EUR 113.45 million has been accounted to a
more integrated purchasing approach. The portfolio analysis is an important tool,
especially for discussing, visualizing and illustrating the possibilities of professional
purchasing and supply management. In the course of time, the purchasing portfolio
approach has contributed to the awareness of purchasing possibilities and the
professionalizing of purchasing within the business units/groups.
5.2.4 Portfolio-based strategies
For the choice of (differentiated) purchasing strategies a combination is developed

between the Kraljic portfolio analysis and a sequential approach, identifying possibilities
from price leverage to system optimization and ultimately, design optimization. In some
cases price and efficiency are the only possibilities for purchasing leverage. In other,
more preferable cases more interesting forms of leverage are possible. The next level
implies system optimization, entering suppliers in DSM’s supply chain. Suppliers become
supply chain partners when they co-ordinate their work with DSM’s production
requirements. These supply chain partners are labeled as ‘partners of convenience’ to
distinguish them from ‘real’ strategic partners who actually contribute to DSM’s design
optimization. This constitutes the final level of the sequential approach. The combination
of the Kraljic portfolio analysis and this sequential approach is very important in
understanding DSM’s strategic purchasing practice and the movements in the Kraljic
matrix.
The general strategic recommendations, as provided by Kraljic, are elaborated in view of

the specific circumstances and conditions at hand. A main principle of DSM is that the
non-critical and the bottleneck cells should be as empty as possible. Obviously leverage
and strategic items are preferred to non-critical and bottleneck items. DSM is always
looking for possibilities to move to other, better positions in the matrix. However, moving
to a more preferable position in the matrix is not always within the bounds of possibility.
In those cases the cell position remains the same.
The in-depth interviews identified the most common strategic switches from one category
to another. The most common strategic changes were identified and in each case the four
generic positions in the (first) Kraljic-matrix were used as points of departure. In a
general sense, the pursued movements in the matrix usually show a clockwise pattern.
From bottleneck to non-critical, from non-critical to leverage, from leverage to strategic.
We have found conditions that lead to the choice of the different purchasing strategies
and the new positions in the matrix.
(a) Conditions and strategies for bottleneck items

For the category ‘bottleneck items’ different purchasing strategies are pursued,
corresponding to different conditions (see figure 5.1). In general terms, DMS aims at
increasing buying power and/or developing new opportunities for bottleneck items,
reducing the dependence on a supplier to an acceptable level.
A very important issue concerns the question whether standardization of purchasing
requirements is possible or not. If standardization is not possible, then in case of
processed materials a capacity deal (1) is explored, concentrating purchases to an
approved supplier. A better deal is made by concentrating regular supply to one supplier
(‘outsourcing’), involving no core competence. Very often, it is a matter of ‘make-or-buy’
certain processed materials. A ‘better’ bottleneck-position is pursued by means of a
capacity deal, reducing supply risk on the one hand and increasing profit impact on the
other hand. Other possibilities in the bottleneck quadrant are: keeping stocks, Internet
buying, widening of specifications, getting access to alternative sources, adapting
technology, risk analysis in combination with contingency planning. The motto is: “stay
in the corner and make the best of it”.
More rigorous is the switch from ‘bottleneck’ to ‘leverage’. Especially MRO-items are
eligible for such a drastic move. Business units/groups have to agree to standardization
and pooling (2) of their purchasing requirements.
Some purchased products are bottleneck items, due to a degree of over specification. In a
technical environment there is a natural drive for over specification, because technical
specialists tend to settle only for ‘the best’. Obviously this results in financial non
attractive deals. A related problem is the incompatibility of equipment and MRO-items,
due to the fact that business units/groups work with their own specifications. This
prevents the use of buying leverage by standardization and pooling of requirements. In
these cases DSM Purchasing Services sets up a team of experts to investigate possibilities
of standardization, following the principle of ‘fit-for-use’. The idea is to make the end-
product less complex: decomplex (3a). What specifications are really necessary to fit the
needs of the business units/groups involved? “Delete the waste of diversity” serves as the
156
leading device. The team chooses the best fitting standards, making the specifications
more generic. This allows for pooling (3b) of requirements across units/groups. There are
more purchasing and supply possibilities in case of a ‘decomplexed’ product and,
obviously, by pooling purchases the buying power is considerably increased. In special
cases, DSM considers joint buying with other companies, for instance in a consortium
structure. To conclude, in a two-step process, buying leverage is established, provoking a
switch from ‘bottleneck’ to ‘leverage’ in the portfolio matrix. Leenders and Fearon (1993)
came to the same conclusion, when they describe ‘standardization’ and ‘simplification’ as
two of the most effective procurement concepts for improving value.
leverage strategic
(2)
(3b)
(1)
(3a)
non-critical bottleneck
current condition 1 condition 2: main products purchasing new position

position standardization? overspecification? strategy (goals)
bottleneck not possible no processed capacity 'better

materials deal (1) bottleneck'
bottleneck possible no MRO-items pooling (2) leverage
bottleneck possible yes equipment, decomplex (3a) non-critical,

MRO-items and pooling (3b) leverage
FIGURE 5.1 Conditions, purchasing strategies and goals for bottleneck items
(b) Conditions and strategies for non-critical items

The main products in the non-critical category are office supplies and services. The
question with these non-production orientated purchases is whether standardization and
pooling are valid options or not. The product category ‘travelling’ is an example for
which pooling (1) is a logical option. A framework agreement (master contract) with a
preferred supplier is a contractual possibility. These arrangements are nowadays
replaced by some form of e-procurement. Almost the entire purchasing procedure can be
completed by some sort of catalogue buying. E-procurement is obviously the most
modern and efficient version of catalogue buying. E-procurement is only feasible when it
is possible to standardize and pool the purchasing requirements, preferably those of
several (if not all) business groups/units. However, sometimes pooling is not an option,
because the product is in some respect unique or business units/groups make a
reasonable case for not pooling their purchases. In that case, the product category
remains ‘non-critical’. Non-critical products are purchased on a transactional basis
(market exchange, non-relational elements). The purchase card (2) is a useful tool for
individual non-important orders.
In a technical environment there is a natural tendency for wanting ‘the best there is’.
Engineers usually prefer a technical tour de force, neglecting advantages of pooling and
standardization. In terms of the Kraljic matrix, from time to time there are unwanted
counter movements from ‘non-critical’ to ‘bottleneck’ positions. Over-specification
condemns a business unit/group to a specific supplier, accompanied by high cost and a
high level of dependence. Caused by the organizational structure and the position of
DSM Purchasing Services these counter-movements are not always avoidable. Business
units and business groups have the last word, although an appeal is possible to the BG’s
Policy Board. A refusal to cooperate can be overruled, when the company interest
exceeds the interest of an individual business group.
158
leverage strategic
(1)
(2)
current position condition 1 main products purchasing strategy new position (goals)
standardization and
pooling ?
Non-critical possible office supplies, pooling (1) leverage

services
Non-critical not possible office supplies, purchasing card (2) ‘better’ non-critical
services
FIGURE 5.2 Conditions, purchasing strategies and goals for non-critical items
(c) Conditions and strategies for leverage items

DSM clearly distinguishes ‘strategic partnerships’ and ‘partnerships of convenience’ (see
figure 5.3). The assessment of the supplier should indicate what kind of partnership is
possible and desirable. The assessment implies the identification of the key buying
criteria and the performance of the supplier on these criteria. A switch from the leverage
to the strategic position in the matrix might be sensible when the supplier has the proper
capabilities for co-design, in view of the qualifying and differentiating factors
(performance criteria). The move from ‘leverage’ to ‘strategic’ is only feasible in case of a
limited number of suppliers with the required capabilities and intentions. The switch
from the leverage to the strategic category might imply supplier development. There
should be a reasonable amount of trust in the dependability of the supplier and in the
supplier having the right attitude and intention of being a real partner. Only then, a
strategic partnerships (1) is considered.
On the other hand, if the supplier does not qualify for being a strategic partner, DSM
focuses on efficiency and cost reductions. Leverage is sought in efficiency and supply
chain optimization, not in design optimization. A partnership of convenience (2) is not
considered as a ‘strategic partnership’, but as an operational solution for a practical
problem. If a ‘partner of convenience’ contributes to the key success factors of one or
more business groups, then the relationship with this supplier obviously is very
important to DSM. Partnerships can be technology driven (joint venture, co-development,
concurrent engineering) or driven by logistics (JIT management). DSM Purchasing
Services considers a logistic based relationship as a partnership of convenience, because it
does not use the advantages of design optimization. In those cases the product category
remains the same: leverage.
leverage strategic
(2)
(1)
current position condition 1 condition 2 purchasing strategy new position (goals)

capabilities for price performance?
co-design?
leverage yes satisfactory strategic partnership (1) strategic
leverage no/n.a. satisfactory or Partnership of leverage

not satisfactory convenience (2)
FIGURE 5.3 Conditions, purchasing strategies and goals for leverage items
160
(d) Conditions and strategies for strategic items

Successful strategic partnerships are rare. In the course of time partnerships may become
unsatisfactory. A position in the ‘strategic’ category means a high mutual dependence. In
some cases the firm is locked in a partnership, from sheer necessity. For instance, because
of an oligopolistic or monopolistic market situation. The development of new suppliers
would solve this locked-in situation. Another non-desirable possibility is that the
supplier does not really want to be involved in co-development. As a result, design
optimization is not achieved. There is always a chance that a partnership evolves into an
indolent, relaxed relationship. Strategic partners should be world-class suppliers, alert
and high performing, not only in a technical but also in an economical sense. This means
that strategic partners should meet external benchmarks with a more than satisfactory
price performance.
In case of these not-optimal ‘partnerships’ with underachievement with respect to the
performance in co-design, a strategy of decomplex and supplier development (1) is pursued.
By making the product less complex, alternative solutions get within reach. If necessary,
new suppliers are developed. Essentially, DSM uses the natural drive to make itself less
dependent on the non-dependable supplier.
leverage strategic
(2)
(1)
current position condition 1 Condition 2 purchasing strategy new position (goals)

price performance? performance in
co-design?
strategic not satisfactory not satisfactory decomplex and supplier leverage

development (1)
strategic satisfactory satisfactory strategic partnership (2) strategic
FIGURE 5.4 Conditions, purchasing strategies and goals for strategic items
Sometimes a DSM business group/unit does not behave like a ‘partner’. Based on the
importance of the purchase for the supplier, an actual strategy of ‘leverage’ and the use of
bargaining power is actually followed. This mismatch, detected by the portfolio analysis,
causes unnecessary problems in buyer-supplier relationships.
5.2.5 Conclusions
DSM is an divisionalized industrial company that is experienced with the use of the
Kraljic portfolio approach. The main objective of the DSM-case study was to identify and
describe an advanced practice in the use of a purchasing portfolio approach. The case
study was conducted at the corporate level, in other words, across the business units. The
study has provided some new perspectives on the Kraljic portfolio approach, dealing
with two major problem areas: (1) measurement issues, and (2) the nature of the strategic
recommendations.
Kraljic’s approach does not provide measurement rules for the assessment of positions in
the matrix. The users have to decide on important measurement issues. Some feel that all
classifications are highly subjective and therefore arbitrary by nature. At DSM however,
this drawback is considered as an attractive benefit. Managers are ‘forced’ to participate
in an open, strategic thinking process, where the portfolio model provides a useful
framework for communication and discussion. The Kraljic approach allows for the
needed customization to this matter.
The in-depth interviews also provided answers to research questions with respect to the
portfolio-based purchasing strategies, their content, goals and conditions. Based on the
results of the case study, we have identified and described:
- the main portfolio-based movements within the matrix (goals and strategies);
- the specific conditions for these different purchasing goals and strategies.
These findings may serve as an illustration of the possibilities of an advanced use of
Kraljic’s portfolio model (Gelderman and Van Weele, 2002b). Kraljic’s strategic
recommendations are elaborated to more practical normative guidelines, based on
prevailing circumstances (conditions) and related goals. It is concluded that the Kraljic
162
approach provides a practical framework, allowing for sufficient customizing and

elaborating purchasing strategies. Kraljic’s purchasing portfolio approach seems to be
very useful for (industrial) companies that are divided into a large number of business
units. It facilitates the development and implementation of specific portfolio-based
purchasing strategies aimed at leverage and synergy across business groups/units. For
DSM, the Kraljic portfolio approach has proved to be a powerful tool for:
- discussing, visualizing and illustrating the possibilities of the development of
differentiated purchasing strategies, and
- coordinating the purchasing strategies of the various business groups/units.
5.3 Akzo Nobel Coatings

Akzo Nobel is made up of three business areas: Pharma, Chemicals and Coatings. This
case-study focuses on Decorative Coatings, an important business unit of the business
area Akzo Nobel Coatings. In more than 30 countries comparable portfolio analyses are
performed for the different sub-business units (area business units). These national
organizations understand their own local markets almost anywhere in the world, which
guarantees expert service close at hand. Akzo Nobel Coatings is among the world leaders
in the development of advanced new coatings. Production is provided by 130 plants, all
over the world. Akzo Nobel Coatings has a registered sales of EUR 5.6 billion, Decorative
Coatings accounts for EUR 1.8 billion in 1999, which corresponds to a 32% share of total
sales in Coatings. The most important product category is raw materials, the vital
ingredients of coatings. The main ingredients are binders, pigments, extenders, additives
and solvents. In financial terms, spendings on raw materials constitute a substantial
share of total sales. Other categories are more or less non-recurring investments and all
kinds of services and supplies. The central purchasing department is responsible for the
procurement of non-production related products. This case study is restricted to the
procurement of raw materials for Decorative Coatings, because this business unit is
experienced in the use of a purchasing portfolio approach.
Akzo Nobel is a very decentralized company that operates on a world wide scale. For the
procurement of raw materials, Akzo Nobel Coatings faces the challenge of finding
balance between global contracting and local opportunities. For certain components the
world market is very concentrated: 5 or 6 suppliers sell and produce 80% of the total
world volume.
For the buying of raw materials three buying systems are being used:
- lead buying (20%),
- main buying (60%), and
- local buying (20%).
Certain raw materials are needed in different plants, all over the world, and can be
delivered by local suppliers. For all business units within Coatings, a lead buying system
is being utilized, in pursuit of savings and synergy. A lead buyer has the responsibility to
develop and implement the overall purchasing strategy for a certain raw material. The
lead buyer draws up the central contract, negotiates prices and has control over volumes
that are bought from different local suppliers. Users in sub-business units can be asked to
switch to another supplier. The lead buyer needs to prove that the best purchasing
strategy is chosen. The main buying systems operates on the business unit level. A main
buyer is responsible for the procurement of a product (group), within a business unit. A
business unit can appoint its own main buyer who cooperates with the main buyer(s) of
other business unit(s). The system of lead buyers and main buyers demands the support
of a computer system that records all purchasing requirements of all business units. For
other product categories the purchasing responsibility is assumed by local plant units.
Local buyers deal with local suppliers. The computer system also supports local buyers,
by giving access to purchasing information with respect to all commodities bought within
Coatings.
An important starting point for the purchasing and supply strategy with respect to raw
materials, is that suppliers should guarantee low cost. Akzo Nobel Coatings does not
demand the lowest prices, but prices that are lower than the ones that are paid by their
competitors. Akzo Nobel Coatings operates from a ‘lower’ and ‘later’- principle:
- Akzo Nobel Coatings wants prices that are lower than the prices paid by competitors,
and
- in case of price-increase Akzo Nobel Coatings wants to endure that rise at a later point
in time.
Another point of interest is the dependence on suppliers (buyer’s dependence). The
business unit is feeling hesitant about being dependent on suppliers. ‘Dependence costs
money’, is the general conviction. Strategic partnerships are rarely an option. As a buyer
of ingredients, it is felt that they are by definition too small to be engaged in strategic
partnerships. The business is in this respect not comparable to the automotive where
strategic alliances with suppliers are more common.
Purchasing strategies are based on the results and the conclusions of the portfolio
analysis. For the development of purchasing strategies it is very important to include
marketing positions and business strategies of Akzo Nobel Coatings as a supplier of
goods. It is very important to create a logical fit between purchasing strategies and
marketing strategies. As a supplier, Akzo Nobel Coatings faces basically two possible
market situations, either a niche market or a commodity market.
164
In a commodity market, Akzo Nobel Coatings has to deal with low margins and large
quantities in aggressive, competitive markets. Specifications are general, resulting in
flexibility in switching from one supplier to another. Contracts are on a short term basis,
price negotiations are tough, and the logistic demands on the suppliers are high. The
same holds for Akzo Nobel Coatings as a manufacturer and seller of products.
In a niche market, Akzo Nobel Coatings operates with relatively high margins. The
delivery times and high quality are important selling points. As a result, these are
important criteria for suppliers too. High product quality in end markets require high
quality ingredients. In return, high margins in end markets allow for expensive raw
materials. Akzo Nobel Coatings is engaged in close relationships with (preferred)
suppliers. Switching cost are relatively high. Purchasing’s job is to maintain the required
quality. In collaboration with suppliers, considerable savings can be gained. The R&D
department will be involved in product improvement and will be guarding the
distinguishing position in comparison with competing manufacturers. If two products are
located in the same quadrant of the purchasing portfolio matrix, it is not concluded
beforehand that the same purchasing strategy is advised. It all depends on the situation
in the corresponding end markets: is it a niche market or a commodity market?
The selection of suppliers should be based on portfolio analysis. Crucial is the question:
“What is the added value of this supplier to our company?” The criteria for the most
important suppliers are set. These preferred suppliers should perform in the areas of
product quality, reliability of delivery, price, technical capabilities, and general
management. In return, Akzo Nobel Coatings enters commission agreements, based on
quantity rebates. Preferred suppliers should have production facilities in several
countries, near Akzo’s plants. Only the suppliers that meet all criteria are preferred
suppliers. It is very important to know what criteria to use (selection) and how to
measure them (operationalization).
For every plant portfolio analyses are performed on a yearly basis. Targets can be
connected to product groups within and across the quadrants of the portfolio matrix. For
instance:
- a certain product should be moved from the strategic quadrant to the leverage
quadrant;
- the number of items in the right quadrants (strategic and bottleneck) should be reduced
by 5%;
- the value of all leverage and non-critical purchases should be at a minimum of 65%.
Akzo Nobel Coatings works with price indices for raw materials. Every year purchasing
plans are developed, including specific goals for specific product categories. Targets and
goals are formulated in terms of these indices. A very important benchmark concerns
prices that are being paid by competitors, although it is very difficult to get a hold on that
information.
The matrices of the different area/country business units are not combined to one joint
purchasing portfolio matrix. An ingredient of coating A might easily be replaced by
another, while the same ingredient in coating B can not be replaced by any other
ingredient. There is a diversity of significance of the same ingredient for different
coatings. The portfolio matrix is completed on the level of individual plants. Given the
fact that local situations are incomparable with respect to the chemical composition of
coatings, portfolio matrices can not be joined. Coordinated sourcing is organized by the
lead buying system and the main buying system.
The portfolio analysis is considered as a very important tool for the development of
purchasing strategies, differentiated to products and suppliers. The portfolio analysis is
being used to indicate the importance of a raw material and its suppliers, and to order
the purchasing value. This results in a clear picture of the own strengths and weaknesses
in purchasing markets.
The main purpose of the portfolio approach is to detect products or productgroups that
cause problems and risks of dependence: bottleneck and strategic items. Considering the
vast number of items that are being bought, it is imperative to use a portfolio-tool.
Otherwise, it would be impossible to gain a clear insight into the problems and
possibilities of the product portfolio.
The results of the portfolio analysis points at the problems and products that need to be
tackled, and to what priority. It focuses on the goals and directions of purchasing
strategies, and the efforts of R&D-departments in their search for alternative solutions.
In addition, the purchasing portfolio provides valuable insights in the balance of power.
It is of critical importance to recognize and formulate questions with respect to
negotiation possibilities. Which party dominates the relationship? Is there a problem,
facing a dominant supplier? If so, what is the problem? Do we want to deal with one or
more suppliers? What would be the advantage of being a dominant party? What goals
would be in reach? Obviously, there is the question “what are the possibilities of
purchasing for influencing the balance of power?”
Akzo Nobel Coatings uses a customized version of the Kraljic portfolio approach. All raw
materials are categorized into four cells, based on:
- the number of suppliers, and
- the value of purchases.
To be precise, the number of suppliers is defined as “the number of suppliers that are
actually used in the last year for the same item”. There is an important difference with
166
the size of the supply base, the potential number of suppliers, which is per definition
larger than the number of actual suppliers. A scale is used that runs from ‘large’ to
‘small’. More specific, the demarcation line between ‘large’ and ‘small’ is drawn by
assessing the dependence on the supplier at hand. In general, the number of one or two
suppliers is considered to be ‘small’. Apart from that, a larger number of suppliers could
create dependence too, in case of mutual agreement and collusion. The number of
suppliers is seen as an operationalization of the original Kraljic-dimension ‘supply risk’.
The value of purchases is measured in money, reflecting the price and the volume (use)
of a raw material. The demarcation line between ‘high’ and ‘low’ is based on a 80-20 rule.
This means that the upper half of the matrix contains all purchases that add up to 80% of
the total purchase value, while the lower half of the matrix holds the remaining 20%. Any
portfolio is to be used from the perspective of the individual users. The implication is that
the demarcation line is drawn from the user’s perspective. The value of purchases is a
relative concept, to be considered from the individual perspective of the local plant
concerned. The reason is that the portfolio matrix is and should be relevant only to the
users. This means for instance that the procurement of 5,000 tons for a small plant A
might be positioned as a ‘high value of purchases’, while 30,000 tons of the same
commodity for a larger plant B is to be seen as a ‘low value of purchases’. Otherwise,
plant A would only have positions in the lower regions of the matrix.
The completion of the matrix can not be carried out by the purchasing department. This
should be done in close concert with the technical and chemical experts (R&D). In
addition, users have information regarding annual use figures, whereas the financial
management could provide information regarding the total value. It is critical that the
portfolio analysis is understood, accepted and supported by all employees of the plant.
Figure 5.5 shows an example of a portfolio matrix, as might be found in a local

organization of Decorative Coatings. What can we learn and conclude from the
information in the portfolio matrix?
- Apparently, the strategic category contains 60% of the value of all purchases. This can
be interpreted as a very high dependence on suppliers, which can only be justified in case
of niche markets for the end products. In a niche markets, products are characterized by
much added value, a high quality and price level, and a drive for new products. For
commodity markets, there would be a misfit of the actual and expected segmentation in
the portfolio matrix. In case of end products for commodity markets, the strategic
quadrant would have to be much smaller.
- In addition, the figure shows a large number of items with a relatively low value. This
implies much administrative work for purchasers. A possible objective could be to lower
the number of bottleneck and non-critical items.
value of purchases
high
leverage strategic
no. of items: 15 no. of items: 25

percentage: 20% percentage: 60%
no. of items: 40 no. of items: 240

percentage: 5% percentage: 15%
Low
large small
number of suppliers
FIGURE 5.5 Example of a raw materials-portfolio
Based on the situations and conditions on the marketing and sales side, purchasing
strategies are focussed on handling costs and strategic vulnerability (dependence on
suppliers). Targets are determined for each product category in each quadrant, dealing
with these issues. Purchasing strategies in general are aimed at adapting and improving
conditions, not so much at changing positions in the portfolio matrix.
However, sometimes movements are possible and desirable in the matrix. The main
movement in the matrix is from strategic to leverage; other switches are rare. The point of
action is the number of suppliers. Sometimes it is possible to enlarge the number of
suppliers, in particular by means of an active strategy of supplier development. The value
of purchases is usually not compliant to intervention, because of the fixed prescribed
composition of coatings.
In practice, there are practically no chosen movements from the left half to the right half
of the matrix. In other words, purchasing strategies are generally not aimed at reducing
the number of suppliers. For raw materials a general rule holds that it is always better to
deal with two or three suppliers, then to deal with a single supplier. The reason is that
any supplier reduction increases dependence which lead to a vulnerability for price
raises. For reasons of flexibility, Akzo Nobel Coatings stresses the importance of
maintaining good relationships with potential suppliers that are not currently contracted.
They can provide useful information to be used in negotiation processes. Moreover, these
suppliers may provide alternative arrangements in cases of emergency or problems with
the current suppliers. Working with a limited number of suppliers is preferably
168
combined with the possibility to fall back on alternative suppliers (flexibility). It is

recognized that there is a huge difference between having a sole supplier of choice and a
sole supplier of necessity (cf. Nicholson, 1993).
There is an area of tension between purchasing and marketing departments. Product and
marketing managers are always looking for possibilities to differentiate products,
whereas purchasing managers are always looking for possibilities to simplify and
standardize products. The demands of marketing and customers limit the number of
possibilities for purchasing in their natural propensity for controlling and reducing cost.
(a) Bottleneck items

For bottleneck items there are concerns and questions with respect to the assurance of
supply. After all, there is just one available supplier for a certain ingredient. The buying
strategy is a forced single sourcing. Generally speaking, negotiating for lower and the
lowest prices is not the main focus of purchasing. Because of the company’s vulnerability,
suppliers of bottleneck items must have contingency plans and emergency stocks. In the
contract there is a clause inserted that compels the supplier to report an intended
termination of production. Otherwise, costly safety stocks would be inevitable. A search
for alternatives only takes place in exceptional cases, because the costs of testing are
several times higher than the expected results. This means that high levels of risk and
dependency have to be accepted to a certain degree.
A consignment system is a practical solution for some bottleneck items. The supplier is
responsible for the continuous availability of certain raw materials that are stocked at the
sites of Akzo Nobel Coatings. Payments are based on actual use, not based on deliveries,
which means that financial risks are taken by the supplier, not by the buyer. It is the
supplier’s responsibility to replenish the stock when and if necessary.
(b) Non-critical items

Non-critical items represent a low value of purchases. The added value is low and the
supply risk is small, because of the large number of suppliers and/or alternative
products. The strategy here is aimed at minimizing the cost of preparing and placing
purchase orders. Possible options are standardization of procedures, combining of orders
and invoices, and e-procurement. These measures reduce administrative costs and the
time-consuming handling of orders to a minimum. On another decision level,
possibilities of outsourcing are to be considered, meaning that parts of the purchasing
process might be outsourced. A possibility would be the contracting of a large
international distributor.
Another option to consider is to ask a supplier of a leverage or a strategic product to
supply a certain non-critical item as well. The same holds for bottleneck items.
(c) Leverage items

Suppliers in the leverage quadrant manufacture ingredients for which alternative
products exist, or for which alternative solutions can be found through a simple
adaptation of the method of preparation. In many cases there is an added value to the
products, for instance just-in-time delivery, consignment stocks, or the delivery in a
special format or packaging. These special features should save production costs.
Supplier selection is often based on the added value in these areas.
Obviously, purchasing is an interesting ‘partner’ for suppliers. The purchasing
department is alert, looking for suppliers that offer more added value and/or that charge
lower prices. Purchasing is continuously monitoring the supplier performance and is
taking action when a supplier deviates from the agreement. A leverage position however,
does not mean that the buyer is the dominant party. The value of purchases is relatively
high from the perspective of the buying company, not from the supplier’s perspective.
Units are usually too small to dominate even leverage relationships.
On an occasional basis, the relationship with a supplier can be transformed from leverage
to strategic. This is only an option if a partnership is expected to add to the competitive
advantages of the firm in end markets. A chosen strategy of increasing the dependence
on a supplier is limited to special circumstances, that is, if the cooperation with a supplier
will result in a new or better product, providing a competitive advantage to the business
unit. A partnership is always on a temporary basis, because after a couple of years the
innovation is diffused and the search for alternatives recommences.
(d) Strategic items

Too often, the supplier is the dominant party in the buyer-supplier relationship. In
practice it is very hard to come to an agreement on the needed requirements. In those
cases, Akzo Nobel Coatings has no choice but to accept that a supplier does not add the
required value. The supplier has a strong position when negotiating the quality, the size
of packaging, the moment of delivery, and so on.
However, exceptions do exist. For instance, if Akzo Nobel Coatings is the major account
for a supplier, then there is naturally room for negotiating a better deal. Another
possibility would be that Akzo Nobel Coatings is considered an important customer for
reasons of image and charisma.
A position in the strategic quadrant is not preferred, because of the risks and
disadvantages of being dependent on a single supplier. Sometimes the number of
suppliers can be enlarged by means of supplier development. Strategic partnerships are
rarely an option, because the business unit is too small and the risks are too high.
Strategic partnerships are only pursued if there is a competitive advantage in end
markets to be gained in a buyer-supplier relationship. These partnerships are always on a
temporary basis.
170
5.3.5 Conclusions
The Akzo Nobel Coatings case study reveals some new insights in the possibilities of a
purchasing portfolio approach at the business unit-level. It is probably a rare example of
a business unit where the portfolio technique is fully integrated in the daily practice of
purchasing and supply management. Purchasing goals and purchasing strategies are
clearly connected to the results of the different portfolio matrices. Every plant completes
a portfolio matrix in a similar way, providing an overview of the purchasing operations
at a business unit level and providing insights in local plant situations. The main purpose
of the portfolio analysis is to detect and to cope with the risk of being too dependent on
suppliers. It provides the basis of purchasing planning and the development of
differentiated purchasing strategies.
For the evaluation of a completed portfolio matrix, it is imperative to have information on

the marketing positions in the connected end markets. For instance, in case of a
commodity end market, it is not acceptable to have a filled-in strategic quadrant. This
means that there is a clear relationship between the purchasing strategy and the marketing
strategy, catalyzed by the use of the portfolio model.
On the level of individual plants, the strategic recommendations are aimed at reducing risk
and dependence on suppliers. Strategic partnerships with suppliers are rare and always
temporary. In special cases, whenever possible, it is recommended to increase the
number of suppliers, for instance by means of supplier development. In more usual cases,
the dependence on suppliers is handled by means of contingency plans and by keeping
safety stocks.
Measurement issues are dealt with in an interesting and remarkable way. The number of
suppliers and the value of purchases are selected as basic dimensions. This choice has a
main advantage over many other operationalizations of the Kraljic-dimensions. The value
of each variable is made measurable in an objective manner. The values are not measured
in terms of perceptions or other proximities of variables. Portfolio matrices are therefore
better comparable, both in time as in comparison to other plants. In addition, the
demarcation problem too is handled in a very practical way. On the basis of a set of clear
rules, it is decided in which category a product is to be placed.
5.4 TE STRAKE - Engineering & Production

The business units of TE STRAKE operate in close cooperation with customers, involved
in the manufacture of machinery for various sectors of the electronics industry, textile
machinery industry, printing industry, agricultural industry, optical media and
automotive industry.
TE STRAKE, located in the Netherlands, is a company specialized in the development

and production of serial products in the field of mechatronics motion controlled systems,
the combination of mechanical and electronic technologies. In addition, TE STRAKE is
involved in the surface treatment of metals for the automobile industry. TE STRAKE
concentrates on narrow market segments, the application of complex technologies, and a
relatively small number of customers per segment. Approximately 80% of the turnover is
generated by 15 customers. In 1999, the company enjoyed a turnover of EUR 52 million,
the operating result being EUR 4.2 million and the net income amounting to EUR 2.7
million. The share of purchases has risen to 68%, which is interpreted as a substantial
increase in flexibility. TE STRAKE has 300 employees.
TE STRAKE is made up of three business units: Engineering & Production, Weaving
Technology, and Corrosion Protection. This case study focuses on the business unit
Engineering & Production, for reasons concerning experience in the use of a purchasing
portfolio approach. With a 75% share in turnover, Engineering & Production is the largest
business unit of TE STRAKE. Engineering & Production is an applicated system supplier
in the true sense of the word, with a very strong customer focus. Engineering &
Production acts as a first tier supplier in a limited number of OEM-markets. Support is
being offered throughout the entire trajectory of product development and production.
Engineering & Production is able to take over management and realization of whole
segments of complete high complex mechatronical modules, units or machines. This is
usually in medium-sized series, from 20 to 7,000 units per year. Engineering &
Production’s expertise is called in especially when a combination of fine mechanics,
electronics and pneumatics is concerned.
Being a main supplier, Engineering & Production aims at a position of applicated systems
supplier, emphasizing that the business unit wants to be a partner in the sequence of
next-generation products and systems. Overall, TE STRAKE is focussed on a very small
number of customers. However, there is mutual dependence, because the machinery
manufacturers are on their turn very dependent on the expertise and products of TE
STRAKE. An overall objective of the company is to reduce associated risks of
dependence. This is attempted by aiming at customers in different industries who are
using the same specialization, namely mechatronic motion controlled systems. The
company positions itself as ‘your competent outsourcing partner’. The business strategy
is based on creating added value for customers, close cooperation and long term
relationships with key customers.
Engineering & Productions has a long term relationship with 4 large, major customers. In
addition there is a limited number of smaller customers with a potential development of
becoming a major customer. That is why they are labeled as ‘Potentials’. The organization
172
of purchasing is completely in line with the overall organization of the business units.
Production and purchasing are managed in customer focus teams, that are dedicated to
these major customers. A customer focus team consists of representatives of the following
business functions: marketing and sales, logistics, assembly, planning, engineering,
purchasing, and quality management.
Purchasing objectives are formulated on the customer level, specified for major
individual customers. There are purchasing and supply targets per customer in the areas
of cost and cost reduction, product quality, reliability of delivery, lead times, flexibility,
and risk management. The dominating customer focus on purchasing practice however
should not imply the negligence of synergy across customer focus teams.
Engineering & Production employs a strategic purchasing function and some senior
buyers, specialized in electronical and mechanical parts and components. Non-
production related products are procured on a central level. In terms of maturity,
purchasing is developing from an administrative function into a more pro-active function
that directly contributes to the competitive position of the firm. A fairly new computer
system supports all purchasing activities.
More and more customers demand very short lead times, presenting complex logistical
challenges for TE STRAKE. In addition, cost effectiveness is important too. Market
conditions are changing, for which supply chain management is helpful. Recently, the
concept of demand chain management has been introduced. Demand chain management
aims at supply solutions that enables the company to react quickly and in an inexpensive
way to changing demands of customers. This implies the disposition of a lean and
flexible network of suppliers.
The management of the supply base implies the development and maintenance of long
term relationships with a set of preferred suppliers. These long standing relationships
have evolved in a period of many years. Technologically advanced suppliers are treated
as strategic partners. However, there is always a field of tension between being a partner
and imposing (needed) cost reductions. Because of changing economic conditions, TE
STRAKE faces the need for continuous cost efficiency. It is difficult to press for lower
prices when dealing with strategic partners.
Sometimes TE STRAKE has to comply to the specific demands of a major customer. These
demands concern not only the specifications of products to be bought by TE STRAKE, but
also the selection of suppliers. In these circumstances TE STRAKE is obliged to enter into
forced ‘partnerships’.
An important starting point for TE STRAKE’s purchasing and supply strategy, is the
involvement of specialized suppliers in development projects. Purchasing faces the
important objective of developing and maintaining of long-term relationships with
technologically advanced suppliers. Open cost calculations and the exchange of technical
information are indispensable for these relationships that thrive on commitment and
trust.
TE STRAKE feels that it has to be engaged in different kinds of relationships, reflecting

the overall business strategy. TE STRAKE’s purchasing and supply strategies have to
support the business strategy. The most important starting points are the demands and
requirements of the major customers. Being a main supplier for knowledge intensive
industries, TE STRAKE addresses complex technological and logistical issues.
The main purpose of the portfolio approach is to gain valuable insights in the profit
impact and the supply risks, connected with the procurement of products that build the
end product for major customers. The results of the portfolio analysis are used for:
- managing the supply base;
- anticipating changes;
- visualizing problems and bottlenecks;
- assessing risks.
Supplier strategies are developed and modified, based on the results of the portfolio
analysis.
TE STRAKE uses the Kraljic portfolio approach on the level of major customers. All
products that are purchased for and used in the end product, manufactured for a major
customer, are categorized into four cells. The dimensions are:
- the profit impact, and
- the supply risk.
The profit impact of each product is determined by two factors:

1. the share of the product value in the price of the end product, and
2. the sensitivity to fluctuations of exchange rates.
The supply risk of each product is determined by four factors:

1. the level of availability (scarcity, number of alternative suppliers);
2. lead times (suppliers who need long order lead times, frustrate the required flexibility);
3. the switching costs (efforts, required to switch to another supplier);
4. the demands of the customer (sometimes customers list the second tier suppliers that
have to be contracted).
174
TE STRAKE assigns different weights to these factors, indicating their perceived

significance. The below-mentioned weights were used for the parts and components,
needed for an important end product.
profit impact
high
leverage strategic
low
low high
supply risk
FIGURE 5.6 Example of a purchasing portfolio at the level of a major customer
Additionally, the scores for the factors are determined, based on a consensus that is
reached by the purchasing professionals who are involved in the measurement process.
The scores are measured on a scale from 1 - 10. Scores are multiplied by their weights and
the weighted scores are added to a single value on the respective dimension. Obviously,
there is a level of subjectivity in the determination of factors, scores, and weights, raising
questions with respect to validity and reliability. These measurement issues are handled
by the process of reaching consensus (intersubjectivity). Moreover, positions in the matrix
are compared with the expected positions, and an assessment is made of the influence of
the measuring method on the qualifications.
Figure 5.6 shows an example of a portfolio matrix that is being made for a very important
end product, representing a large share of TE STRAKE’s turnover. The figure illustrates
the areas where the different commodities can be found:
- most items are located at the right side of an imaginary diagonal;
- there are hardly any leverage items;

- the non-critical items have a relatively low purchase value;
- the bottleneck items are concentrated in the upper left corner of the quadrant;
- the strategic items are connected with the highest supply risks.
What can we learn and conclude from the information in the portfolio matrix? For that
purpose we have to include additional background information on the specific
circumstances. In this case, the supply base for a large number of strategic and bottleneck
items has been stable for many years. Relationships with key suppliers have developed
over many years of close cooperation. Production and planning processes are
coordinated, advancements in technology are shared in a true sense of partnerships.
Long-standing business relationships result in commercial and relational bonding. These
highly committed relationships imply much dependence on suppliers and high supply
risks. A number of items in the right half of the matrix are supplied by these long-term
key suppliers (second tier). In some instances the customer exercised influence on TE
STRAKE’s relationships with key suppliers. For a number of components the customer
prescribed which (preferred) suppliers had to be contracted. This has resulted in some
forced ‘partnerships’ that can be located in the strategic quadrant and in the bottleneck
quadrant. Under these circumstances there are limited possibilities for improving
purchasing conditions. On the short term, locked-in situations have to be accepted. This
practice is quite common in many industries. Many large companies actually specify
which suppliers are to be used by their first-tier suppliers, mainly because particular
critical components have to fit with other components (cf. Johnsen et al. 2000).
In addition, the figure shows a large number of items with a relatively low value. This
implies much administrative work for purchasers. A possible objective could be to lower
the number of bottleneck and non-critical items.
Depending on the classification, a differentiated set of supplier strategies are determined.

Purchasing and supplier strategies are fairly in line with Kraljic strategic
recommendations. Bottleneck and non-critical quadrants should be as empty as possible,
but most attention is paid to the strategic items.
(a) Bottleneck items

The problem with many bottleneck items is that they can only be procured in remote
countries. If an American or Japanese supplier does not deliver on time, it is very difficult
to speed up the supplier’s production. After all, the items are of relatively low value. The
supplier might want to wait until a larger production batch can be made.
176
The usual approaches when dealing with bottleneck problems are entering into long term
contracts, keeping large stocks, and accepting consignment systems. However, sometimes
it is possible to pool the requirements and reach a better position in the matrix, in the
direction of the leverage quadrant. For instance, products that can be grouped as
‘springs’ and ‘pieces of sheet iron work’, are usually located in the non-critical quadrant
or round the borderline of the bottleneck and the non-critical quadrant. Contracts for
these products have been awarded to preferred suppliers. The objective is the assurance
of supply and the reduction of administrative work load.
(b) Non-critical items

As a rule of thumb, 80% of the buying efforts are spent on non-critical items. The focus is
always on reducing the number of suppliers, diminishing the administrative work load
and facilitating logistical processes. Whenever possible, leverage is pursued by clustering
goods and services. The recommended strategy is to cluster items into groups that are
procured at a single source. Obviously, this strategy enhances the buying power of the
firm.
(c) Leverage items

For leverage items an approach of competitive bidding is recommended. Suppliers are
interchangeable and the supply risk is low. As a rule, no long-term contracts are closed.
Buying criteria are formulated in terms of prices, delivery, and quality. However, it
appears that not in every end product leverage items are being used.
(d) Strategic items

Different strategic options are possible for the strategic items, depending on specific
conditions and circumstances. In case of problems and unsatisfactory performance of a
supplier, corrective actions are attempted. Improvements are explored, in close
cooperation with the supplier. Targets are being set or reset. But in essence, the idea of
being together in the same value chain is put forward, convincing the supplier of a
mutual interest. However, if the supplier does not succeed in improving his performance,
the relationship has to be broken off. More in general, supplier development is important
in case of poor supplier performance and in case of unacceptable supply risks.
Sometimes there is a much more adversarial relationship in case of a locked-in
‘partnership’ that is imposed by a major customer. Commitment is low and suppliers are
in the comfortable position of being irreplaceable. TE STRAKE has limited possibilities of
influencing the behavior and attitude of these suppliers. Convincing the customer of
other possibilities might be an option, but only for the long term.
Generally however, technologically advanced suppliers are very important for TE
STRAKE’s business strategy in its condition of applicated system supplier. Based on open
cost calculation, long-term relationships are pursued. These relationships are
characterized by mutual commitment. To reduce supply risks, TE STRAKE invests in
supplier relationships. On the other hand the firm pays attention to contractual issues in
order to avoid problems. These preferred key suppliers are very important, because they
contribute heavily to the success of the firm.
5.4.5 Conclusions
Being a first tier supplier in a selected number of OEM-markets, TE STRAKE’s business

unit Engineering & Production operates in close cooperation with major customers. The
necessary involvement of specialized suppliers is in line with the organizational structure
of customer focus teams. Purchasing needs to develop and maintain long-term
relationships with technologically advanced suppliers. Obviously, not all suppliers are
partners, meaning that TE STRAKE is engaged in different kinds of relationships.
TE STRAKE is aware of the consequences of its position as a first tier supplier in the
supply chain of a small number of industrial manufactures. The strong customer focus is
of decisive importance to purchasing and supply strategies. The portfolio analysis is
conducted on the level of major customers. All products that are purchased for and used
in a specific end product, are categorized into the four Kraljic-cells. The level of analysis
emphasizes the strong customer focus and the relationship with the overall business
strategy. A point of special interest is the development and creation of purchasing
synergy across customer focus teams.
The portfolio approach has proven its worth for TE STRAKE. The results of portfolio
analysis are used for the development and adjustment of differentiated purchasing and
supply strategies. In most situations the standard Kraljic recommendations are followed.
Interesting are the different kinds of partnerships with suppliers in the strategic
quadrant. In an ideal context mutually committed relationships are maintained with
innovative, reliable key suppliers who contribute heavily to the firm’s success. It can be
very difficult to pass on the high demands of the customer to these long term partners,
especially in the field of logistics (short lead times) and prices (low costs). A related
challenge is the handling of key suppliers that are prescribed by the customer. In all cases
the advantages of tight relationships with specialized suppliers imply a high level of
dependence on these suppliers. In terms of the Kraljic portfolio, much items have to be
located in the bottleneck and the strategic quadrant.
The basic Kraljic-dimensions are used: profit impact and supply risk. The values of these
dimensions are determined through an arithmetic method of weighted scores. This
includes successively the determination of factors, weights, and scores. For all
measurement issues consensus has to be reached, in pursuit of intersubjectivity and
validity. The results of the TE STRAKE-case study show that the portfolio approach is
and can be used on the level of a major customer, focussing on all the items that are
procured for the production of complex customized modules and machines.
178
5.5 Reactions to the dependence model

In section 4.4 we have developed a conceptual (portfolio) model that includes Kraljic’s
categories, and that is constructed by two new dimensions: buyer’s dependence and
supplier’s dependence. The model hypothesizes on the determinants of buyer’s
dependence and supplier’s dependence. This conceptual model is symmetrical in the
sense that the same general factors are included:
- financial magnitude;
- criticality;
- number of buyers and suppliers respectively;
- switching cost.
These explaining factors were derived on the basis of an analysis of conceptual and
empirical studies to the determinants of dependence.
In the case studies reactions were gathered on the tentative determinants of buyer’s and
supplier’s dependence. Are these the most appropriate determinants, in view of one’s
own practice? Or are there other, more relevant determinants?
Generally, the basic idea of mutual buyer/supplier dependence was well received by the
respondents. The model made sense, considering the possibilities of the interpretation of
‘dependence’ and the application of purchasing strategies. There appears to be much
support for the relationship in the model between categories and dimensions. From this
perspective, the conceptual model can be considered as an improvement of the original
Kraljic model.
Respondents were asked to react to the listed determinants in the model. Appendix C
summarizes the reactions that were gathered in the interviews. Respondents were invited
to a critical assessment of the model, providing sufficient time for reflection and reaction.
In some cases respondents needed some days to formulate their opinions.
Based on the insights derived from the three case studies, modification of the conceptual
model is necessary (see figure 5.7). Starting with the buyer’s dependence, we must conclude
that the number of alternative suppliers and the switching cost are generally perceived as
important determinants. On the other hand, financial magnitude and criticality (as such)
appear to be not or not very relevant. Market relations and competitive positions often
require the synchronization of production systems. Just-in-time delivery and the
reduction of the supply base are well known characteristics of modern business. These
circumstances cause a logistics-based dependence on suppliers. In addition, there are
high levels of technology-based dependence. Industrial firms have to rely more and more
on technologically advanced (key) suppliers. Interorganizational relationships can be an
effective means of transferring knowledge across firms (Barringer and Harrison, 2000).
Companies need the technological expertise, capabilities and resources of their suppliers.
This need may add greatly to the buyer’s dependence. In their study of characteristics of
customer-supplier relationships, Johnsen and Ford (2001) identified capability as one of
the most a prominent relationship characteristic. Therefore, the modified model includes
logistical indispensability as well as the need for the technological expertise of suppliers
as determinants of the buyer’s dependence.
Respondents report high levels of supplier’s dependence, mainly due to the financial
magnitude of transactions. Basically, the financial magnitude of purchases should be
assessed from the supplier’s position, not the buyer’s position. For the supplier’s
dependence it is important to know what the share is of a supplier’s output taken by a
particular buyer. The number of alternative buyers and the switching cost are considered
to be important determinants as well. Criticality has to be redefined as a need for the
buyer’s technological expertise. In some cases suppliers need the technological input of
the buying firm and require a transfer of know-how.
The modified model is shown in figure 5.7
financial supplier's
leverage strategic
magnitude dependence
technological
expertise non-
critical bottleneck
alternative
buyers
switching buyer's
cost dependence
logistical technological alternative switching

indispensability expertise suppliers cost
FIGURE 5.7 Modified conceptual model of buyer’s and supplier’s dependence
Based on the assumed relations between variables and dependence, it is possible to

formulate eight hypotheses.
180
1. For buyer’s dependence:
Hypothesis B1 The level of dependence on a supplier is positively related to

the logistical indispensability of the purchases.

the need of the buyer for the supplier’s technological expertise.
Hypothesis B3 The level of dependence on a supplier is negatively related to

the number of alternative suppliers.

the switching costs, incurred in case the supplier is replaced by
another
Hypotheses for supplier’s dependence:
Hypothesis B5 The level of dependence on a buyer is positively related to the

financial magnitude of the purchases.

need of the supplier for the buyer’s technological expertise.
Hypothesis B7 The level of dependence on a buyer is negatively related to the

number of alternative buyers.

switching costs, incurred in case the buyer is replaced by
another.
These hypotheses are to be tested by the data that will be gathered in the survey.
5.6 Answers to the third research question: comparison of the

case studies
In this section the findings of the case studies are further analyzed. By means of a
comparative analysis of the case studies answers to the research questions are
formulated. We recall that in section 5.1 the following questions were posed, regarding
the case studies:
3a) Considering the unclear guidelines and the unanswered questions with respect to the
measurement of (composite) dimensions and the weighting of factors in the use of a
purchasing portfolio approach, how are these issues handled to the satisfaction of
experienced purchasing professionals?
3b) What kind of specific strategies of purchasing and supply are based on Kraljic’s
portfolio matrix?
3c) What kind of movements are considered in the Kraljic matrix, in terms of current
positions, future positions (goals) and means (strategies)?
To summarize, the questions refer to issues of use, measurement and strategies. We will
start with a comparison of the differences in business context.
5.6.1 Specific business context
Obviously, there is a different business context for the three cases. Figure 5.7 summarizes
the main situational factors that describe the most notable case specific circumstances.
These factors
enlighten about the reasons behind the various ways in which the purchasing portfolio
analysis is being used.
182
DSM Akzo Nobel TE STRAKE

Coatings BU - E&P
Investigated unit/ corporate level, across business area business unit

level of analysis business units Coatings engineering & production
Value proposition operational excellence product leader customer intimacy
Supply chain position main manufacturer main manufacturer first tier supplier
Main products chemical, biotechnical decorative and mechatronical modules,

products and plastics industrial coatings units and machines
Customers large number of industrial mass markets of small number of

markets and professional users industrial customers
customers and consumers
Sales EUR 6.4 billion EUR 5.6 billion EUR 39.2 million
Organization of basically decentralized system of lead buying, customer focus teams

purchasing with a centralized main buying and
purchasing services unit local buying
Main spend groups chemicals, raw materials, raw materials electrical and mechanical
technical parts and components
products/services, and
physical distribution
Purchase spend EUR 4.9 billion raw materials: EUR 26.6 million
(78%) a substantial share of (68%)
sales
FIGURE 5.8 Context of the investigated cases
The unit of analysis in the three cases is rather different. The DSM-case investigates the
use of a purchasing portfolio approach on the corporate level, aimed at synergy and
leverage across business units. For different kinds of products and product groups the
portfolio analysis serves as a framework for strategic discussion and ultimately for
starting joint operations. As a consequence purchasing decisions are made that concern
large amounts of money. The scope of the Akzo Nobel Coatings case is a large global
business unit. In more than 30 countries comparable portfolio-analyses are performed for
the different sub-business units (area business units). The portfolio analysis concerns the
procurement of ingredients (raw materials), to be used in the end product (coatings). The
third case is performed at the business unit level as well. However, the business context
differs to a large extent. TE STRAKE is a relatively small, basically national manufacturer
of technologically advanced modules. The importance of the (limited number of)
customers is omnipresent. As a first tier supplier the portfolio analysis is performed on

the level of the customer: all products that are used for a specific customer are positioned
in the matrix. From the various scopes of the case studies it can be concluded that a
purchasing portfolio approach can be applicated on many different levels of aggregation.
The generic nature of the Kraljic approach allows for customization, implying that users
have to make all kinds of decisions, implementing the portfolio analysis. We have
concluded that measurement issues and strategic issues are handled in different ways.
Some of the differences can be explained by differences of scale, scope and organizational
level. We can better understand the content and the directions of the portfolio-based
strategies, if we take a look at differences in supply chain position and value proposition.
In terms of Treacy and Wiersema (1993) DSM aims at operational excellence, Akzo Nobel
Coatings wants to be a product leader, and TE STRAKE’s strategy is based on customer
intimacy. The conditions on end markets, the requirements of customers, and the overall
business strategy are very important circumstances for the selection of purchasing and
supplier strategies.
5.6.2 Measurement and use
In the case studies purchasing professionals were interviewed on their experience with
and views on the application of the purchasing portfolio analysis. It was clear that they
faced common problems and that they had to answer the same questions. Based on their
experience, an outline is made for the application of purchasing portfolio analysis,
covering the main decisions and choices (see Appendix D: Outline for the application of
purchasing portfolio analysis).
Practitioners who want to applicate the portfolio analysis have to answer some basic
questions of scope and design:
- what will be positioned?
- what will be the level of aggregation?
- for what (organizational) unit will the analysis be performed?
The case studies illustrate that many choices have to be made. In our cases products were
positioned in the matrix. From other sources it is known that suppliers or supplier
relationships are sometimes positioned in the matrix (for instance Van Weele, 2000;
Åhman, 2002). The portfolio analysis might be restricted to certain types of products, for
instance raw materials (Akzo Nobel Coatings), logistical services, or non-product related
products.
The level of aggregation refers to the question whether individual items, or smaller or
broader product groups will be positioned in the matrix. What is the unit of analysis? The
positioning of individual items usually implies that data are needed on thousands of
items, some of which will be very similar and many of which will be of very low value.
On the other hand, a high level of aggregation will reduce the usefulness of the analysis.
184
For instance, if all components are grouped into a single category ‘components’, then it
will not be possible to determine the level of supply risk and to select an overall
purchasing and supplier strategy. In general, the level of aggregation should be linked to
the level on which it is useful to select (differentiated) purchasing strategies. This means
that major or unique purchases should always be classified individually, while other
purchases could be classified by commodity groups or type of buy. If there are important
individual items within a commodity group with significant other values on the two
dimensions of the matrix, those should be broken up and analyzed separately (conform
Ellram, 1996).
Another question relates to the organizational unit of analysis. The investigated portfolio
approaches in the case studies were all connected to different organizational units: the
corporate level (DSM), the level of area business units (Akzo Nobel Coatings), and the
level of a major customer (TE STRAKE). Other options are conceivable. To conclude, the
purchasing portfolio analysis allows for very different modes of application. There is an
infinite number of possibilities, combining choices with respect to products/product
types, levels of aggregation, and organizational units. Choices of design are important for
the scope of the strategic recommendations and the specific portfolio-based strategie that
can be selected.
The case studies illustrate difference in use intensity, occasion, and purpose. It can be
concluded that there is always an advocate of the technique, introducing and supporting
the portfolio analysis. In the case studies, these ‘product champions’ were the highest
purchasing professionals in the organization.
Figure 5.9 summarizes some of the most significant characteristics of the investigated
portfolio approaches, examining use issues and measurement issues.
Use issues DSM Akzo Nobel Coatings TE STRAKE

BU - E&P
Frequency, occasion irregularly, in response regularly, fully integrated incidentally, on major

to changes with daily practice customer level
Main advocate and director purchasing purchasing vice strategic buyer of the
project manager services president of each BU business unit
Main purpose to identify and to to detect and to cope with to assess risk and to
develop synergy and supplier dependence identify possiblities
leverage across BU's
Measurement issues
Dimensions 1 strategic importance 1 value of purchases 1 profit impact

2 supply risk 2 number of suppliers 2 supply risk
Determination of factors during the analysis, in advance, factors in advance,

basically unlimited are dimensions limited number
Measurement of factors consensus based objective consensus based
Determination of weights implicitly, n.a. explicitly,

during the analysis in advance
Aggregation of subscores consensus based n.a. arithmetic

(additive model)
FIGURE 5.9 Characteristics of the purchasing portfolio approaches
When using the portfolio approach, it is imperative to recognize that the positioning of
items does not complete the portfolio analysis. In all case studies, the positioning of items
in the quadrants (the measurement) was followed by a process of reviewing the positions
in the matrix and a process of reflection on the consequences. Whatever method is
selected, there are always subjective choices, limitations and elements that influence the
actual positioning in the matrix. On closer consideration, questions have to be answered
for each position that is found in the matrix:
- why is an item/product positioned in this specific spot?
- is the found position in line with previous expectations?
- are positions unintentionally and wrongfully influenced by the measurement method?
- are therefore re-adjustments necessary?
In other words, after the matrix is filled, users reflect on the results. If necessary, manual
adjustments are made. In-depth discussions on the positions in the matrix are considered
186
as the most important phase of the analysis. Strategic discussions provide deeper insights
and may lead more easily to consensus-based decisions. It is felt by the users that the
Kraljic framework facilitates these important discussions to a large extent.
Additionally, a reflection on the consequences is needed. How are the positions in the
matrix to be viewed and assessed? Leading questions are:
- what is the actual meaning of the different positions in the matrix?
- what is the interpretation of the results?
- are there any points of intervention? which risks are (un)acceptable?
In section 5.1, the first research question refers to measurement problems, associated with
the use of a portfolio approach: how do experienced professionals handle issues with
respect to the measurement of dimensions and factors? We will now try to answer this
question, describing the solutions developed and used by different professionals.
The case studies identified three kinds of different approaches to the measurement issues,
connected to every purchasing portfolio analysis:
- DSM uses a ‘consensus-based’ method,
- Akzo Nobel Coatings uses a ‘one-by-one’ method,
- TE STRAKE uses a’ weighted factor score’ method.
The ‘consensus-based’ method is predominantly based on a process of reasoning and

discussing. The reaching of consensus is very important when choices are made with
respect to the measurement of variables and factors, and ultimately for the positioning of
items/products in the matrix. Advocates of this approach regard this as a very attractive
feature of the portfolio analysis that is being used. Profound open discussions about
purchasing issues are considered as the most critical part of strategy development.
Differences of opinions become very clear, allowing for a true strategic discussion. As a
rule, points of view always have to be substantiated by facts. DSM has used this method
for years. Users are content with the flexibility and possibilities of this consensus-based
approach.
Quite a different approach is the ‘one-by-one’ method that is used by Akzo Nobel
Coatings. Just one key variable is selected per dimension. Variables should meet high
requirements:
- they are measurable in an objective way;
- they are excellent proximities for ‘profit impact’ and ‘supply risk’.
The financial value of items comes very close to the profit impact; the supply risk is
usually operationalized by the number of (alternative) suppliers. As a result, positions in
the matrix can be determined in a rather quick and unambiguous way. A related benefit is
that it is possible to compare matrices, because they are all based on a rather objective
measurement method. In addition, the method allows for identifying to what extent
products can shift to another quadrant. For instance, suppose product A is a raw material
that is only available from one mine in the world, owned by one supplier. Suppose
product B is a raw material of which the buying company requires its specifications to
meet extremely high quality elements. Obviously, products A and B will both be
positioned in the strategic quadrant. While product A can hardly be shifted, product B
could be moved towards the leverage quadrant, provided that its specification are less
strictly defined, opening the market to more suppliers. In other words, additional
background information is needed on products, markets and suppliers, so that
opportunities or threats are not neglected.
TE STRAKE uses a ‘weighted factor score’ method that includes a number of factors for
each dimension. The method allows for a completely customized approach, deciding on
factors, weights and (usually) scores. Total scores per dimension are calculated in an
additive model through the multiplication of scores and weights. The sub scores are
added to a single value.
It is assumed that a lower score on a factor can be compensated by a higher score on
another factor. The user of an additive model should ask himself if this is an acceptable
line of reasoning. For instance, if there is just a single supplier delivering a certain
product, then there is a maximal dependence on this supplier. Would it be possible for
the resulting supply risk to be compensated by other factors? Other disadvantages of the
weighted method are:
- depending on the level of aggregation, it could be necessary to dispose of a large
number of quantitative data that are rarely available in a purchasing information system;
- working with constituent factors, the overall picture can be hard to see, especially when
dealing with large numbers of factors and weights.
On the other hand, there is a major advantage, recognized by the users of the weighted
factor method. The portfolio analysis can be fully customized, according to one’s own
views and requirements. Preferably, all the relevant factors are included in the analysis.
5.6.3 Strategic directions
The investigated cases made clear that the development of portfolio-based strategies
requires additional information. In all case studies it was found that the following factors
were included:
- the (relationship with the) overall business strategy,
- the situations on supply markets, and
- the performance, capacities and intentions of (individual) suppliers.
The business strategy of TE STRAKE focuses on technological innovations, as first tier
supplier. Purchasing and supply have to connect with these basic principles, partnering
key suppliers for early involvement in product development and product improvement.
188
The basic points of departure of TE STRAKE are for instance rather different from the
ones at DSM, a firm that operates from an operational excellence perspective, always
looking for
cost reductions and efficiency. The marketing requirements on end markets are clearly
translated by Akzo Nobel Coatings in guidelines for the development of purchasing
strategies. Commodity markets are distinguished from niche markets, which affects the
selection of purchasing objectives and strategies to a considerable extent. Obviously, the
purchasing professional will take into account the situations on specific supply markets
and the assessments of individual suppliers. Items with high supply risks will be treated
differently, according to the reliability, the performance, the competences, and the
intentions of the connected suppliers.
DSM Akzo Nobel Coatings TE STRAKE

BU - E&P
Connection between meet the key success match the situations meet the demands and
purchasing strategies factors of the BU's in end markets requirements of major
and business strategy (commodity vs niche) customers
Portfolio based objectives 1. item-level 1. item-level 1. item-level

2. category-level 2. category-level
3. matrix-level 3. matrix-level
Strategies and conditions S1 pooling S1 forced single sourcing S1 contracts, stocks,

for bottleneck items C standardization (consignment, stocks) consignment
C one available supplier C high, unavoidable
S2 decomplex and dependence on suppliers
pooling S2 supplier development
C1 standardization C right cost-benefit ratio S2 pooling
C2 overspecification C grouping of items
Strategies and conditions S2 purchase card S1 minimalize order cost S1 separate ordering
for non-critical items C standardization and C low value items with C pooling is not
pooling are impossible low supply risk possible
S1 pooling S2 cross sourcing S2 single sourcing

C standardization and C availability of supplier C pooling is possible
pooling are possible
Strategies and conditions S1 partners of S1 maximize added value S1 competitive bidding

for leverage items convenience C many alternatives C interchangeable
C1 no co-design suppliers
S2 partnership (rare)
S2 strategic partners C competitive advantage
C1 co-design
C2 price performance
Strategies and conditions S1 strategic partners S1 accept suppliers' terms S1 strategic partnership
for strategic items C1 co-design C dominant supplier C1 technologically
C2 price performance advanced suppliers
S2 partnership (rare) C2 mutual trust and
S2 decomplex and C competitive advantage exchange of information
supplier development
C failing in co-design S3 supplier development S2 supplier management
and/or price (rare) C low performance
performance C right cost-benefit ratio
S3 accept the terms
C locked-in-relationship
Si = strategic response I Cj = condition j
FIGURE 5.10 The handling of strategic issues in the investigated cases

190
The second research question that we wanted to answer in the case studies refers to the
different kind of purchasing strategies, based on a portfolio approach. Figure 5.10 shows
the handling of strategic issues that are observed in the three cases. Portfolio-based
strategies should be connected to portfolio-based objectives. The research revealed three
levels of portfolio-based objectives:
- the item-level,
- the category-level, and
- the matrix-level.
In all of the three investigated cases, objectives were formulated at the item-level. For
individual items statements were made, regarding the optimal positions in the matrix.
Sometimes these objectives imply a movement in the matrix, in other instances it is
decided to hold a certain position. We will come back to this later.
On a category-level, objectives can be formulated for the four quadrants. For instance,
DSM wants to empty the non-critical category as much as possible. Akzo Nobel Coatings
employs very detailed measurable objectives for the categories in the matrix. For
example:
- reduce the number of items in the bottleneck quadrant by 5%, and
- increase the value of all leverage items by 50%.
Finally, it is possible to make statements at the level of the whole matrix. In general terms,
DSM prefers a matrix that is filled in a particular way:
- the bottleneck and non-critical categories should be as empty as possible, by means of
standardization and pooling of requirements;
- the leverage category should be filled with ‘partners of convenience’, meeting key
success factors of business units (always price and logistics);
- the strategic category should be filled with ‘strategic partners’, with the proper
capabilities for co-design;
- in addition, even for strategic items DSM rejects positions at the right side of the
quadrant, implying high levels of dependence and high levels of supply risk.
Akzo Nobel Coatings uses an even more sophisticated system of developing portfolio-
based objectives. For each business unit the whole matrix is assessed, mainly based on
the situations on their end markets. For instance, niche markets require high-quality
ingredients. A logical consequence is that the strategic quadrant will be filled with a
relatively large number of key suppliers, with whom close relationships are maintained.
Figure 5.10 has summarized the most common strategies and conditions that were found
in the case studies. This brings us to the question of how to find common ground between
those different kind of strategies and conditions. At first sight they are incomparable. If
we take a closer look and take up a higher level of abstraction, there are some striking
similarities, despite the differences in the level of the investigated cases. Conditions
usually refer to (im)possibilities to reduce the dependence on a supplier and to

(im)possibilities to increase buying power. We will come back to these issues later on.
The third research question explicitly refers to possibilities of moving in the matrix. Figure
5.10 provides an overview of portfolio-based strategies, albeit a rather unclear one. The
case studies revealed that, additionally to Kraljic’s theory, experienced practitioners were
very much aware of the different choices within each quadrant. Based on the interviews
and the overview of selected strategies, we concluded that for each category two different
kinds of strategic directions can be distinguished:
1. actions to pursue other positions in the matrix, and
2. actions to hold the same positions in the matrix.
Holding on to a position implicitly means that current circumstances are taken for
granted. We have observed that a position in the matrix can be accepted for different
reasons, sometimes referring to a positive, sometimes to a negative choice. There might
be a preference for a position because a firm is convinced that it is the best position for a
certain item. In other cases a position might be accepted, because there are no realistic
possibilities for change. The first type of strategies are of a more active, radical nature.
When possible and desirable, other positions in the matrix are identified and pursued.
This dichotomy between ‘holding position’ and ‘moving to another position’ has laid the
foundation of the conceptual model of strategic directions in the matrix, as is visualized
in figure 5.11. We will illustrate and amplify on the dichotomy for each product category.
These matters will be examined more closely in the survey.
192
Leverage items strategic items

(1)
(1)
(2)
(2)
Leverage items strategic items
(2)
(2)
(1) (1)
(1) = holding the same position

(2) = moving to another position
FIGURE 5.11 Conceptual model of strategic directions for all categories
If we take a look at the bottleneck and the strategic quadrant at the right side of the
matrix, those movements are pursued that reduce the supply risk. In terms of the matrix,
this means a movement to the left. Non-critical items are moved upwards, leverage
positions could be exchanged for strategic positions. We will elaborate on the strategic
directions that can be identified in the conceptual model.
Bottleneck items
(1) holding the position: “keep safety stocks”
If no other options are possible, then the category remains the same. Common responses
to unfavorable bottleneck-positions are: keeping (extra) stocks, consignment systems and
long term contracting.
(2) moving to another position: “decomplex the product and find a new supplier”
Bottleneck items are by definition of low value and of high risk. It should be interesting
enough, especially from an economic point of view, to look for alternatives. The most
common alternatives refer to the product (broadening specifications/decomplex) or to the
supplier (searching, managing and developing suppliers). These measures must lead to a
lower level of supply risk and a lower level of the dependence on a supplier.
Non-critical items
(1) holding the position: “individual ordering”
Whenever it is not possible to pool the purchasing requirements, the only remaining
option is some type of individual ordering, for instance by means of a purchase card.
(2) moving to another position: “pooling of requirements”

Preferably, non-critical items are put together in large quantities, increasing the buying
power of the firm. If necessary, a process of standardization is pursued. The strategic
direction is toward the leverage quadrant.
Leverage items
(1) holding the position: “maintaining a partnership of convenience”
The generally preferred leverage position can be used for a rather aggressive supplier
management. Competitive bidding and short term contracts are feasible options to exploit
the leverage position. The dominant power position allows for a command strategy. In
one of the investigated cases leverage suppliers are adequately referred to as ‘partners of
convenience’.
(2) moving to another position: “developing a strategic partnership”

Exceptionally, the leverage position is abandoned in search for a more strategic
partnership with a supplier. A cooperative strategy is only pursued if the supplier
involved is willing and capable of contributing to the competitive advantage of the firm.
Such a new role is only feasible for technologically advanced suppliers. The case studies
revealed that the move from ‘leverage’ to ‘strategic’ should be considered as an exception
to the rule.
Strategic items
(1a) holding the position: “maintaining a strategic partnership”
Long term relationships with key suppliers should always contribute to the competitive
advantage of the firm, as we have underlined. Relationships that include mutual trust,
mutual commitment and an open exchange of information are rare. A successful
partnership can be very valuable for both parties.
194
(1b) holding the position: “accepting a locked-in partnership”

On the other hand, a position in the strategic quadrant may be due to unchosen and
unfavorable conditions. The resulting ‘locked-in’ situation is commonly caused by a
patent position, a monopoly position, high switching costs (asset specificity) or by the
directions of a major customer. These circumstances produce an involuntary stay in the
strategic quadrant.
(2) moving to another position: “terminating a partnership, finding a new supplier”

A partnership may develop in an undesirable way. A supplier’s performance may
become unacceptable and incorrigible. This may start a painful process of reducing the
dependence on the supplier involved. The firm will have to search for, develop and
contract other suppliers, while bringing the relationship with the non-performing
supplier to an end.
With these generic descriptions of purchasing strategies, based on portfolio analysis, we

have answered the third research question: what kind of movements are considered in
the Kraljic matrix? We have filtered and analyzed the responses in the interviews and
summarized the findings. This process has resulted in an overview of possibilities,
visualized and represented in figure 5.11. We feel that a first and elegant solution is
found to the unanswered question of goals (future positions) and means (strategies),
allowing for further study. The survey will investigate the alternative conditions under
which these strategies are selected in practice.
5.7 Conclusions
The case studies began with the contention that we need to gain a better understanding
of how purchasing portfolio models are being used in practice and how they could be
used by purchasing professionals in order to pursue effective differentiated purchasing
strategies. The literature study has identified a number of problems and unanswered
questions. Publications however, do not reveal how purchasing professionals actually
handle those issues. This study has clarified these issues, describing advanced practices
with respect to purchasing portfolio models. The research questions referred to
measurement issues and portfolio-based strategies. The investigated cases provided useful
insights in the possibilities and actual use of purchasing portfolio analysis. In the case
studies we found a variety of approaches and differences of scope and scale, which has to
be viewed in the specific business context. The research questions were mainly answered
by interviewing experienced professionals.
The study was aimed at exploring and explaining the third major research question: How
are portfolio models employed by experienced purchasing professionals? A subset of
research question referred to measurement issues and portfolio-based strategies.
The case studies revealed three rather distinctive methods of measuring variables and
weighting factors:
(1) consensus method
(2) one-by-one method (1 factor per dimension)
(3) weighted factor score method.
Each method satisfies the needs and expectations of the different users. The reason for
this can be found in the additional steps that have to be taken in the portfolio analysis
(see Appendix D). Before strategic actions are determined, it is imperative to complete a
further process of interpreting and reflecting on the results. The positioning of items in the
matrix should be considered as the starting point of portfolio analysis, definitely not the
finishing point. After the matrix is filled, it is imperative that users reflect on the results.
If necessary, manual adjustments should be made. In-depth discussions on the positions
in the matrix are considered as the most important phase of the analysis. Strategic
discussions provide deeper insights and may lead more easily to consensus-based
decisions. It is felt by the users that the Kraljic framework facilitates these important
discussions to a large extent.
Some argue that the complexity of business decisions does not allow for simple
recommendations. How could one deduce strategies from a portfolio analysis that is
based on just two basic dimensions (e.g. Dubois and Pedersen, 2002)? Actually, the
answer is simple: one cannot! In addition to the various factors that constitute the two
dimensions of any matrix, it was found that experienced portfolio users always included
additional information on:
- the overall business strategy (related situations on end markets),
- the specific situations on supply markets and
- the capacities and the intentions and competences of individual suppliers.
The selection of portfolio-based strategies was explored as well.

Based on the case studies, a conceptual model of strategic directions has been presented,
providing insights and overview of the main strategic choices for the categories in the
matrix. In addition to Kraljic’s strategic recommendations, different kinds of strategic
responses were identified and described for each category. A dichotomy was identified
between:
- strategies to hold a position (1) and
- strategies to move to another position (2).
At the right side of the matrix (in the bottleneck and the strategic areas) movements are
pursued in order to reduce a high level of supply risk. In terms of the matrix, this means
moving to the left. Non-critical items are preferably moved upwards, exceptionally
leverage positions are exchanged for strategic positions. These are the most common
movements within the matrix.
196
From the buyer’s perspective a new classification of partnerships was found, related to
the portfolio matrix:
- partners of convenience, located in the leverage quadrant, where relationships are
dominated by the buyer;
- strategic partnerships, located somewhere in the middle of the leverage and strategic
quadrant, further characterized as balanced relationships based on a high level of mutual
dependence;
- locked-in ‘partnerships’, located at the right side of the strategic quadrant, where
relationships are dominated by suppliers, who are indispensable for the buyer.
Based on the case studies, a conceptual model of strategic directions is presented,
providing insights and overview of the main strategic directions for the categories in the
matrix. Variables and relations in this model will be quantified and tested with the data
that will be gathered with the survey.
197
6 The survey
This research project includes a literature study, case studies and a survey. The literature
study has answered questions regarding the various portfolio models in purchasing,
identifying differences and similarities. The case studies have addressed questions and
problems with respect to the actual use of portfolio models by experienced purchasing
professionals, referring to important issues such as the measurement and weighting of
factors and the selection of portfolio-based purchasing strategies. In this chapter we will
report on the design and the results of a survey among purchasing professionals.
6.1 Introduction and objectives

To a large extent, the survey is aimed at answering the second, the fourth and the fifth
major research questions:
Which factors would explain the utilization of the purchasing portfolio analysis?
and
Under which conditions are the various portfolio-based strategies pursued in purchasing
management?
and
What is the role of power and dependence in the Kraljic approach?
Based on our analysis of literature, a use model was developed, identifying factors that
would explain the use of the portfolio analysis. Also the model includes the direct
accountable impact of portfolio use. The sub-questions aim to quantify the relationships
with the central use-variable. In addition other sub-questions refer to the number of users
and differences between users and non-users. In the case studies we have identified what
kind of purchasing strategies can be based on a portfolio approach. Tentative insights are
gathered with respect to conditions and circumstances, leading to certain strategic
choices. The possibilities for selecting purchasing strategies are obviously limited by
external conditions. The survey is aimed at identifying and quantifying these conditions
which will result in additional insights for understanding and explaining the selected
strategies in
198
different situations. A specific set of conditions will be referred to as a relationship-

dependence profile, which will include the determinants of buyer’s dependence and the
determinants of supplier’s dependence. In addition, other factors will complete the
profiles.
Based on literature and case studies we have developed:

- a conceptual model for the use of a purchasing portfolio approach;
- a conceptual model of strategic directions for all categories in the Kraljic matrix;
- a conceptual model of determinants of buyer’s and supplier’s dependence.
This survey is intended to:
- measure variables and quantify the relationships between variables in the conceptual
models,
- test the hypotheses that have been formulated.
6.2 The conceptual models

The use-model identifies variables in order to explain the use of purchasing portfolio
analysis in companies (see chapter 3). Based on the literature study, five groups of
variables have been identified as independent explanatory variables: company size, share
of purchasing, position of the purchasing department, purchasing’s professionalism and
orientation of purchasing. The position, the professionalism and the orientation of
purchasing are constructs, which will be elaborated in section 6.6, which deals with
issues of measurement and operationalization. In chapter 3 we have formulated a set of
hypotheses for the (positive and negative) relationships between the explanatory
variables and portfolio use.
Subsequently, the use is expected to have a positive impact on the understanding of the
overall purchasing situation in companies, and on the development of differentiated
purchasing strategies. These hypotheses will be tested too with the data of the survey.
The conceptual model of strategic directions is based on a comparative analysis of the case
studies. Altogether, 9 scenarios are described, identifying different strategic directions
(actions), and the circumstances that accompany these purchasing strategies (see
Appendix F: “Elaboration and overview of the 9 scenarios”). A dichotomy of strategic
directions was found, distinguishing two different kinds of strategic directions for each
category:
1. actions to pursue other positions in the matrix, and
2. actions to hold the same positions in the matrix.
Holding on to a position implicitly means that current circumstances are taken for
granted, or have to be taken for granted. We have observed that a position in the matrix
can be accepted for different reasons, sometimes referring to a positive, sometimes to a
negative choice. A position might be preferred because a firm is convinced that it is the
best position for a certain item. In other cases a position might be accepted, because there
are no realistic possibilities for change. The first type of strategies are of a more active,
radical nature. When possible and desirable, other positions in the matrix are identified
and pursued. The dichotomy for each category is elaborated in a limited set of purchase
situations. Each scenario contains a description of:
- the selected portfolio-based purchasing strategy,
- Kraljic’s dimensions (profit impact and supply risk), and
- additional specific circumstances for the selection of the purchasing strategy.
This means that in the conceptual model, the selection of a (portfolio-based) purchasing
strategy is connected to these factors. Therefore, each purchasing situation (scenario)
combines a (portfolio-based) purchasing strategy, a position in the matrix (values on the
Kraljic dimensions, ‘treatment variables’), and a set of specific circumstances. In this
study we will introduce and use the concept of relationship-dependence profile, as an
instrument to measure the values and impact of conditions that accompany a scenario. A
relationship-dependence profile should include a limited number of key factors
appropriate to characterize different types of buyer-supplier relationships. Based on
former literature study and based on the results of the case studies, four main groups are
set to form a relationships-dependence profile:
- determinants of buyer’s dependence,
- determinants of supplier’s dependence,
- relationship characteristics: trust and commitment,
- the nature of specifications (product customization, possibilities for standardization).
A combination of these factors will be addressed to as a relationship-dependence profile.
Obviously, different profiles are expected in different buyer-supplier exchange

relationships. Purchasing strategies are considered as responses to different situations.
From the buyer’s perspective the recurring question is: ‘what kind of relationship is
possible and/or desirable with a specific supplier?’ In the upper half of the matrix, based
on the case studies, a clear distinction can be made between:
- partners of convenience, located in the leverage quadrant where relationships are
dependence;
In the lower half of the matrix strategic responses are related to the nature of
specifications.
Based on a literature study, it is expected that the occurrence of different types of buyer-
supplier relationships can be explained by some key constructs and variables. Besides
factors relating to buyer’s and supplier’s dependence, the survey includes
200
operationalizations of product customization, possibilities for standardization, trust and

commitment that are connected to relationships with suppliers.
The prominent role of power and dependence in buyer-supplier relationships has
directed the literature study to the determinants of organizational dependence. Analysis
of literature has resulted in a tentative model of buyer’s and supplier’s dependence. This
model is presented to all respondents in the case studies. The reactions to the tentative
model are incorporated in the modified model of buyer’s and supplier’s dependence. To
gain more insights in the selection of portfolio-based strategies, the scenarios will be
further analyzed and described by the quantification of conditions in the profiles.
Quantification of buyer’s and supplier’s dependence for the different purchasing

situations (scenarios) allows for:
- the identification of specific dependence profiles, describing and explaining differences
between strategic choices and circumstances in the portfolio matrix;
- the explanation of buyer’s dependence and the explanation of supplier’s dependence, in
other words the testing of hypotheses with respect to the impact of the determinants on the
buyer’s and the supplier’s dependence.
6.3 Relationship-dependence profiles: a condition-seeking

research strategy
Consequentially we have identified two different types of portfolio-based strategies in
each quadrant of the matrix; in case of strategic items we found three strategic responses.
It is possible to conceptually explain the selection of a strategy in a quadrant by pointing
at differences in corresponding relationship-dependence profiles. Each portfolio-based
strategy constitutes a unique fingerprint that is expected to be significantly different from
all other profiles, but most importantly, from the profile that belongs to the other strategy
in the same matrix category. Otherwise, the preference for strategy A to strategy B in a
quadrant can not be attributed to the explanatory variables. This means that a set of
hypotheses can be formulated with respect to the different relationship-dependence
profiles. The hypotheses focus on the question whether or not there are significant
differences between two relationship-dependence profiles in the same category. The
found differences will be used for a better understanding of the selection of different
strategies under different conditions in the same quadrant.
1. Hypothesis for the bottleneck category

Hypothesis C1 For bottleneck products, there is a significant difference
between the relationship-dependence profile that fits a strategic
choice for ‘keeping safety stocks’ and the relationship-
dependence profile that fits a strategic choice for
‘decomplexing the product and finding a new supplier’.
2. Hypothesis for the non-critical category

Hypothesis C2 For non-critical products, there is a significant difference
between the relationship-dependence profile that fits a strategic
choice for ‘individual ordering’ and the relationship-
dependence profile that fits a strategic choice for the ‘pooling of
requirements’.
3. Hypothesis for the leverage category

Hypothesis C3 For leverage products, there is a significant difference between
the relationship-dependence profile that fits a strategic choice
for ‘maintaining a partnership of convenience’ and the
relationship-dependence profile that fits a strategic choice for
‘developing a strategic partnership’.
4. Hypotheses for the strategic category

Hypothesis C4 For strategic products, there is a significant difference between
for ‘maintaining a strategic partnership’ and the relationship-
dependence profile that fits a strategic choice for ‘accepting a
locked-in partnership’.

for ‘maintaining a strategic partnership’ and the relationship-
dependence profile that fits a strategic choice for ‘developing a
new supplier’.

for ‘accepting a locked-in partnership’ and the relationship-
dependence profile that fits a strategic choice for developing a
new supplier’.
202
In addition to the testing of the 6 hypotheses C1 to C6 ,we will take a closer look at the
composition of the relationship-dependence profiles. The quantification of variables
allows for analyzing the differences in more detail, on a lower level of abstraction,
namely the level of the dependent variables. Pairwise comparisons of variables between
profiles will be used in our search for conditions under which the various purchasing
strategies are selected. With this we propose to discover which of the many conditions
are indeed necessary or sufficient to produce a certain result, namely the selection of a
certain portfolio-based strategy. The aim of this part of our study is therefore not to
answer questions like “Does strategy Si occur?” or “How often is strategy Si being
selected?” In contrast, we follow a condition-seeking research strategy which answers
questions of the form “Under which conditions strategy Si will be selected?” The results
are of the form “Strategy Si is selected if conditions C1, C2 and C3 are met”. In other
words, the answers must specify conditions under which a given result is obtained
(Greenwald et al., 1986).
In section 6.5 we will elaborate on the design of the study and the implications of the data
analysis.
6.4 Population and sample

The population for the sample consists of Dutch manufacturing companies. The study is
limited to industry for several reasons. In most companies the share of purchasing is
relatively high, putting purchasing and supply management on the strategic agenda.
Purchasing is always important, purchasing strategies are likely to be developed, and
purchasing professionals are likely to be knowledgeable about the selected strategies and
reasons accompanying those strategies. Furthermore, limiting the study to industrial
companies increases the chance of gathering comparable data that can be combined, and
the chance of finding significant relations between variables.
The target respondents for the survey are purchasing professionals within manufacturing
companies. Purchasing professionals are in the best position to answer the questions of
this survey, because of their experience, expertise and insights with respect to the use of
purchasing portfolio models and the development of differentiated purchasing strategies.
Questionnaires are administered to 1,153 individuals, mostly purchasing managers, from
companies who are member of the Dutch Association of Purchasing Management (NEVI).
The companies are all manufacturing companies. Manufacturing can be defined as the
physical or chemical transformation of materials or components into new products, in
accordance with the United Nations Classifications Registry
(http://esa.un.org/unsd/cr/registry/regcs.asp?Cl=2&Lg=1&Co=D).
Excluded are companies that are in the business of agriculture, construction, wholesale,
retail, and (public and private) services. The NEVI-industry database consists of the
following categories:
- electro technical industry
- metal products industry (metallurgical)
- chemical industry
- machine industry
- wood, furniture and paper industry
- metal basic industry
- means of transport industry
- building materials, glass and pottery
- instrument or optical industry
- textile and clothing industry
- graphic industry
- petroleum and coal processing industry
- other industry.
6.5 Design and procedure

In accordance with the research objectives, a written questionnaire has been used,
gathering information from a relatively large number of respondents. The survey
procedure included four mailings, in chronological order:
1. the pilot study (pre-test): the questionnaire has been sent to a limited number of
respondents,
2. the initial mail-out: administering the actual survey to the selected respondents
(sample),
3. the follow-up mailing: a replacement questionnaire has been sent to all non-
respondents,
4. an electronic mailing: a replacement questionnaire has been sent to all non-respondents
with known e-mail addresses.
Different sources of knowledge and expertise have been used to develop the
questionnaire. Principally, the questionnaire is aimed at measuring the variables and
relationships in the conceptual models. These models are based on the insights from the
literature study and the case studies. Furthermore, the construction of items is largely
based on literature review (operationalization). This has resulted in a first draft
questionnaire, that was discussed with a focus group of academics. Finally, the modified
questionnaire was pre-tested by a small number of individuals from four industrial
companies (GTI, EMTEC, DSM, and Akzo Nobel). The pre-test group were asked to
review the survey primarily for the clarification of scenarios, the questions and the time
required to complete the survey. Pre-testing provided the opportunity to improve the
questionnaire, and to enhance its reliability and validity.
204
The questionnaire consists of three sections:

A. general purchasing issues,
B. portfolio-based purchasing strategies,
C. purchasing portfolio analysis.
The questionnaire and the letters of recommendation for the first and second mailing can
be found in Appendix G.
The first section gathers information about the respondent, his/her company and the
purchasing department/function within the company. In the second section respondents
are asked to assume the role of a purchasing expert in their own company where they are
actually employed. To play this role, it is important to have an overview of
circumstances, accompanying and resulting in the selection of certain pre-specified
purchasing strategies. The respondents are asked to evaluate a series of questions related
to different purchasing situations (scenarios). The procedure is that all respondents are
asked to respond to all the elaborated scenarios. This is done by means of closed items,
measuring the reactions on statements on a 5-point Likert scale. The third section focuses
on the actual use and attitudes regarding the purchasing portfolio analysis. The answers
to the items in part A and C will be used for the quantification of variables and
relationships between variables, all within the model with respect to the use of a
purchasing portfolio approach.
Part B of the questionnaire can be typified as a repeated measures design, because all the
respondents participate in all the conditions of the experiment. The experiment involves
the manipulation of treatments, namely the 9 scenarios (9 levels of treatments). Each
scenario describes a situation in terms of the Kraljic dimensions (profit impact and
supply risk, ‘high or low’) and in terms of a the selection of a corresponding pre-specified
purchasing strategy. The respondents are consequently asked to answer the same 14
questions that constitute the 14 dependent variables of the relationship-dependence profile.
The construction of this experimental design is motivated by the following considerations:
- the objective of the survey is to describe and to explain the selection of differentiated
purchasing strategies, based on portfolio analysis;
- it is impossible to ask respondents about specific circumstances leading to (unspecified)
strategic choices with respect to the thousands of products and/or the hundreds of
suppliers they are usually dealing with;
- a self-selection of products by respondents would not lead to representative samples of
purchase situations.
A repeated measures study means that measurements are taken on each respondent under
each of several conditions (treatments). In other words, the same subject participates in
all conditions of an experiment. The levels of the conditions describe the within-subjects
(W-S) variable. There is no between-subjects (B-S) variables, since the respondents are not
classified into groups. The design can be further characterized by the fact that there are
multiple dependent variables.
A repeated measures design has a distinct advantage over other experimental designs.
The main reason for within-group variability is that there are individual differences
among the respondents. Even though the subjects receive the same treatment, their scores
on the dependent variable(s) can differ considerably, because of differences on relevant
background variables, such as I.Q., education, motivation, etc. A common solution in
experimental designs is through block on such variables, which means that subjects are
blocked into (more) homogenous groups. In a repeated measures design, blocking is
carried to its extreme, by blocking on each subject. Variability among the subjects to
individual differences is removed from the error term, which makes the design more
powerful than randomized designs, where subjects are randomly assigned to different
treatments (Stevens, 2001).
However, there is a disadvantage. The accuracy of the F-test depends upon the
assumption that scores in different conditions are independent. Obviously, when repeated
measures are taken, the independence-of-observations assumption is violated. The scores
taken under different scenarios are likely to be related, because they come from the same
respondents. A conventional F-test will lack accuracy and an additional assumption is
required: the assumption of sphericity, which states that the level of dependence between
experimental conditions is (roughly) equal. The correlation between scores on the
dependent variable in our study for scenario #1 and #2 should be about the same as the
correlation between scenario #5 and #6, or #7 and #8, etc. Beforehand, this will be very
unlikely. If the assumption of sphericity is violated, appropriate measures are necessary.
Tabachnick and Fidell (2001) suggest the use of one of the significance tests, which are
adjusted for violation of the assumption, such as Greenhouse-Geisser (G-G) or Huynh-
Feldt (H-F) with adjusted significance levels.
Significance tests for repeated measures and (M)ANOVA are based on the assumptions
of multivariate normality. Normality requires that the sampling distributions of means of
the various dependent variables at each level of the W-S variables are normally
distributed (Weinfurt, 2000). In our study normality is not likely to produce problems.
Generally speaking, repeated measures analysis is considered robust to violations of the
normality assumptions (Stevens, 2001). This means that the Type I and Type II error rates
for the F test are significantly distorted only when the distribution of the data is an
extreme deviation from normal (Weinfurt, 2000). Moreover, the central limit theorem
suggests that the sampling distribution of means approaches normality (even when raw
scores do not) in case of large samples (Tabachnick and Fidell, 2001). As will be
elaborated later, the number of cases in our study exceeds the number of ‘20 cases in the
smallest’ cells, which should ensure robustness.
206
By comparing two profiles of corresponding scenarios in the same quadrant, we aim to

search for statistically significant differences between profiles. However, a series of
independent t-tests will not do, because we are dealing with a repeated measures design.
The option Estimated Marginal Means (EMMEANS) will be used for the multiple
comparisons within the GLM-repeated measures SPSS-procedure. As with the overall F
test for a W-S effect, post hoc analyses are vulnerable to the violation of the sphericity
assumption. When sphericity is violated, the Bonferroni method is considered to be
generally the most robust of the available techniques especially in terms of power and
control of the Type I error rate (Field, 2000). By applying a Boneferroni correction to each
test, one controls the family wise Type I error (Tabachnick and Fidell, 2001). In our study
we will use the Bonferroni adjustment when testing for differences of variables between
profiles.
In his comment on multiple comparison procedures with repeated measures, Howell
(2002) stresses that the test statistic itself is not the issue, because the basic underlying
statistic is the t-test for the different procedures. The different tests, such as a standard
contrast, a Bonferroni test, a Tukey test, and a Scheffé test, all result in the same t-value.
The only difference is the critical value required for significance. The Bonferroni test uses
a t-test, but then evaluates that t at alpha = .05/c, where c is the number of comparisons
and .05 is the desired alpha. When dealing with 84 comparisons (6 pairs of scenarios x 14
dependent variables), the critical p-value has to be adjusted to:
.05/84 = .0006
Because of the relatively large number of comparison, the correction in our study will be
rather severe. If we could select a limited number of comparisons on the basis of theory,
the correction would obviously have been less severe. However, we have to use post hoc
procedures, including a large number of multiple comparisons, since we do not have a
set of a priori comparisons. Therefore, an explorative investigation will be conducted, in
search for statistical differences between profiles. The found differences will be traced
back to the variables that constitute a profile (e.g. factors relating to dependence and
other conditions).
6.6 Measurement and operationalization

The operationalization of a concept translates the conceptual definition into measurable
terms. An operationalization specifies what the researcher must do in order to measure
the construct concerned. In this study, for some constructs it is not very complicated to
define and measure them (for instance ‘purchasing share’). However, most of the
constructs in the conceptual models are abstractions, posing researchers for more difficult
choices and dilemma’s. In those cases, special attention is paid to the insights and
operationalizations that are used in other empirical and conceptual studies.
6.6.1 Variables and constructs in the use-model
variables operationalizations and scales

company size number of employees
number on a ordinal scale
purchasing share purchasing spend as a percentage of total turnover

number on an interval scale
position of purchasing * (1) top management recognizes Purchasing as an important

contributor to the competitive position
(2) purchasing function reports directly to top management
5-point Likert scale: (dis)agree
purchasing’s professionalism* (1) skills of purchasing personnel are adequate for working in cross-
functional teams
(2) skills of purchasing personnel are adequate for developing
purchasing and supplier strategies strategies
purchasing orientation * (1) in our company, purchasing aims at collaboration with suppliers
(2) in our company, purchasing aims at tough negotiations with
suppliers in pursuit of the lowest prices
(3) in our company, purchasers mainly relate to clerical and
operational activities
use is the purchasing portfolio analysis being used?

dichotomous variable: yes, no (nominal)
use intensity how often is the purchasing portfolio analysis being used?
ordinal scale: never, less than/approximately/more than once a year
understanding of situation in our company, there is sufficient understanding of the possibilities

and problems of the products that are bought
5-point Likert scale: agree-disagree
development of strategies in our company, there is sufficient differentiation with respect to the
ways of handling suppliers
5-point Likert scale: agree-disagree
* cf. Rozemeijer (2000)

208
Rozemeijer (2000: 170, 232-233) identified ‘position of purchasing’, ‘purchasing’s

professionalism’, and ‘purchasing orientation’ as the main factors, constituting the
construct variable ‘purchasing maturity’. His study indicated additional factors to
explain a company’s purchasing maturity. However, these factors are not included in our
conceptualization of ‘maturity’, because they are tailored to the specific objectives of
Rozemeijer’s study: creating corporate advantage in purchasing. The factors refer to
differences between purchasing departments of fairly autonomous business units.
6.6.2 Variables and constructs relating to buyer’s and supplier’s dependence
Preparing the survey, an extensive literature search is conducted to the measurement of

‘organizational dependence’. Points of interest and study were:
- the perspective (‘whose dependence?’);
- measurement method: directly (by means of a single item or a proxy measure) or
indirectly (by means of multiple items);
- measuring perceptions or other approaches;
- the formulation of questions and items;
- the number of items (in case of multiple items)
- the scales (open/closed items, Likert scales, 5- or 7 point scales).
The results are summarized in the appendix H, “Measurement and operationalizations of
organizational dependence”. The main conclusions are:
- there is much diversity in perspective;
- ‘dependence’ as a construct is being measured indirectly (via multiple items and
questions) and directly (by means of a single items or by means of a proxy measure);
- all studies employ (subjective) perceptions from the perspective of one trading partner;
- in most studies ‘organizational dependence’ is used as an explanatory variable, rather
than a variable that is to be explained by a set of independent variables;
- ‘vulnerability’ and ‘replaceability’ seem to be recurring terms, looking for the core of the
dependence-construct. How important is a trading partner to the other? How important
are the ‘resources’ involved? Are there alternative buyers/suppliers? What are the
consequences of terminating a relationships (switching cost)?
In the literature study we have used some basic thoughts of the resource-dependence
theory. However, when implemented in a purchasing context, we have deviated from the
original work. Pfeffer and Salancik (1978: 51) made a firm statement with respect to the
relationship between the three acknowledged dimensions of dependence: “A resource
that is not important to the organization cannot create a situation of dependence,
regardless of how concentrated control over the resource is. Also, regardless of how
important the resource is, unless it is controlled by relatively few organizations, the focal
organization will not be particularly dependent on any of them.” In their view every
dimension constitutes a necessary, but by itself insufficient condition for producing
‘dependence’, and when one is absent, an organization will be not be ‘particularly

dependent’. We have chosen for a somewhat different approach. We have considered
dependence not as a dichotomous variable (‘dependent’ versus ‘independent’), but as a
continuous variable, allowing for different degrees of dependence. Furthermore, we have
not proposed an operationalization of the ‘dependence’-concept by measuring the
determinants of that dependence. Some argue that it seems good to view dependence as a
multidimensional construct, allowing for closer examination of each of its determinants
(Heide and John, 1988). A reason would be that the different aspects of dependence do
not necessarily covary, or have the same effect on outcome variables. To grasp the overall
meaning and implication of ‘dependence’, many researchers propose and prefer a
multiple-items construct, capturing all relevant dimensions. The degree of dependence is
usually assessed by combining values of relevant items/variables into one constructed
item. This seems a reasonable approach, since dependence is by nature determined by
several factors and variables, and it is consistent with a general feeling that it is possible
to consider ‘degrees of dependence’. Moreover, it leaves the possibility open for a
compensatory decision rule: a low score on a certain variable can be compensated by a
high score on another variable. However reasonable this approach might be, in our study
we have selected another approach, conform to Noordewier et al. (1990) and Berger et al.
(1995), for example.
In our study we propose to measure ‘dependence’ as a single item construct. This means
that respondents are directly asked to assess the buyer’s dependence on a supplier and
the supplier’s dependence on the buyer. In addition, determinants of dependence are
measured in a direct way. The results of the case studies indicated that expert
respondents do not have any problems in assessing dependence, when asked directly.
For purchasing professionals ‘dependence’ is a concept that is easily understood and that
plays a prominent role in the development of purchasing strategies. Based on these
empirical experiences, we do not expect to encounter any major validity problems to this
respect. Purchasing professionals are able to assess dependence-situations. However, to
avoid related validity problems, respondents will indicate the level to which they
recognize different (dependence) situations.
The determinants of dependence are based on literature study and interviews in the case
studies. The different single dependence-items in other studies coincide to a large extent
with the determinants in our study. By measuring and using the determinants separately,
it is possible to find out which of the variables has a significant influence on buyer’s
dependence and supplier’s dependence. In addition, these various impacts will be
calculated for the 9 different scenarios, allowing for an in-depth analysis of
organizational dependence. The results will constitute corresponding dependence profiles
that are related to
210
the portfolio-based purchasing strategies, which matches the core of the overall research
project. If a multiple-item approach had been selected, important information would have
been lost.
variables operationalizations and scales

buyer’s dependence * the buyer is dependent on the supplier
logistical indispensability reliability of delivery of the product is important for an uninterrupted flow of
manufacturing
need for supplier’s technological the buyer needs the technological expertise of the supplier
expertise **
alternative suppliers the product can be bought from other suppliers
switching costs for buyer *** replacing the supplier, the buyer will incur high switching cost
supplier’s dependence * the supplier is dependent on the buyer
financial magnitude **** the buyer is an important customer for the supplier, considering the volume
of trade
need for buyer’s technological the supplier needs the technological expertise of the buyer
expertise **
alternative buyers the product of the supplier can be sold to other customers
switching costs for supplier *** the supplier will incur high switching cost, replacing the buyer by other
buyers
All variables are measured on a 5-point Likert scale (agree - disagree).
* cf. Noordewier, John and Nevin (1990) and Berger, Noorderhaven and Nooteboom (1995)
** cf. ‘need for supplier’s skills’, as used by Campbell and Cunningham (1983) and ‘knowledge
exchange’, as used by Berger, Noorderhaven and Nooteboom (1995)
*** cf. ‘difficulty and cost in replacing a partner’, as used by El-Ansary and Stern (1972), Etgar (1976),
Brown, Lusch and Muehling (1983), Heide and John (1988), Sriram, Krapfel and Spekman (1992),
Lusch and Brown (1996), Dant and Gundlach (1998)
**** cf ‘share of supplier’s output taken by the buyer’, as used by Campbell and Cunningham (1983); in
contrast to Kraljic’s conceptualization, ‘financial magnitude’ should refer to the supplier’s position, not
the buyer’s position, because the variable is considered here as a determinant of supplier’s dependence
6.6.3 Variables and constructs relating to other conditions
We have stated earlier that a number of factors is considered to be important, explaining

differences in the selection of purchasing strategies. We have identified ‘product
customization’, ‘commitment’ and ‘trust’ as key constructs, explaining the selection of
different purchasing strategies. Product customization indicates the level of adaptation
that is made by the supplier. Trust is seen as a prerequisite for the development of a
fruitful relationship. Successful relationships require commitment and trust. Literature on
partnership, cooperation and buyer-supplier relationships reports extensively on trust

and commitment. Therefore, it is possible to use and adopt existing conceptualizations
and operationalizations of constructs.
Product customization refers to adaptations that are especially made by a supplier to

serve the needs of a specific, important customer (Hallén et al., 1991). This product
specification dimension is frequently used in the automotive industry, where a limited
number of suppliers deliver highly customized products (Dyer, 1996; Dyer and Chu,
2000). The level of product customization indicates the supplier’s willingness to invest in
the relationship (Hartmann et al., 2001). On the other hand it shows the willingness or
the need of the buying company to rely on a single source (Morgan and Hunt, 1994). We
operationalize product customization as the level to which product specifications are
tailored to the specific needs and wants of the buyer.
Commitment and trust have been identified as essential prerequisites for building and
developing customer-supplier relationships (De Ruyter et al., 2001). Successful
relationships require commitment and trust (Morgan and Hunt, 1994).
The concept of trust is central in understanding the development of interorganizational
relationships. Trust is generally considered as a necessary ingredient for the development
of long-term buyer-supplier relationships (Ganesan, 1994). Research suggests that trust in
supplier-buyer relationships is an important source of competitive advantage. Dyer and
Chu (2000) concluded that trust (1) lowers transaction costs (Sako, 1992; Barney and
Hansen, 1994; Dyer, 1996), (2) leads to superior information sharing routines, and (3)
facilitates investments in relationship assets (Asanuma, 1989; Dyer, 1996). Zaheer et al.
(1998) found that trust is an important determinant of supplier performance, examining
trust from both buyer and supplier perspective. There is much support for the importance
of trust in building and sustaining buyer-supplier relationships. At the same time, trust is
a diffuse concept, defined in different ways (Schary and Skjøtt-Larsen, 2001). Blomqvist
(1997: 271) points at “the many faces of trust”, referring to the various dimensions and
levels of trust, and to the many disciplines that incorporate the concept of trust (social
psychology, philosophy, economics, contract law, and marketing). Blois (1999) examined
this lack of clarity in the conceptualization of trust.
Trust can be seen as a necessary condition for accepting high levels of dependence. Zand
(1972: 230) defined ‘trust’ as the conscious regulation of one’s dependence on another: “
(…) one who does not trust will try to minimize his dependence on others.” Ganesan
(1994) predicted that firms with high dependence will seek constantly to escape from this
dependence. However, with trust, things will be different. Trust is necessary for the
perception of a fair division of the pie of resources in the future. The need for trust will be
high in situations that are perceived as risky. When trust is high, monitoring costs go
down, the number of safeguards that need to be put in place to prevent opportunism can
be reduced, and governance becomes less of a salient issue (Barringer and Harrison,
212
2000). The development of trust over time is an area of ongoing interest and research
(Nooteboom et al., 1997).
The connotations of trust and dependability may seem quite congruent. However, trust
and dependability are not interchangeable, because they do not have identical meanings.
As Kumar (1996) appropriately remarked, a partner that promises to punish and always
follows through, is dependable, but is not a company in which one places trust. Trust
refers to the willingness to rely on exchange partners in whom one has confidence
(Moorman et al. 1992). Basically, the most important aspect of trust is a positive belief, an
affective sentiment about an exchange partner. Morgan and Hunt (1994: 23)
conceptualized trust as existing when one party has confidence in an exchange partner’s
reliability and integrity. This definition parallels that of Anderson and Narus (1990),
Moorman et al. (1993) and Ganesan (1994). Some include trust as risk-taking behavior or
a willingness to engage in such behavior (e.g. McAllister, 1995). To be clear, we have
limited the conceptualization of ‘trust’ to a cognitive expectation, an affective sentiment,
and we have excluded behavioral intentions and trusting behaviors. Trust results from
the belief that the other party is reliable and has high integrity. Sako (1992) underscored
that trust refers to an expectation that a trading partner behaves in a predictable and
acceptable manner. Since ‘predictability’ exists for different reasons, different types of
trust should be distinguished. In fact, we have identified two different dimensions of
trust: ‘reliability’ and ‘integrity’. Reliability and credibility refer to the extent to which an
exchange partner has the required expertise to perform the job effectively (Ganesan,
1994). This dimension focuses on the expectancy that the partner’s word or statement can
be relied on (Doney and Cannon, 1997). This type of trust is referred to as competence
trust: the ability of an exchange partner to perform according to agreements (Nooteboom,
1996). Competence implies that a partner has the required technical capabilities, skills
and know-how (Blomqvist, 1997). Another type of trust is goodwill trust, which refers to
the integrity and benevolence of parties. Some propose that the true meaning of trust
implies a ‘leap of faith’: parties believe that each is interested in the other’s welfare and
that neither will act without considering the action’s impact on the other (Kumar, 1996:
95). Goodwill trust reflects the belief that one’s partner is interested in the other’s welfare
and will not take actions which will negatively impact the firm (Anderson and Narus,
1990; Geyskens et al., 1996). A similar conception is confidence trust, which refers to the
confidence that the other party in the exchange relationship will not exploit its
vulnerabilities (Dyer and Chu, 2000). Goodwill or confidence trust is not based upon
contracts or sanctions, but rather on non-contractual mechanisms. To conclude, we have
operationalized goodwill trust as the belief that the supplier will not misuse his position
and actually takes the buyer’s interests into consideration. The willingness to accept
dependence will be higher in case of goodwill trust, than in case of competence trust.
The contractual trust referred to by Sako (1992) has been omitted. In conformity with
Blomqvist (1997), contract is seen as an alternative or additional means of coordination,
but not as another form or dimension of trust.
Commitment too has emerged in literature as a critically important element in business

relationships. Moorman et al. (1992: 316) defined commitment as an enduring desire to
maintain a valued relationship. This corresponds with the belief that relationship
commitment only exists when the relationship is considered important (Morgan and
Hunt, 1994). Although different conceptualizations exist, commitment has typically been
defined as the intention of an exchange partner to continue a relationship (Anderson and
Weitz, 1989; Dwyer et al., 1987).Geyskens et al. (1996) underscored that different
motivations can underlie this intention, and thus various types of commitment. Of these,
affective commitment and calculative commitment appear most frequently and also seem
to be the most relevant (cf. Mathieu and Zajac, 1990; Kumar et al., 1994; Geyskens et al.,
1996; De Ruyter et al., 2001). An affective committed partner desires to continue its
relationship because it likes the partner and enjoys the partnership (Geyskens et al.,
1996). Affective commitment expresses the extent to which a party likes to maintain a
relationship with another party. This kind of commitment is based on a general positive
feeling towards the exchange partner. In this study we conceptualize ‘affective
commitment’ according to De Ruyter et al. (2001: 286) as the level to which “it is pleasant
working with our supplier, that is why we stay with our supplier”.
Calculative commitment, in contrast, is the extent to which an exchange partner perceives
the need to maintain a relationship. The relationship results from a ‘cold’ calculation of
costs and benefits (Geyskens et al., 1996). Therefore, calculative commitment refers to a
negative motivation for continuing a relationship. In line with De Rutyer et al. (2001: 286)
we conceptualize ‘calculative commitment’ as the level to which “there is just too much
time, energy, and expense involved in terminating our relationship with this supplier.”
214
variables operationalizations
product customization product specifications are tailored to the specific needs and wants of the
buyer
possibilities for standardization it is possible to use standardized specifications for comparable items,
working with less specific specifications
competence trust * we believe that the supplier will keep his promises and agreements
goodwill trust* we believe that the supplier will not misuse his position and actually takes
our interests into consideration
calculative commitment ** we are doing business with this supplier, mainly because too much time,
energy and expense would be involved in terminating the relationship
with the supplier
affective commitment ** we are doing business with this supplier, mainly because it is pleasant
working with the supplier
All variables are measured on a 5-point Likert scale (agree - disagree).
* cf. Ganesan (1994), Kumar et al. (1995), Nooteboom (1996), Doney and Cannon (1997), De Ruyter et al.
(2001): ‘competence trust’ refers to the perceived credibility, the words and statements of the supplier can be
relied on, the supplier performs according to agreement; ‘goodwill trust’ refers to the benevolence, the
supplier is genuinely interested in the partner’s welfare, the belief that the supplier will not choose to employ
opportunities for defection.
** cf. Geyskens et al. (1996), De Ruyter et al. (2001): ‘calculative commitment’ refers to a negative motivation
to continue a relationship (‘need to maintain a relationship’), ‘affective commitment’ refers to a positive
motivation to continue a relationship (‘like to maintain a relationship’).
6.7 Methodological concerns

As in all empirical studies, there are concerns with respect to the methodology. In our
study there are issues concerning the validity and the non-repsonse.
6.7.1 Validity
Gellner (1978) characterizes the role playing method as “passive” and cautions for
experiments where respondents have no prior experience with the role that they are
playing. In such circumstances the quality of data and therefore the overall validity are
questionable. Given that the respondents in our study should be actual purchasing
professionals, lack of identification with the role will not be a concern in the research. In
addition, respondents are to mark their job title, which allows for assessing the
proportion of purchasing professionals.
Another, more serious concern of this study might be the lack of identification with the
scenarios in the experiment. To ensure the validity, a rating will be obtained for assessing
the recognizability of the scenarios. To avoid the gathering of unreliable data, for each
scenario respondents are asked to assess the level to which they recognize a familiar
situation. Analyzing the data, respondents with low scores on ‘recognizability’ will be
removed from the data base.
Measurement should ensure the content validity. In this study, content validity is based
on grounding the survey questions in the literature and on eliciting the expert opinion of
knowledgeable researchers. Moreover, the content validity is also based on the case
studies and the pilot study, interviewing potential respondents. All this to ensure that the
items in the survey adequately cover the domain of the subject. In other words, to ensure
that we are actually measuring what we are supposed to measure.
6.7.2 Buyer’s perceptions
All questions are answered by the buyer and from the buyer’s perspective. Assessing
respondent’s perceptions is the most meaningful way to measure variables, since decisions
will be made on managers’ perceptions of the relevant conditions (McCutcheon and
Stuart, 2000). This is the main argument for measuring perceptions. We posit that all
decisions are made on the basis of interpretations and perceptions of conditions and
circumstances. Studies focussing on environmental uncertainty, as perceived by decision
makers, have generally made use of subjective indicators of the environment (Aldrich
and Mindlin, 1978). As Bacharach and Lawler (1981) make clear, decision-makers’
perceptions of dependence play a large part in determining their response to a situation
of dependence. A ‘reputational’ method of measuring ‘power’ and ‘dependence’ is one of
the most common measurement methods in organizational studies (Gaski, 1986). It can be
argued that the perception of power and dependence is tantamount to their existence.
By nature Purchasing has a boundary spanning role within organizations. This implies
that a great deal of relevant information at the supply side of organizations passes
through the hands of purchasing professionals. In terms of research strategy, the
perceptions of boundary spanning employees are most relevant in order to assess the
actual state of dependence within buyer-supplier relationships. An important objective of
the survey project is to describe, to amplify, and to explain the selection of differentiated
purchasing strategies. It is not about analyzing buyer-supplier relationships: the study
focuses on the buyer’s side of the relationship. What kind of (portfolio-based) purchasing
strategies are being developed and under which circumstances? To conclude, in line with
the objectives and the perspective of the overall study, it is more sensible to measure
buyer’s perception.
6.7.3 Non-response
A mail survey is very useful, gathering information from a relative large number of
respondents. The method allows researchers to obtain a large amount of information
from a large sample, gives respondents time to answer, allows respondents to remain
anonymous and helps reduce interviewer bias (Mangione, 1995). Response rates have
216
become one of the primary yardsticks for judging successful survey research. Frohlich
(2002) concluded that response rates are important for three main reasons:
1. When the percentage of non-respondents is high, there is real risk that the data will be
biased.
2. Many statistical tests require a suitable number of cases.
3. High response rates indirectly reflect the relevance and rigor of the study. Respondents
are more likely to return a questionnaire if they perceive that the study is important and
warrants cooperation.
Business researchers are always concerned with the issue of non-response. Erdogan and
Baker (2002) concluded that the most threatening issue for researchers conducting mail
surveys is non-response. In general, the response rate depends on the motivation of the
recipient to answer the questionnaire and to send it back. Literature review suggests
support for prior notification, anonymity, university sponsorship, stamped return
envelope, incentives, and follow-up questionnaires (see for instance Jobber and O’Reilly
(1996) who reviewed a number of techniques in postal surveys that could raise response
rates). Other factors, such as appeals, handwritten postscripts, personalization and cover
letters, were found to be ineffective. Erdogan and Baker (2002) found in an experimental
design of a mail survey, that the ‘original replacement condition’ was the most effective
follow-up technique (compared to: a photocopy replacement, a follow-up post card, and a
follow-up letter). Greer et al. (2000) studied the respondents’ willingness to respond to a
questionnaire in a business context. The results indicate that the content of the study is
the most important factor in stimulating response participation, followed by survey
sponsorship and postage paid reply envelopes. Pre-notification and follow-up are
considered the least important factors. It is advised that researchers pay attention to the
format of the questions and the length of the questionnaire. Obviously, shorter
questionnaires are likely to produce better response rates than longer ones, especially
because the questionnaires will be completed during company time.
It is recognized that the questionnaire is relatively long, containing more than 175
questions. The pilot test provided information on the actual time needed to complete the
questionnaire: an average of 35 minutes. As a result of the pilot study, corrective actions
were undertaken, to resolve unclarities and to reduce the time needed for the completion
of the questionnaire.
In general, the following measures were taken, aimed at reducing non-response:

- University sponsorship was emphasized.
- For the initial mailing, a letter of recommendation was added, written by the president of
the NEVI (the Dutch Association for Purchasing). The follow-up mailing was completed
with a letter of recommendation that was signed by the researchers, emphasizing
academic relevance and including a university logo. Both letters contained sincere and direct
requests for help and cooperation. In addition, the letters of recommendation emphasized
that the objective of the research was to investigate and explain the selection of
purchasing and supplier strategies. The letters deliberately did not reveal that one of the
objectives was to determine the use frequency and the use intensity of the portfolio tool,
because that might deter non-users from completing the questionnaire. Also, the
questions about the portfolio approach were deliberately put at the end of the
questionnaire, to avoid that people might decide not to respond, because they do not use
a portfolio approach.
- A summary of results was offered to the respondents, reporting on the main conclusions
of the study.
- The first 200 respondents are entitled to a free copy of a book on purchasing management
(‘Samenwerkend Ondernemen’).
- Business reply envelopes were used (freepost).
- The questionnaire was pre-tested to improve readability, question order and to improve
ambiguous questions.
- The format and lay out of the questionnaire were aimed at reducing the number of pages.
The questionnaire was printed on both sides of the paper, which reduced the total
number to 4 pages.
- Six weeks after the initial mail-out a replacement questionnaire was mailed to all non-
respondents (follow-up mailing).
- Four weeks after the follow-up mailing remaining respondents received an e-mail,
asking them for the third and last time to participate in the survey. A replacement
questionnaire was added as an attachment.
Even though pre-notification is likely to affect the response rate positively, in this study
respondents were not pre-notified because of time and financial constraints.
6.8 Results
After an extensive analysis of response and non-response, we will present the results of
the survey study. Consistently, we will focus on answering the research questions.
6.8.1 Response
In this research, the response rate is defined as the percentage of total usable
questionnaires returned by respondents (Wiseman and Billington, 1984). A total number
of 248 reactions were received of which 10 could not be used, because the questionnaires
were not completed. If at least 90% of the questionnaire was completed, a response was
seen as valid. This means that the total valid response of the survey is 238 to be used for
analysis. The sample consisted of 1,153 companies, resulting in a response rate of:
238 / 1,153 = 20,6%.
Chronologically, the following stages were undertaken, as shown in figure 6.1.
218
Week in 2002 Activity

Results
10 (March) first round of survey sent out by regular mail 121 valid responses
16 (April) second round of surveys sent out by regular mail to 109 valid responses
remaining respondents
20 (May) third round of surveys sent out by e-mail to remaining 8 valid responses
respondents with known addresses
22 (May) Mail box closed n = 238
FIGURE 6.1 Chronological overview of survey activities and results
More than 100 respondents contributed to the first and the second round of surveys. The
majority of respondents were collected by means of the postal mailings. In contrast, the
response to the electronic mailing was low, much lower than expected. An explanation
might be that the electronic mailing was the third and last attempt to get cooperation. In
addition, a number of purchasing professionals might object to unsolicited e-mail (or
‘spam’). This objection to unsolicited e-mail had been observed in other studies (Boyer et
al., 2002). A total number of 276 e-mails resulted in 8 valid responses (2.9%). In his
research Mol (2002) used the same NEVI industry database. He indicated that the
database should be updated, because a relatively large number of firms no longer resided
at the given address. Illustrative for our study is the percentage of e-mails that turned out
to be ‘undeliverable’: 21%. Fifteen reactions gave some insight in reasons for not
participating in the study: 12 indicated to have ‘no time’, 3 indicated that they did not
practice a purchasing function.
Low response rates are typical of industrial mail surveys and 20-30% response rates are
considered to be good in Europe (Erdogan and Baker, 2002). Given the enormous length
of the questionnaire (8 pages, 175 questions), the response rate should be considered as
rather high. For instance, Frohlich (2002) analyzed 233 Operations Management research
papers that were published over the last 12 years. Most frequently found was an average
rate of 20% (mode), and an arithmetic mean of 32%. He concluded that a survey should
not contain more than 125 items. Yu and Cooper (1983) found that a questionnaire length
of 40-50 items delivered the highest average response rate. If a survey is under 4 or 5
pages, then resistance will be lower and the response rates tend to be higher. Obviously,
these conditions were not met in the present case. Without doubt, the resistance of
respondents was considerably lowered by the free offers that were made. Over 78% of the
respondents were interested in the book on purchasing management and 86% indicated
that they would
like to receive a summary of the main findings. It can be expected that all the other
measures have contributed to the response rate, albeit on a lower level of influence
(sponsorship, letters of recommendation, appeals, pre-paid postage, formatting, lay-out).
Some 30% of the respondents bear ‘Director Purchasing’ as their job title. The other
purchasing professionals accounted for almost 60% of the respondents. They were
classified according to the NEVI-clustering (Wesselink et al., 2001): purchasing managers,
(senior) buyers and purchasing assistants. Remaining respondents were managers of
logistics, supply chain managers and a small group of ‘others’ (see figure 6.2). Based on
their job titles, the respondents can be considered as being well-informed about the
purchasing operation in their companies.
Job title Frequency Valid percent

Director Purchasing 70 30
Purchasing Manager 79 33
Senior Buyer 23 10
Purchasing Assistant 37 16
Manager of Logistics 10 4
Supply Chain Manager 4 2
Other 12 5
Total 235 100
FIGURE 6.2 Jobtitle of respondents

220
Industries Frequen- Percen- NEVI- Percen- Frequen- Percen-

cy in tage of mem tage of cy in tage
Dutch Dutch bership NEVI- sample of
industry * industry figures figures sample
electro technical industry 9,122 15.2 221 21.2 45 18.9
metal products industry 9,205 15.4 193 18.5 51 21.4
chemical industry 1,144 1.9 180 17.3 34 14.3
machine industry 5,158 8.6 164 15.8 30 12.6
wood, furniture or paper industry 12,636 21.1 70 6.7 17 7.1
metal basic industry 329 .1 59 5.7 10 4.2
means of transport industry 3,214 5.4 39 3.7 10 4.2
building materials, glass, pottery 2,365 3.9 34 3.3 4 1.7
instrument and optical industry 832 .1 28 2.7 7 2.9

textile industry, clothing industry 5,189 8.7 22 1.9 7 2.9
graphic industry 10,568 17.6 19 1.8 7 2.9
petroleum, coal processing 170 .0 12 1.2 4 1.7

industry
Total 59,932 100.0 1,041 100.0 100.0
Other 112 12
Total 1,153 238
* In 1999, source: http://www.cbs.nl/nl/cijfers/statline/index.htm
FIGURE 6.3 Industry-by-industry composition of the sample, NEVI-membership figures,

and the Dutch manufacturing sector at large
Figure 6.3 allows for a number of comparisons. Firstly, we could compare the structure of
the NEVI-database with the corresponding numbers in the Dutch manufacturing
industry. Obviously, some industries are over-represented, especially the chemical
industry and the machine industry, and to a lesser extent the electronic industry and the
metal products industry. Other industries are clearly under-represented, such as the
wood, furniture or paper industry and the graphic industry. The composition of the
NEVI database reflects the motivation of firms and purchasing managers from different
industries to join a professional association. Generally speaking, members of a
purchasing organization tend to be larger firms and firms that purchase more. This
proposition was confirmed by Mol (2002) who analyzed the NEVI-database and
concluded that NEVI-members are generally much larger and more purchasing intensive.
Secondly, another comparison is of importance, addressing the issue of the
representativeness of the sample. Considering the different types of industry, it is
concluded that the sample is quite representative for the structure of the sample frame
(NEVI-database). As a consequence, the sample contains some under- and over-
represented industries. However, the composition of the sample is to a large extend
representative for the NEVI-database. The same holds true for company size,
operationalized as the number of employees. A firm with fewer than 100 employees is
categorized as a SME (small and medium sized enterprise). In the sample 28.6% of the
companies are SME’s, the larger companies account for the remaining 71.4%. In Dutch
industry however, the proportion between SME’s and larger companies is 96.7% and
3.3% (CBS statistics, 1999). This picture confirms to the fact that larger companies are
more inclined to join a professional association than smaller companies. We will take this
into account, when interpreting the data, for instance with respect to the number of
portfolio users.
We have recognized the risks and limitations of non-response. If persons who respond
differ substantially from those who do not, the results do not directly allow one to say
how the entire sample would have responded (generalization). In this study a number of
measures have been taken in order to protect against non-response. Obviously, non-
response is an issue in this study. One way of dealing with non-response is the
comparison of results with known values for the population. This is what we have done
so far, comparing sample characteristics with known CBS statistics (namely, industry
type and company size). Another approach to the non-response problem is to estimate
non-response bias. Armstrong and Overton (1977) have introduced an extrapolation
method in which the data are classified into a first category of returned questionnaires
(first-wave, early respondents) and a second category of returned questionnaires (second-
wave, late respondents). To establish the presence of non-response, first-wave
respondents are compared with second-wave respondents on relevant variables. The
most common type of extrapolation is carried over successive waves of a questionnaire.
‘Wave’ refers to the response generated by a certain stimulus, for instance sending a
questionnaire by mail. Persons who respond in a later wave are expected to be similar to
non-respondents (Armstrong and Overton, 1977).
In this study the non-response issue is addressed by comparing 121 first-wave

respondents to 109 second-wave respondents. For a large number of variables we have
tested whether there are statistical significant differences between the first and second
wave. Chi-square tests were used for the background variables and the use-variables
222
which were measured on a nominal or ordinal scale. At the .05 level of significance, no
differences were found in:- job position,
- company size,
- turnover,
- use, and
- use intensity.
A chi-square test for the ‘type of industry’ variable was not permitted, due to the
presence of 12 cells that have an expected count of less than 5. However, we could
roughly say that the frequency distributions displays no important differences (see figure
6.4). The small differences in the (absolute) frequencies give the impression of uniformity.
In relative terms the second wave contains more firms from the metal products industry
and fewer firms from the textile and clothing industry.
Industries absolute frequencies relative frequencies
first wave second wave first wave second wave
electro technical industry 24 18 19.8 16.6
metal products industry 22 26 18.2 25.7
chemical industry 17 17 14.0 15.6
machine industry 14 15 11.6 13.8
wood, furniture or paper industry 11 6 9.1 5.5
metal basic industry 4 6 3.3 5.5
means of transport industry 5 5 4.1 4.6
building materials, glass, pottery 2 2 1.7 1.8
instrument and optical industry 5 2 4.1 1.8
textile industry, clothing industry 6 1 5.0 .9
graphic industry 3 2 2.5 1.8
petroleum, coal processing industry 1 3 .8 2.8
other industry 7 4 5.8 3.7
Total 121 109 100.0 100.0
FIGURE 6.4 Distribution of absolute and relative frequencies of the type of industries `
for the first and second wave of respondents
Subsequently a series of t-tests were done for a large number of variables that are
measured on an interval or ratio scale. First, we examined the variables that are
connected with the use of the portfolio analysis, with an exception for the variable
‘company size’ (see above). At the .05 level of significance, no differences were found in all
of the explanatory variables:
- purchasing share
- the position of purchasing (two items)
- purchasing’s professionalism (two items)
- orientation of purchasing (three items).
The same conclusion can be drawn for the variables which are operationalized to
measure the impact of portfolio use (e.g. the understanding of problems and possibilities
of purchasing and the development of differentiated strategies).
Second, a series of t-tests were used for the variables that are connected to the 9 scenarios.
There are 17 variables in the questionnaire for each scenario, resulting in a total number
of 153 variables. It was found that 92.2% of these variables showed no significant
differences between the first and the second-wave group. This means that there are
statistical significant differences in merely 12 out of the total number of 153 variables. On
closer consideration 6 of these 12 variables are related to scenario 7 which refers to
maintaining a good relationship with an excellent supplier in the strategic quadrant. The
data indicate that the second-wave respondents especially estimated the buyer’s
dependence to be lower than the first-wave respondents: an average of 4.06 versus 4.36
(on a 5-point Likert scale). This could explain the differences in some other variables
which are most likely related to the perceived buyer’s dependence (i.e. logistical
indispensability of the supplier, the need for the supplier’s technological expertise, the
level of product customization and the possibilities for standardization).
In conclusion, not many statistically significant differences were found between the first
wave and the second wave of respondents. This conclusion is valid for nearly all
variables: 4 background variables, 13 variables related to portfolio use and 141 (of 153)
variables that are associated with the 9 purchasing scenarios. On the assumptions that
late respondents are similar to non-respondents, this leads to the important conclusion
that the study does not suffer from non-response bias.
6.8.2 Explaining the use of the portfolio analysis: the second research question
In this section we will present the answers on the research questions with respect to the
second core issue of this study: “Which factors would explain the utilization of the
purchasing portfolio analysis?” At the end of chapter 3 we derived four research sub-
questions and a number of hypotheses, relating to the use of a portfolio approach by
purchasing professionals.
224
Research question (2)

How many firms actually use a purchasing portfolio approach?
The most straightforward answer to this research question is that in the survey 73.8% of
the respondents indicated that they have been using portfolio analysis in their
purchasing practice; the remaining 26.2% indicated otherwise. This means that almost 3
out of 4 respondents can be classified as ‘users’. Therefore, we must conclude that the use
in manufacturing firms is much higher than estimates for other industries (Kamann: 20%,
Boodie: 44%). This conclusion remains valid, taking differences in company size into
account. It was recognized that the proportion of Dutch manufacturing SME’s is much
higher than the proportion in our sample: 96.7% versus 28.6%. The use frequency of the
SME’s in the sample is 57.4%; for the larger companies a number of 80.5% is found.
Multiplying the population proportions with the average use of SME’s and larger
companies, we can calculate a weighted average use frequency
(96.7% x 57.4%) + (3.3% x 80.5%) = 57.7%.
The use frequency in the sample indicates the percentages of users (figure 6.5a). This
information could rise in importance if we break down the use percentage into three
categories, according to the intensity of use (see figure 6.5b). One might wonder whether
a portfolio use of ‘less than once a year’ could contribute to the daily practice of
purchasing. If we would exclude the respondents that do not use the portfolio on a
regular basis, another number will be found. Eliminating the incidental users (‘less than
once a year’), a regular use percentage of 53.6% remains. Within a total of 73.8% a
substantial heavy user group can be detected, representing almost 40% of all respondents.
use frequency
Yes 73.8%
No 26.2%
FIGURE 6.5a Use frequency (n = 237 respondents)
use intensity
less than once a year 20.3%
approximately once a year 28.3%
more than once a year 25.3%
FIGURE 6.5b Use intensity (n = 175 using respondents)

The second sub-question related to reasons for use:
Research question (2b)

Why is the purchasing portfolio approach being used in practice?
The users were asked what the most important reasons would be for using the portfolio
approach. Respondents were allowed to mark more than one of the prespecified reasons.
According to the users, portfolio analysis contributes to:
- the insights into suppliers and products: 58.3%
- purchasing cost savings: 38.9%
- the identification of problems and possibilities in purchasing: 50.3%
- the information for the development of differentiated
purchasing strategies: 70.9%
These findings seem to confirm, albeit to a limited extent, that the core purpose of
portfolio analysis is the development of differentiated purchasing and supplier strategies.
What is more is that the reason ‘development of differentiated strategies’ was seldom
combined with other reasons. It could be interpreted that the development-reason refers
to a separate category of respondents who value and emphasize this core purpose.
Summarizing the findings in order of frequency, the following reasons were directly
marked for the use of the purchasing portfolio approach:
1. the development of differentiated strategies
2. the gaining of additional insights and knowledge
3. the realisation of cost savings.
Following on the reasons for portfolio use, the satisfaction of users has been determined.
From the case studies it became clear that three different measurement methods were
being used in practice. Most handbooks and papers on the portfolio do not reveal how
the positions in the matrix could or should be determined. Sometimes a weighted factor
score method is assumed, which allows for a large number of factors to be dealt with in a
customized way. However, in our study it was found that the weighted factor score is the
least frequently-used measurement method. Kamann (2000a) assumed that practitioners
‘often’ use ‘the number of suppliers’ on the horizontal axis and ‘financial value’ on the
vertical axis. This corresponds to, what we have labeled, as the 1-by-1 method. Although
this method is quite ‘often’ used, it is not the most frequently-used method. Almost half
of the respondents preferred the consensus method. The following picture can be drawn:
- consensus method 47.7%
- 1-by-1 method 28.2%
- weighted factor scores 11.5%
- other method 12.6%.
Users of another method, generally indicated that they used a combination of two or
more methods, for instance, adding a consensus element in the 1-by-1 method.
226
In addition, users were to a great extent satisfied with their portfolio use, scoring an
average of 4.0 on a 5-point scale. This comes as no surprise because if professionals were
not satisfied they would not be using the portfolio tool. It is more remarkable that
between the three methods no statistically significant differences in satisfaction were
found. This finding tempts us to conclude that the measurement method has no
significant impact on the perceived satisfaction. In other words, apparently it does not
matter which method is being selected.
In a way, all the foregoing is related to the impact and results of portfolio use. Chapter 3
unfolded a line of reasoning, addressing the question ‘what could be expected from
portfolio use?’ The discussion has led to the formulation of two hypotheses, positing that
the use of the purchasing portfolio is positively related (A1) to the level of understanding
of the problems and possibilities of the purchasing function and (A2) to the extent to
which differentiated purchasing and supplier strategies are developed.
To test these hypotheses, two (independent sample) t-tests were run, with ‘use’ as
grouping variable (see figure 6.6). For both performance variables, the figure shows that
the mean scores of the users are higher than the mean scores of the non-users. Moreover,
the differences are both statistically significant at the p<0.05 level. This means a
confirmation of the two hypotheses: portfolio users report higher levels of understanding
and higher levels of differentiation in strategies.
user’s mean non-user’s mean t-value P-value

score score
understanding of problems and 3.63 3.28 2.24 .026

possibilities
differentiated purchasing and 3.44 2.87 3.35 .001

supplier strategies
FIGURE 6.6 Results of the t-tests for the direct impact of portfolio use
To conclude, comparing the average scores of users and non-users, statistically significant
differences were found at the p<0.05 level, therefore indicating that the portfolio
technique ‘delivers’ what it is supposed to:
- it adds to the understanding of purchasing problems and possibilities, and
- it assists the purchasing professional in developing differentiated purchasing strategies.
This means that we can accept hypotheses A1 and A2 which assumed a positive
relationship between the use of the purchasing portfolio analysis and the two impact
variables.
With these conclusions, we have answered the 2b-research question. From our analysis it
becomes clear that there are good reasons for using the purchasing portfolio analysis in
practice. However, there are professionals who have reasons for not using the portfolio.
The non-users were asked what the most important reasons would be for their point of
view. Respondents were allowed to mark more than one of the pre-specified reasons. A
distinction can be made between company-internal reasons and the perceived limitations
of the portfolio analysis:
Company-internal reasons:
- takes too much time 39%
- absence of required knowledge 27%
- unwillingness of purchasing director 3%
The case studies pointed at the importance of (top) management support for the adoption
of portfolio analysis. The results of the survey indicate that a possible negative attitude of
the purchasing director does not impede the adoption of the portfolio in practice. Only
3% of the non-users made reference to the unwillingness of their purchasing director.
However, this figure should be interpreted with some prudence, because it is not clear
whether the other non-respondents (a) do not perceive ‘unwillingness’ or (b) are not
aware of their director’s opinions on this matter. Another cautionary note should be
added. Professionals who claim that the use of the portfolio analysis takes too much time,
might not be able to provide a reasonable assessment of the required time, due to a lack
of first-hand experience. In addition, another reason for caution might be that a lack of
time in some cases should be interpreted as ‘has no priority’ or as ‘lack of knowledge’.
However, restricting ourselves to the numbers, the conclusions are inevitable: lack of time
and lack of knowledge are perceived as the most important company-specific reasons for
not using the portfolio.
Boodie (1997) assumed that larger companies would probably employ higher educated
purchasers, in comparison with smaller companies. The result of this difference would
logically be that smaller companies are less capable of using the portfolio analysis.
However, this assumption can not be confirmed by our study. We found no relationship
between company size and the lack of knowledge, as a reason for non-use of the
portfolio.
Perceived limitations of the portfolio analysis:

- produces unusable results 15%
- includes not enough factors 11%
- too simple for complex decisions 16%
Taking a closer look at the limitations, it turned out that 31% of all non-users have
marked one or more of the possible motivations for their non-use.
To conclude, answering the 2c-research question (‘why not?’), the main reasons for not
using the portfolio refer to a lack of time and/or knowledge (company-specific) and to the
perceived limitations of the tool.
228
6.8.3 Users versus non-users
Based on literature study a number of variables and factors were found, to describe and
explain the differences between users and non-users. In this section, we will address
research question 2d: What are the differences between users and non-users of a
purchasing portfolio approach? First, we ran a number of statistical tests to determine
whether the differences are statistically significant. The variable ‘company size’ is
measured on an ordinal scale, therefore a chi-square test is used. Independent sample t-
tests are used for the other explanatory variables. In each case the differences are tested
separately. Second, to gain further insights in the matter, we ran a logistic regression
analysis with ‘use’ as the (dichotomous) outcome variable and a number of continuous
and categorical predictor variables.
The company size has been operationalized by the number of employees, measured on an
ordinal scale. This variable has five categories, the first of which coincides with the
official CBS-definition of SME’s (less than 100 employees). In addition and in agreement
with the foregoing, a recoded variable is constructed: a dichotomous variable that
classifies companies either as SME’s or as larger companies. Both variables are used in
the chi-square analysis, to detect differences between the use and the non-use group. In
both cases the value Pearson chi-square was statistically significant at the .000 level,
indicating significant differences (see figure 6.7). The conclusion is: the distribution of
respondents on ‘company size’ is disproportionate to the ratio between users and non-
users in the sample. More specific, we accept hypothesis A3 which stated that ‘the
company size of users will be larger than the company size of non-users of the
purchasing portfolio analysis’. The positive relationship between company size and
portfolio use is in line with Boodie’s findings in his 1997-study. Larger companies are
more likely to deal with a larger number of products, more suppliers and more complex
purchasing situations. Management tools, such as the portfolio analysis, will soon be
beneficial.
Value degrees of freedom P-value
company size on a 5-category ordinal scale
Pearson chi-square 23.794 4 < .001
contingency coefficient .302 < .001
company size:
SME-larger companies
Pearson chi-square 13.419 1 < .001
contingency coefficient .231 < .001
N of valid cases 237
FIGURE 6.7 Results of the chi-square tests: company size x use
The explanatory variables were subjected to a number of successive t-tests, summarized

in figure 6.8. The direction of the differences between user’s and non-user’s mean scores
on the explanatory variables are fully in line with the prior expectations. However, not all
differences proved to be statistically significant. We found no significant differences for
three variables:
- the extent to which purchasing reports to the top management (position of purchasing)
- the extent to which purchasing aims at collaboration with suppliers (orientation)
- the extent to which purchasing aims at tough negotiations with suppliers in pursuit of
the lowest prices (orientation).
This means that two hypotheses that correspond with ‘orientation of purchasing’ have to
be rejected: collaboration and negotiation/lowest prices (A7a and A7b). At this point the
results are inconclusive with respect to hypothesis A5 which claims that ‘the position of
purchasing is better at companies where the purchasing portfolio analysis is used, than at
non-using companies’. We did not find significant differences between users and non-
users on the variable ‘reports to top management’.
On the other hand, we did find significant differences for the other explanatory variables
in the use model. This means that we found a confirmation for the corresponding
hypotheses that can not be denied:
- A4: purchasing share of users will be higher than the purchasing share of non-users of the
purchasing portfolio analysis,
- A6: the professionalism of purchasing is higher at companies where the purchasing
portfolio analysis is used, than at non-using companies and
- A7c: the users of the purchasing portfolio analysis will be less involved with clerical and
operational activities than non-users.
230
In addition, more users reported that purchasing contributes to the competitive position of
the firm than non-user. Apparently, there are significant differences with respect to this
element of the ‘position of purchasing’-construct.
user’s mean non-user’s mean t-value P-value

score mean score difference
purchasing share 54.9 48.8 6.1 2.20 .029
contributes to competitive 3.83 3.42 .41 2.64 .009

position
reports to top management * 4.03 3.66 .37 * 1.74 .083
skills for cross functional teams 3.64 3.18 .46 3.23 .001
skills for developing strategies 3.57 3.19 .38 2.86 .005
orientation on collaboration * 3.63 3.48 .15 * 1.14 .254
orientation on negotiation and 3.11 3.31 - .19 * - 1.31 .192

lowest prices *
orientation on clerical and 2.76 3.18 - .41 - 2.53 .012

operational activities
* not significant at p<0.05
FIGURE 6.8 Results of the t-tests for the differences between users and non-users
(n=236)
On the whole the results of the tests confirm our prior expectations and hypotheses.
Users of the portfolio are employed with larger companies and have to deal with higher
purchasing shares. All in comparison with non-users. We presented the respondents with a
number of propositions that refer to the position of purchasing in the company, the
professionalism of purchasing and the orientation of purchasing. Comparing the results,
it was found that users of the portfolio reported that they:
- contribute more to the competitive position of their company
- have more skills in working in cross functional teams and in developing purchasing and
supplier strategies
- are less involved in clerical and operational activities.
In other words, the user group distinguishes itself in a positive way from the non-user
group. Their position is better, regarding their contribution to the competitive position of
the company. Their professionalism is related to important attributes, namely the skills of
professional purchasers to work in cross-functional teams and to develop strategies.
Therefore, it is not surprising that these professionals are not (heavily) engaged in clerical
activities and day-to-day routine work.
With the separate tests we determined what the significant differences are between users
and non-users. In addition a logistic regression analysis was run, which is multiple
regression but with an outcome variable that is a categorical dichotomy and predictor
variables that are continuous or categorical. With logistic regression we can predict to
which of two categories (users and non-users) a respondent is likely to belong, given
certain other information. The analysis can be used to establish which variables are
influential in predicting the correct category. Answers can be found to the question:
which variables are appropriate for predicting whether a respondent (purchasing
professional) will use the portfolio approach or not?
The selected method of conducting the logistic regression is Backward Stepwise. This
method starts with all predictors included. The computer then tests whether any of the
predictors can be removed from the model without having a substantial effect on the fit
of the model. The selection of this stepwise method is advised when used in situations in
which no previous research exits and in situations where one merely wishes to find a
model to fit the data. These conditions apply to our situation.
The main results of the logistic regression are shown in figure 6.9. A number of variables
were left out of the final equation. Their inclusion would not significantly affect the
predictive power of the model. The following variables were taken out (in order of
removal):
- orientation on clerical and operational activities
- purchasing reports to top management
- orientation on collaboration with suppliers
- skills for working in cross functional teams
- orientation on tough negotiation and lowest prices.
232
main statistics chi-square df P-value
omnibus test of model coefficients 25.208 4 < .001
Nagelkerke R square .226
correctly predicted group membership 78.5%
variables in the equation coefficients: df Wald statistic: P-value

Exp(B) chi-square
company size 2.423 1 4.085 .043
purchasing share 1.026 1 4.364 .037
contributes to competitive position 1.567 1 4.581 .032
skills for developing strategies 1.811 1 5.497 .019
constant .031 1 7.439 .006
FIGURE 6.9 Results of the logistic regression with ‘use’ as dichotomous outcome
variable
The final model includes only variables with a significant contribution to the explanation
and prediction of the outcome variable. The crucial statistic is the Wald statistic which has
a chi-square distribution which tells us whether the coefficients for the predictors are
significantly different from zero. Even more crucial to the interpretation of the results are
the values of the coefficients Exp (B). The value of Exp (B) is an indicator of the change in
odds resulting from a unit change in the corresponding predictor variable. If the value is
greater than 1 then it indicates that as the predictor increases, the odds of the outcome
occurring (user or non user) increase. The higher the value of Exp (B), the larger the
contribution to the prediction of the outcome variable. Conversely, a value less than 1
indicates that as the predictor increases, the odds of the outcome occurring decreases. The
remaining variables in the model are, in ascending order of impact on the outcome
variable:
1. purchasing share
2. contribution to competitive position
3. skills for developing strategies
4. company size (dichotomous variable).
For a further assessment of the results, the main statistics have to be evaluated. The
Nagelkerke R-square of 226 points at a relatively low percentage of explained variance. The
fit of the model is therefore rather limited. The predicted group membership is a specific
result of logistic regression, indicating the overall accuracy of the model. The predicted
group membership predicts to which of the two categories (users and non-users) a
respondent is most likely to belong, based on the model. We found a correctly predicted
group membership of 78.5%. This might be interpreted as a rather small improvement,
compared to the classification of the beginning block (77.8%), which is the result of a
model where only the constant is included (step 0). However, the value of the model chi-
square statistic is statistically significant. The chi-square measures the difference between
the model and the model when only the constant is included. This means that overall the
model is predicting use and non-use significantly better than a model with only the
constant included.
a,b
+
company size
a,b
+
purchasing share
a, b
+
contribution to
competitive position
position of
purchasing +
reports to top
c
management
portfolio
use
+
skills for cross
a
functional teams
purchasing’s
professionalism +
skills for developing
a, b
strategies
+
orientation on
c
collaboration
orientation of orientation on
-
c
purchasing negotiation and prices
orientation on clerical and

-
a
operational duties
Significant at the p<0.05 level:

a b c
in the separate t-tests, in the logistic regression, not significant
FIGURE 6.10 Signs and significance in the use model

234
To conclude, the statistical tests for the separate variables have resulted in a list of 6
variables that make a significant difference between users and non-users. In addition, the
logistic regression resulted in a smaller list of 4 variables, predicting use and non-use as
outcome variables. These results underscore the most important conclusions of this
section: users contrast sharply with non-users of the portfolio approach for their scores on
a number of variables (see figure 6.10). Company size and purchasing share are of a more
general nature; other variables make the distinction in favor of using purchasing
professionals who contribute more to the competitive position of the company and have
more skills for the development of purchasing and supplier strategies.
6.8.4 Answers to the fourth research question: conditions for the selection of
strategies
In section 6.7. we have raised the issue of validity. Respondents were asked to assume the
role of a purchasing expert in their own company and to evaluate questions relating to
different purchasing scenarios. To avoid respondent bias, due to a lack of recognition of
scenarios, a total number of 22 respondents are removed from the database. These
respondents scored an average recognition of lower than 3 on a 5-point Likert scale,
resulting in an effective response of 216 respondents. By excluding respondents with
relatively low recognition of the scenarios, the validity is enhanced. Obviously a small
number of 22 does not allow for extensive analysis. On the face of it, there is a near
resemblance between the characteristics of the non-recognition group and the other
respondents. For instance, the proportions of use and use intensity of portfolio analysis
are quite similar to the frequencies found in the overall analysis. From the total of 22
companies, 8 are SME’s and 14 are larger companies. A 36.4% of SME compares too the
28.6% of all SME-respondents. Apart from that, we found no significant differences
between the recognition scores of users and non-users of the portfolio analysis.
By comparing two relationship dependence profiles of two purchasing scenarios in the

same quadrant, it is possible to test whether these profiles differ significantly. In section
6.3 we have formulated 6 hypotheses, concerning differences between the relationship-
dependence profiles in the matrix. The pairwise comparisons of relationship profiles
indicated that all differences are significant at the p<0.05 level, using the adjusted
significance levels of the Greenhouse-Geisser test and the Huynh-Feldt test. This means
that the hypotheses C1 until C6 are confirmed by our study: there are statistically significant
differences between the profiles that fit alternative strategic choices in the same portfolio
quadrant.
In the following sections we will formulate the answers to the fourth research question:
under which conditions are the various portfolio-based strategies pursued in purchasing
management?
For each quadrant we will look for relatively high and relatively low scores on variables
that therefore shape the corresponding relationship-dependence profiles. Subsequently,
the variables in the profiles will be further analyzed, indicating which variables attribute
to the overall difference between the relationship-dependence profiles. The differences
between profiles on this variable-level will be used to explain the selection of strategies.
Figure 6.11 provides an overview of the means and the standard deviations of the 14
dependent variables in the 9 scenarios. Due to the missing values a number of 171 cases
could be used for the pairwise comparisons of variables.
236
bottleneck non-critical leverage strategic
scenarios #1 #2 #3 #4 #5 #6 #7 #8 #9
variables
logistical 4.26 4.24 3.22 2.96 3.93 4.15 4.63 4.52 4.48
indispensability (.929) (.801) (1.286) (1.250) (1.032) (.868) (.541) (.597) (.636)
supplier’s 3.24 3.09 1.59 1.76 2.40 3.44 4.08 3.78 3.56
technological (1.146) (1.111) (.717) (.911) (1.120) (1.163) (.868) (1.119) (1.085)
expertise
alternative 2.94 3.60 4.71 4.39 4.56 3.79 2.81 2.63 3.26
suppliers (1.223) (1.055) (.637) (.863) (2.469) (.915) (1.168) (1.250) (1.104)
switching cost 3.80 3.25 1.74 1.73 2.22 3.14 4.08 4.12 3.94
buyer (1.173) (1.162) (.870) (.866) (.973) (1.097) (1.074) (1.067) (.980)
financial 2.71 2.74 2.62 2.07 3.57 3.85 3.95 3.24 3.21
magnitude (1.191) (1.104) (1.184) (1.049) (.964) (.797) (.761) (1.141) (.971)
buyer’s 2.30 2.13 1.63 1.55 2.28 2.95 3.32 2.57 2.59
technological (1.100) (1.049) (.781) (.729) (1.013) (1.126) (1.038) (1.095) (1.039)
expertise
alternative 3.75 3.75 4.43 4.37 3.89 3.40 3.27 3.67 3.51
buyers (1.142) (1.137) (.751) (.840) (.927) (1.071) (1.079) (1.006) (1.002)
switchting cost 2.29 2.27 1.89 1.73 2.66 3.02 3.36 2.84 2.82
supplier (1.157) (1.084) (.997) (.887) (1.069) (1.071) (1.131) (1.108) (1.044)
product 3.56 3.49 1.97 2.09 2.91 3.70 4.20 3.81 3.80
customization (1.298) (1.195) (1.155) (1.192) (1.264) (1.143) (.926) (1.139) (1.090)
standardization 2.47 3.22 3.70 3.41 3.02 2.74 2.04 2.23 2.32
possibilities (1.276) (1.167) (1.217) (1.268) (1.225) (1.120) (.984) (1.014) (1.038)
competence 3.40 3.53 3.82 3.66 4.05 4.16 4.25 2.76 2.50
trust (.968) (.856) (.731) (.855) (.688) (.627) (.631) (.968) (.935)
goodwill trust 3.19 3.29 3.41 3.44 3.91 4.04 4.08 2.67 2.49
(.994) (.878) (.944) (.964) (.813) (.754) (.755) (.945) (.910)
affective 2.65 2.73 3.27 3.13 3.08 3.47 3.60 1.87 1.99
commitment (1.048) (1.006) (1.106) (1.093) (1.037) (1.013) (1.049) (.789) (.840)
calculative 3.11 2.88 2.22 2.25 2.16 2.82 3.25 3.67 3.33
commitment (1.208) (1.208) (1.032) (1.032) (1.002) (1.200) (1.270) (1.232) (1.147)
FIGURE 6.11 Means and (between parentheses) standard deviations of the variables
(n=171)
6.8.4.1 Profiles in the bottleneck quadrant
In the bottleneck quadrant the logistical indispensability is very high, considering the
absolute values (4.26 and 4.24). Apparently, the supply risk and the bottleneck positions
are caused by logistical dependencies. The poor power position of the buyer is further
enhanced by the large number of alternative buyers and the relatively low switching cost
for the supplier. Buyers and suppliers have a limited need for each other’s technological
expertise.
In descending order, the following significant differences (p <0.05) were found in the
bottleneck quadrant:
possibilities for standardization - .75

alternative suppliers -. 66
switching cost buyer +. 55
Mean differences are calculated as (#1) – (#2).

Scenario #1 = keep safety stocks
Scenario #2 = find other solutions: decomplex the product, if necessary find new supplier
238
low medium high mean

difference
1 2 3 4 5 (#1) – (#2)
1a
logistical -.02
indispensability
1b
supplier’s .15
technological
expertise
1c
alternative suppliers -.66
1d
switching cost buyer .55
2a
financial magnitude -.03
2b
buyer’s technological .07
expertise
2c
alternative buyers .00
2d
switching cost .02
supplier
3a
product .07
customization
3b
standardization -.75
possibilities
4a
competence trust -.13
4b
goodwill trust -.10
4c
affective commitment -.08
4d
calculative -.10
commitment
Scenario #1 = keep safety stocks

Scenario #2 = find other solutions: decomplex the product, if necessary with new supplier
multivariate test of significance: F-value = 6.04, P-value < .001
FIGURE 6.12a Relationship-dependence profiles in the bottleneck quadrant

Analysis of differences shows that the possibilities for standardization and the availability
of alternative suppliers are the most salient differences between the two scenarios. The
significant lower level of switching cost in scenario #2 indicates that the buying company
has made less relationship-specific investments, resulting in lower barriers for making
the switch to a routine position in the matrix. Not surprisingly, the buyer would consider
a strategic move to the routine quadrant if he experiences less dependence on the
supplier. What is remarkable is that none of the other 11 factors in the relationship-
dependence profiles show any significant differences. For instance, trust and
commitment have no bearing on the decision-making process. Key parts are played by a
very limited number of factors (see figure 6.12a.)
Main implication
In general the switch from ‘bottleneck’ into ‘routine’ in the matrix is desirable, but not
always possible (scenario #2). Such a strategic move is only an option if the conditions
concerning the possibilities for standardization and the availability of alternative suppliers
are satisfied. Consequently, the switch is associated with lower switching cost for the buyer.
6.8.4.2 Profiles in the non-critical quadrant
In general we found a relatively high number of alternative suppliers in the non-critical

quadrant, in fact for both scenarios. From the supplier’s perspective the alternative
number of buyers is also high. Buyers and suppliers do not need each other’s technical
expertise for non-critical items. For both partners relatively low switching costs were
found, indicating a market exchange relationship (cf Bensaou, 1999). Considering the
specifications it is clear that the product customization is of a low level, with relatively
many possibilities for standardization. This conclusion holds for both scenarios (see
figure 6.12b).
Pooling of requirements offers more possibilities and implies lower costs and more
efficiency for relatively low value items. However, within organizations there might be
good reasons for the individual ordering of certain items. Alternatively, there might be
ineffective barriers to purchasing efficiency.
non-critical quadrant:
financial magnitude +. 55
alternative suppliers +. 32

Scenario #3 = pooling of requirements
Scenario #4 = individual ordering
240

difference
1 2 3 4 5 (#3) – (#4)
1a
logistical .26
indispensability
1b
supplier’s -.17
technological
expertise
1c
alternative suppliers .32
1d
switching cost buyer .01
2a
financial magnitude .55
2b
buyer’s technological .08
expertise
2c
2d
switching cost .16
supplier
3a
product -.12
customization
3b
standardization .31
possibilities
4a
competence trust .16
4b
goodwill trust -.03
4c
affective commitment .14
4d
calculative -.03
commitment
Scenario #3 = pooling of requirements

Scenario #4 = individual ordering
FIGURE 6.12b Relationship-dependence profiles in the non-critical quadrant

The results of the study clearly point at the importance of differences in the financial
magnitude of the product. The key factor, discriminating between a pooling strategy and
a strategy of individual ordering, turns out to be the financial magnitude of the
transactions for the supplier. In addition differences were found in the availability of
alternative suppliers and the connected level of the buyer’s dependence. The other 12
factors in the profiles do not seem to contribute significantly to the selection of strategic
choices in the routine quadrant.
Main implication
From a purchasing point of view pooling of requirements (scenario #3) is preferable to
individual ordering (scenario #4). However, the switch towards the leverage quadrant is
only feasible if there is a sufficient volume of trade (financial magnitude) from the
supplier’s perspective and an ample availability of alternative suppliers.
6.8.4.3 Profiles in the leverage quadrant
Perhaps surprisingly, the logistical indispensability of leverage products turns out to be

rather high. These products are obviously important for an undisturbed production flow.
This finding is underscored by the relatively high levels of trust (competence and
goodwill trust). However, the buyer’s dependence is still low, caused by a large number
of alternative suppliers, accompanied by a low level of switching cost. The buying
company typically has not invested much in its relationships with leverage suppliers.
It was clear in advance that we could expect many differences, comparing scenario #5
(maintaining a convenience partnership) and scenario #6 (developing a strategic
partnership). This is confirmed in our study: there are far more differences than
similarities in relationship profiles. Exceptions are competence trust and goodwill trust,
which do not attribute to the overall difference between the two profiles. The average
scores on these variables are relatively high for both scenarios (4 on a 5-point scale). Trust
is equally important in the leverage quadrant, probably due to the logistical
indispensability. Although the switching costs are low, a leverage position does not
imply that the qualities and competences of the selected supplier are unimportant. Both
scenarios show comparable high levels of logistical indispensability and financial
magnitude.
leverage quadrant:
supplier’s technological expertise -. 104

switching cost buyer - .92
product customization - .78
alternative suppliers +. 77
buyer’s technological expertise -. 67
242
calculative commitment -. 67
alternative buyers +. 49
affective commitment -. 39
switching cost supplier -. 36
financial magnitude -. 28

Scenario #5 = maintaining a partnership of convenience
Scenario #6 = development towards a strategic partnership

difference
1 2 3 4 5 (#5 – #6)
1a
logistical -.22
indispensability
1b
supplier’s -1.04
technological
expertise
1c
alternative suppliers .77
1d
switching cost buyer -.92
2a
financial magnitude -.28
2b
buyer’s technological -.67
expertise
2c
2d
switching cost -.36
supplier
3a
product -.79
customization
3b
standardization .28
possibilities
4a
competence trust -.11
4b
goodwill trust -.1
4c
affective commitment -.39
4d
calculative -.70
commitment
Scenario #5 = maintaining a partnership of convenience

Scenario #6 = development towards a strategic partnership
FIGURE 6.12c Relationship-dependence profiles in the leverage quadrant

244
The main question in the leverage quadrant is: when do we want to maintain a
convenience partnership and when do we want to develop the relationship into a
strategic partnership? Focussing on the largest differences, it is found that the needed
supplier’s technological expertise is significantly higher in scenario #6, in combination with
higher switching costs for the buyer. The availability of alternative suppliers is high, but
not as high as in scenario #5. The choice of entering a strategic partnership is stimulated
by the fact that the buyer’s dependence is already rather high. However, the supplier’s
dependence is higher as well in scenario #6. Therefore, before entering in a real strategic
partnership there should (already) be a power balance, which is expressed by comparable
levels of buyer’s and supplier’s dependence. This supports the proposition that, from the
buyer’s perspective, there is trade-off: a higher dependence is only acceptable if the
supplier’s commitment is also high, supplying technological expertise and a customized
product.
Main implication
A stay in the leverage quadrant is usually considered as the most preferable position in
the matrix. Therefore, the switch to the strategic quadrant is likely to be the exception to
the rule. The main condition for engaging in a partnership with a supplier is related to
already higher levels of supplier’s dependence and especially buyer’s dependence. In
addition, there is more (affective and calculative) commitment towards the future partner.
6.8.4.4 Profiles in the strategic quadrant
The three scenarios in the strategic quadrant all indicated a very high buyer’s
dependence. All constituting variables contribute to this dependence situation for the
buyer:
- high logistical indispensability,
- high need of supplier’s technical expertise,
- low number of alternative suppliers, and
- high level of switching costs.
In addition, product customization is relatively high, with relatively few possibilities for
standardizations in all three scenarios. In contrast, more differences between the
relationship-dependence profiles were found for the determinants of supplier’s
dependence and the variables in the ‘trust and commitment’-cluster (see figure 6.12d).
low medium high mean mean mean

diff. diff. diff.
1 2 3 4 5 #7 - #8 #7- #9 #8 - #9
1a
logistical .11 .15 .04
indispensability
1b
supplier’s .30 .52 .22
technological
expertise
1c
alternative suppliers .18 -.45 -.63
1d
switching cost buyer -.04 .14 .18
2a
financial magnitude .71 .74 .03
2b
buyer’s technological .75 .73 -.02
expertise
2c
alternative buyers -.40 -.24 .16
2d
switching cost .52 .54 .02
supplier
3a
product .39 .40 .01
customization
3b
standardization -.19 -.28 .09
possibilities
4a
competence trust 1.49 1.75 .26
4b
goodwill trust 1.41 1.59 .18
4c
affective commitment 1.73 1.61 -.12
4d
calculative -.42 -.08 .34
commitment
Scenario #7 = maintaining a strategic partnership; Scenario #8 = accepting a

locked-in partnership; Scenario #9 = finding a new supplier
multivariate tests of significance: F-value = 29.28, P-value < .001 (#7 vs. #8),
F-value = 35.71, P-value< .001 (#7 vs. #9), F-value = 5.25, P-value< .001 (#8 vs. #9)
FIGURE 6.12d Relationship-dependence profiles in the strategic quadrant

246
Comparing scenario #7 (strategic partnership) with scenario #8 (locked-in ‘partnership’),

the largest differences were found in trust and commitment. In comparison with other
situations, extreme differences were revealed. Beforehand, we might have expected that
the main differences would refer to differences in buyer’s dependence and/or supplier’s
dependence. Nonetheless, affective commitment, competence and goodwill trust clearly
exceed these clusters of explaining variables. On a lower level of importance, we have
found significant differences in the supplier’s dependence between a strategic partnership
and a locked-in partnership, indicating that a strategic partnership requires investments
and commitments from the supplier’s side. In comparison, it is perceived that suppliers
in a strategic partnership experience a higher financial magnitude, higher switching cost,
a small number of alternative buyers and a higher need for the buyer’s technical
expertise. In contrast, hardly any significant differences were found in the buyer’s
dependence. In both scenarios, we found a relatively high dependence of the buyer on the
supplier.
In descending order, the following significant differences (p <0.05) were found, between
scenario #7 and scenario #8:
affective commitment + 1.73

competence trust + 1.49
goodwill trust + 1.41
buyer’s technological expertise + .75
switching cost supplier +. 52
calculative commitment - .42
product customization + .40
alternative buyers - .39
supplier’s technological expertise +.29

Scenario #7 = maintaining a strategic partnership
Scenario #8 = accepting a locked-in partnership
Broadly speaking, we have come to similar conclusions for the comparison of #7 (strategic
partnership) with scenario #9 (finding a new supplier). An exception has to be made for
the lower level of buyer’s dependence in case of scenario #9, which is associated with
differences in the need for the supplier’s expertise and the availability of alternative
suppliers, all compared to scenario #7. Evidently, the main point of discrimination is the
difference between a satisfactory and an unsatisfactory relationship, as in scenario #8 and

scenario #9. The quality of a true partnership is expressed by much higher average scores
on:
- goodwill trust: the supplier does not misuse his position and actually takes our interests
into consideration.
- competence trust: the supplier keeps his promises and agreements
- affective commitment: we are doing business with this supplier, mainly because it is
pleasant working with the supplier
Unsatisfactory relationships with suppliers in the strategic quadrant are associated with
‘partners’ who are less dependent on the buying company, compared to satisfactory
relationships (from the buyer’s perspective).
In descending order, the following significant differences (p <0.05) were found, between
scenario #7 and scenario #9:
competence trust + 1.74

affective commitment + 1.61
goodwill trust + 1.60
buyer’s technological expertise + .73
switching cost supplier +. 54
supplier’s technological expertise +. 52
alternative suppliers - .45
product customization + .41
possibilities for standardization - .28

Scenario #7 = maintaining a strategic partnership
Scenario #9 = terminating partnership, finding a new supplier
Main implication
From the buyer’s perspective, satisfactory partnerships in the strategic quadrant are
distinguished from unsatisfactory ‘partnerships’ by higher levels of supplier’s dependence,
trust and commitment.
Finally, comparing scenario #8 (locked-in ‘partnership’) with scenario #9 (finding a new

supplier), we found that the corresponding profiles have many points in common. In fact,
there are far more similarities than differences. Close collaboration is out of the question,
in both scenarios. Affective commitment is extremely low, indicating that the relationship
is not based on a spirit of cooperation: it is not pleasant working with the supplier. On
the other hand, switching costs are high. A main difference between the two profiles is
248
the availability of alternative suppliers which is much higher in case of scenario #9 where
an unsatisfactory business partner is replaced by a new supplier. The availability of
alternatives in scenario #9 results in a lower level of buyer’s dependence. For scenario #8
the average score on calculative commitment is relatively high: terminating the relationship
would take too much time, energy, and expense. Perhaps surprisingly, there is no
significant difference between the switching costs of the buyer, which are relatively high,
in both scenarios. Finally, terminating a partnership, as in scenario #9, is associated with
a lower level of competence trust.
The following significant differences (p <0.05) were found, between scenario #8 and
scenario #9:
alternative suppliers - .63
calculative commitment + .34
competence trust + .26
Scenario #8 = accepting a locked-in partnership
Scenario #9 = terminating partnership, finding a new supplier.
Main implication
If a purchasing manager decides to terminate an unsatisfactory ‘partnership’ and find a
new supplier, beforehand he has a favorable judgement on the availability of alternative
suppliers. Accepting a locked-in situation (scenario #8) is accompanied by a higher level of
calculative commitment, whereas terminating a partnership (scenario #9) is associated with
a lower level of competence trust.
6.8.5 Answers to the fifth research question: power and dependence in the matrix
Based on the case studies, 9 scenarios were developed that all focus on the selection of a
purchasing strategy, albeit under different conditions. There are fixed treatment variables
(the positions in the Kraljic matrix) and other variables that allow for the measurement of
a large number of conditions (relationship-dependence profiles). In this section we will
focus on the buyer’s dependence and the supplier’s dependence for the 9 different
scenarios. First, we will analyze the mean scores on buyer’s and supplier’s dependence in
the Kraljic matrix. In addition we will analyze the net dependence between them, thus
providing a measure for the balance of power. The results will be compared to prior
expectations. Secondly, a multiple regression analysis is conducted to explain the level of
buyer’s dependence from the four determinants of buyer’s dependence. The same
analysis is conducted for the supplier’s dependence in the nine scenarios.
6.8.5.1 Power and total interdependency in the matrix
In this section the answers will be presented to research question 5b:
“What are the levels of power and (total) interdependence in the categories of the Kraljic matrix?”
Our analysis of literature in chapter 4 has resulted in the conclusion that the Kraljic
categories correspond to four basic power positions:
- asymmetrical relationships in the bottleneck and leverage quadrants, and
- balanced relationships in the non-critical and strategic quadrants.
The survey data make it possible to compare these theoretical propositions with the
empirical findings, from the buyer’s perspective (see figure 6.13).
scenario supplier’s buyer’s net inter-

dependence dependence dependence dependence
(1) * (2) * (1) – (2) ** (1) + (2) ***
bottleneck #1 safety stocks 2.09 3.93 - 1.84 6.02

quadrant
#2 decomplex product 2.20 3.42 - 1.22 5.62
non-critical #3 pooling 2.01 1.49 + .52 3.50

quadrant
#4 individual ordering 1.86 1.67 + .19 3.53
leverage #5 partner of convenience 2.77 1.94 + .83 4.71

quadrant
#6 develop partnership 3.04 2.98 + .06 6.02
strategic #7 maintain partnership 3.25 4.22 - 0.97 7.47

quadrant
#8 accept locked-in 2.44 4.31 - 1.91 6.75
#9 terminate partnership 2.53 3.87 - 1.34 6.40
* supplier’s and buyer’s dependence are measured on a 5-point Likert scale

** the difference between supplier’s and buyer’s dependence is put on a scale from –4 (maximum supplier’s
dominance) to + 4 (maximum buyer’s dominance)
*** the total interdependence between buyer and supplier is put on a scale from +2 (minimum
interdependence) to + 10 (maximum interdependence)
FIGURE 6.13 Power and dependence in the matrix
Figure 6.13 shows the average supplier’s dependence and the average buyer’s
dependence in the nine scenarios (measured on a 5-point Likert scale). The average scores
of the buyer’s dependence are perfectly in accordance with our prior expectations: a high
dependence on the right-side of the matrix (bottleneck and strategic quadrant), a low
dependence on the left-side of the matrix (non-critical and leverage). However, this
250
conclusion can not be drawn for the supplier’s dependence. On average, the dependence
of the supplier is only in accordance with the expected values in the bottleneck and the
non-critical: relatively low. Remarkably, the supplier’s dependence in the leverage
quadrant is much lower than we might have expected in advance: medium scores (2.8 and
3.0 on a 5-point scale). The same conclusion holds for the strategic quadrant, where the
supplier’s dependence is medium in satisfactory partnerships (#7) and low in
unsatisfactory relationships (#8 en #9). Especially scenario #7 is in deviance with priori
expectations: we would have expected to find more balanced power-dependence
relationships. From a buyer’s perspective this has resulted in some striking conclusions
for the issue of power and dependence in the portfolio matrix: satisfactory partnerships are
dominated by the supplier. Obviously, once a buyer has entered a partnership this results in
a disproportionate raise in the dependence of the buyer on the supplying partner.
We have defined the relative power position of the buyer as the net dependence: the
difference between the supplier’s dependence and the buyer’s dependence (see the
second last column of figure 6.13). According to the calculating method, the difference
between supplier’s and buyer’s dependence is put on a scale from –4 (maximum
supplier’s dominance) to + 4 (maximum buyer’s dominance. In our study it is found that
buyer-supplier relationships are dominated by the supplier in the bottleneck quadrant
and in all of the scenarios in the strategic quadrants. In contrast, the buyer is dominant in
the non-critical and the leverage quadrants. In between a balance between buyer and
supplier is attained. A move from leverage to strategic, as in scenario #6, is indicative of a
balanced situation. Additional t-tests showed that in most of the scenarios the differences
between supplier’s and buyer’s dependence were statistically significant with exceptions
for scenario #6 with equal mutual dependence.
The differences between the theoretical and the found balances of power are outlined in
figure 6.13. The buyer dominance in the non-critical quadrant seems to be caused by
slightly lower switching costs for the buyer and by a slightly higher availability of
alternative trading partners, both in comparison with the supplier. The unexpected
supplier dominance in case of the satisfactory partnerships (as in scenario #7) can be
traced back to extremely high levels of logistical indispensability for the buyer. In
addition, the buyer perceives that he:
- has more need for the supplier’s technological expertise than vice versa, and
- faces higher switching costs than the suppliers, and
- has fewer alternative trading partners than the supplier.
We have compared scenario #7, which reflects a longer-lasting partnership, with scenario
#6 which corresponds to the beginning of such a partnership. From the buyer’s
perspective the development of a strategic partnership with a supplier typically means
that the buyer’s
dependence increases significantly, much more than the supplier’s dependence. The
result is a change from a balanced power position in scenario #6 into a situation of
supplier’s dominance in scenario #7.
The last column of figure 6.13 reveals the level of interdependence in the various scenarios.
Because of the calculation method, the interdependence scores are measured on a
constructed scale that runs from +2 (minimal interdependence) to +10 (maximum
interdependence). In chapter 4 we have posited that the total interdependence of a
relationship can be measured as the sum of buyer’s dependence and supplier’s
dependence. In the Kraljic matrix we are expecting:
- high levels of interdependence in the strategic quadrant,
- moderate levels of interdependence in the bottleneck and leverage quadrant, and
- low levels of interdependence in the non-critical quadrant.
The results that were found in our study confirm these prior expectations. The average
scores in the strategic quadrant are set between 6.4 and 7.5. These values are all higher
than the measured interdependence in the asymmetrical relationships in the bottleneck
quadrant and the leverage quadrant: between 4.7 and 6.0. The lowest levels of
interdependence are found in the non-critical quadrant (3.5).
6.8.5.2 Determinants of dependence
In this section we will test the hypotheses, related to research question 5c:
“What are the determinants of buyers’ dependence in the categories of the Kraljic matrix? Idem
for supplier’s dependence.”
In chapter 5 we have formulated a set of hypotheses, connecting buyer’s dependence and

supplier’s dependence to a selected number of their determinants.
We have related buyer’s dependence to:
- the logistical indispensability of the purchases (B1);
- the need of the buyer for the supplier’s technological expertise (B2);
- the number of alternative suppliers (B3);
- the switching costs of the buyer, incurred in case the supplier is replaced by another
(B4).
The supplier’s dependence is related to:

- the financial magnitude of the purchases (B5);
- the need of the supplier for the buyer’s technological expertise (B6);
- the number of alternative buyers (B7);
- the switching costs of the supplier, incurred in case the buyer is replaced by another
(B8).
252
From our differentiated portfolio perspective it is clear that relationships between

variables will differ in different scenarios and quadrants.
* Hypothesis B1 was confirmed only for the right-side of the matrix: the scenarios in the
bottleneck and strategic quadrant. In other words, the logistical indispensability has a
positive impact on the (relatively high level of) buyer’s dependence in case of bottleneck
and strategic items. In contrast, for the non-critical and the leverage quadrant no
significant effects were found, indicating that the logistical indispensability has no
impact on the (relative low level of) buyer’s dependence. There seems to be a conditional
relationship between logistical indispensability of the product and the buyer’s
dependence: a positive relationship in cases of high supply risk and no relationship in
cases of low supply risk.
* Hypothesis B2 was confirmed in all scenarios, with a notable exception for scenario #1
and #2 in the bottleneck quadrant. In general, the need for the supplier’s expertise has a
positive impact of the buyer’s dependence on that supplier. Especially in the non-critical
quadrant a relative strong impact was found, indicating that the differences between
companies with respect to their buyer’s dependence could be explained by the
differences between their need for the supplier’s expertise.
* Hypothesis B3 was confirmed in almost all scenarios: a negative relationship between the
availability of alternative suppliers and the buyer’s dependence. No significant relationship
was found for the scenarios which are related to the so-called partnership of convenience
in the leverage quadrant: scenario #3 means pooling of requirements, entering a
convenience partnership and scenario #5 means maintaining a convenience partnership.
Obviously the relatively low level of buyer’s dependence is not influenced here by the
availability of alternative suppliers.
* Hypothesis B4 was confirmed too in most scenarios. With the exception of scenario #3
(pooling), there is a positive significant relationship between the buyer’s switching cost
and his dependence.
scenario dependent variable:

buyer’s dependence
logistical need for alternative switching intercept R2

indispens- supplier’s suppliers cost buyer
ability expertise
1 safety stocks .281 a .044 - .132 a .153 a 2.400 a .190
2 decomplex product .354 a -.023 -.187 a .294 a 1.711 a .255
3 pooling - .041 .411 a -.126 . 096 1.356 a .257
4 individual ordering .051 .413 a - .266 a .224 a 1.552 a .472
5. partner of convenience .032 .189 a -.060 .396 a .218 .302
6 develop partnership .105 .196 a - .211 a .318 a 1.710 a .314
7. maintain partnership .348 a .173 a - .125 a .172 a 1.567 a .313
8 accept locked-in .410 a .126 a - . 126 a .135 a 1.775 a .339
9 terminate partnership .491 a .272 a - .126 a .185 a .385 .384
a
significant at p<0.05
FIGURE 6.14 Explaining buyer’s dependence in the matrix: results of the regression
analysis
* Hypothesis B5 proposes a positive relationship between the financial magnitude and the
supplier’s dependence. This hypothesis was confirmed in all scenarios. Moreover,
considering the value of the response parameters, the financial magnitude is the most
important variable for the explanation of the supplier’s dependence. It is remembered
that financial magnitude is considered from the supplier’s perspective: the level to which
the buyer is an important customer for the supplier, considering the volume of trade.
* Hypothesis B6 refers to the need of the supplier for the buyer’s technical expertise. In 7 of
the 9 scenarios a positive relationship was found between the need for buyer’s expertise
and the supplier’s dependence. Only in scenarios where the final position of the items in
the non-critical quadrant, variance in need for expertise does not produce dependence:
scenario #2 where the product can be ‘decomplexed’ and scenario #4 where individual
ordering is an option.
* The results with hypothesis B7 are less unambiguous. In exceptional cases a negative
relationship can be established between the availability of alternative buyers and the
supplier’s dependence. From the buyer’s perspective only in cases of ‘real’ partnerships
(#6 and #7) a significant impact was found.
254
* Hypothesis B8 was confirmed in practically all scenarios, indicating a positive

relationship between the switching costs of the supplier and the supplier’s dependence.
Switching costs are associated with a need for the buyer’s technical expertise. Therefore,
it seems that suppliers who have made a commitment to the relationship actually need
the buyer’s expertise, which all leads to switching costs and supplier’s dependence.
scenario dependent variable:

supplier’s dependence
financial need for buyer’s alternative switching intercept R2

magnitude expertise buyers cost
supplier
1 safety stocks .342 a .265 a - .030 .210 a .188 .50

8
2 decomplex product .265 a .105 .095 .462 a .577 .19
4
3 pooling .339 a .250 a -.209 a .134 a 1.352 a .40
1
4 individual ordering .409 a .136 -.092 .104 1.003 a .28
2
5 partner of convenience .440 a .298 a .014 .222 a - .123 .39
1
6 develop partnership .295 a .222 a - .146 a .233 a 1.072 a .34
7
7 maintain partnership .358 a .243 a - .135 a .218 a .735 .42
4
8 accept locked-in .296 a .248 a - . 094 .226 a .558 .52
1
9 terminate partnership .455 a .202 a - .016 .157 a .169 .52
9
a
significant at p<0.05
FIGURE 6.15 Explaining supplier’s dependence in the matrix: results of the regression
analysis
6.9 Conclusions
The survey was aimed at answering the research questions with respect to:
- the explanation of portfolio utilization (#2),
- the conditions under which portfolio-based strategies are selected (#4), and
- the role of power and dependence in the Kraljic matrix (#5).
A questionnaire was developed, based on literature study, on the insights from the case
study, and on the results of the pilot study. The fourth research question required the
development of 9 scenarios, based on the main strategies that were found in the case
studies. The research design included a quasi-experiment where respondents were asked
to evaluate a series of questions for the scenarios, relating to the conditions that constitute
a relationship-dependence profile.
Answering more than 175 questions is likely to produce non-response problems.
However, the special measures turned out to be rather effective, resulting in a response
rate of 20.6% (n=238). Additional statistical analysis indicated that the study does not
suffer from non-response bias.
With respect to the explanation of the use of the purchasing portfolio tool, the most
important findings are:
- Portfolio analysis delivers what it is supposed to: provide additional understanding of
problems and possibilities of purchasing and provide assistance in the process of
developing differentiated purchasing strategies.
- Users contrast in a positive way with non-users of the portfolio, especially on their
purchasing’s professionalism (skills) and their contributions to the competitive position
of the company. In addition it was found that the portfolio was relatively more used by
larger companies with higher purchasing shares.
With respect to the conditions-research question, statistical tests revealed the significant
differences between the relationship-dependence profiles of the scenarios. The analysis of
differences was conducted at the quadrant-level within the Kraljic matrix. A pair-wise
comparison of profiles was used in our search for conditions under which the various
purchasing strategies are selected. The main findings of this condition-seeking research
strategy are:
- A move from the bottleneck quadrant to the non-critical quadrant is only an option if the
conditions are satisfied, concerning the possibilities for standardization and the availability
of alternative suppliers.
- A move from the non-critical quadrant to the leverage quadrant is only feasible if there is
a sufficient volume of trade, as perceived by the supplier (financial magnitude) and an
ample availability of alternative suppliers.
- In the leverage quadrant: the main condition for engaging in a partnership with a
supplier is related to already higher levels of supplier’s dependence and especially buyer’s
dependence. In addition, there is more commitment towards the future partner.
- Satisfactory partnerships in the strategic quadrant are distinguished from unsatisfactory
‘partnerships’ by higher levels of supplier’s dependence, trust and commitment.
- An unsatisfactory ‘partnership’ in the strategic quadrant will be terminated, if the buyer
is convinced that there are relatively more alternative suppliers. Accepting a locked-in
situation (scenario #8) is accompanied by a higher level of calculative commitment,
terminating a partnership (scenario #9) is associated with a lower level of competence
trust.
256
Next we have analyzed the issue of power-and-dependence in the portfolio matrix. We

have calculated interdependencies and the power positions of the buyer in the Kraljic
categories and we have compared them with prior theoretical expectations. Most
remarkable was the observed supplier dominance in case of all of the 3 scenarios in the
strategic quadrant. The results may shed a different light on the buyer’s view on issues of
power and dependence: even satisfactory partnerships are dominated by the supplier.
The observed buyer dominance in the non-critical quadrant seemed to be caused by
slightly lower switching costs for the buyer and by a slightly higher availability of
alternative trading partners, both in comparison with the supplier. We found an expected
supplier dominance in the bottleneck quadrant, and an expected buyer dominance in the
leverage quadrant. The observed levels of total interdependence were completely in
accordance with prior expectations: ‘highest’ in the strategic quadrant, ‘moderate’ in the
bottleneck and leverage quadrant, and ‘lowest’ in the non-critical quadrant.
Finally, we have explained the buyer’s dependence and the supplier’s dependence from a
limited number of explanatory variables (‘determinants of dependence’). We have
established different conditional relationships in categories of the Kraljic matrix for most of
the explanatory variables:
- A conditional relationship between the logistical indispensability of a product and the
buyer’s dependence: a positive relationship in cases of high supply risk (on the right side
of the matrix) and no relationship in cases of low supply risk (on the left side of the
matrix).
- A conditional relationship between the need for the supplier’s expertise and the buyer’s
dependence: no relationship in case of bottleneck items and a positive relationship in case
of non-critical, leverage and strategic items.
- For practically all scenarios it was found that both the availability of alternative suppliers
and the switching costs of the buyer have a significant impact on the buyer’s dependence.
- A positive relationship between the financial magnitude and the supplier’s dependence in
all scenarios.
- There is a positive relationship between the need for the buyer’s technical expertise and
the supplier’s dependence, except for non-critical items for which no impact was
established.
- For practically all scenarios it was found that both the availability of alternative buyers
and the switching costs of the supplier have a significant impact on the supplier’s
dependence.
257
7 Conclusions and recommendations
In this final chapter we will present the main conclusions, by summarizing the answers
to the research questions of this research project. In addition, recommendations will be
made for practitioners and for academics. The recommendations for business are
intended to provide guidance for the application of a purchasing portfolio approach.
Researchers with an interest in buyer-supplier relationships may benefit from the
insights and the experiences of this research project. The chapter is completed by some
suggestions for further research.
7.1 Conclusions and main findings

Kraljic published his seminal paper in 1983, introducing a comprehensive portfolio model
for purchasing management. Although other models have been developed, Kraljic’s
approach subsequently became the dominant approach in the profession. This study
started with the observation that there is a striking contrast between the growing use and
acceptance of purchasing portfolio models and research into the actual use of such
models. Most publications are conceptual or anecdotal by nature. In addition there are
unanswered questions with respect to the theoretical foundations of the Kraljic approach.
The main objective of this research project is to gain a better understanding of:
- how they could be used by purchasing professionals in order to pursue differentiated
In line with these objectives, we have posed five main research questions:
1. What are the differences and similarities of the various purchasing portfolio models?
2. Which factors would explain the utilization of the purchasing portfolio analysis?
3. How are portfolio models employed by experienced purchasing professionals?
4. Under which conditions are the various portfolio-based strategies selected in
purchasing management?
5. What is the role of power and dependence in the Kraljic approach?
In order to answer these questions, we have combined three research methods: a

literature study, a series of explorative case studies and a large scale survey among
purchasing professionals at industrial companies in The Netherlands, successively. The
methods were used in a complementary way. The results of the literature study were
used in the case studies and in the survey, findings of the case studies were input for the
258
design of the survey and the questionnaire. Each research method has its own
characteristics and its own strong points, which makes it more appropriate for answering
certain types of research questions.
In this section we will present the main conclusions of this research project, organized
according to the research questions.
Research question 1
What are the differences and similarities of the various purchasing portfolio models?
The comparison of portfolio models has concentrated on the following four elements: (1)
dimensions, (2), categories, (3) strategic recommendations, and (4) acceptance and
adoption of the various models. The analysis of literature made clear that Kraljic (1977,
1983) introduced the first comprehensive portfolio approach for purchasing and supply
management. Many years after the introduction, a reasonable amount of evidence can be
found that Kraljic’s basic ideas and concepts represent the dominant approach in the
profession. Other authors have used Kraljic’s basic ideas for the development of rather
similar models. Some changed the labels of categories (e.g. Elliott-Shircore and Steel,
1985), some adapted the dimensions (e.g. Hadeler and Evans, 1994; Olsen and Ellram,
1997), others elaborated on the recommendations for the leverage, the non-critical and the
bottleneck category (Van Weele, 1992; Syson, 1992). The comparison of the various
portfolio models has shown that there are more similarities than differences. Most
approaches use practically the same dimensions as the Kraljic matrix, the same categories
and the same recommendations.
FIGURE 7.1 Kraljic’s categories and strategic recommendations
Kraljic’s approach includes the construction of a portfolio matrix that classifies products
on the basis of two dimensions: profit impact and supply risk (‘low’ and ‘high’). The
result is a 2x2 matrix and a classification in four categories: bottleneck, non-critical,
leverage and strategic items. Each of the four categories requires a distinctive approach
towards suppliers (see figure 7.1). Non-critical items require efficient processing, product
standardization, order volume and inventory optimization. Leverage items allow the
buying company to exploit its full purchasing power, for instance through tendering,
target pricing and product substitution. Bottleneck items cause significant problems and
risks which should be handled by volume insurance, vendor control, security of
inventories and backup plans. A further analysis of the strategic items is recommended.
We agree with Nellore and Söderquist (2000) who contended that all purchasing portfolio
approaches basically use the same three steps: the analysis of products and their
classification in a matrix (diagnosis), the analysis of required supplier relationships to
deliver the products (objectives), and the development of action plans in order to bridge
the gap between current and required supplier relationships (strategies). In addition, the
analysis of literature has indicated that no significant differences have been found in the
general approach of the various portfolio models in purchasing. The Kraljic matrix is the
dominant approach in the profession.
260
Research question 2
Which factors would explain the utilization of the purchasing portfolio analysis?
A number of sources reported on the growing use of the portfolio analysis in practice
(e.g. Lamming and Harrison, 2001; Van Weele, 2000; Kamann, 2000a; Boodie, 1997), while
other publications claim that the opposite is true (e.g. Christopher and Jüttner, 2000;
Olsen and Ellram, 1997). Most of these claims do not refer to concrete numbers of
portfolio users in practice. The survey resulted in facts and figures about the actual
utilization of the portfolio analysis. It was found that 73.8% of the respondents are using
portfolio analysis in their purchasing practice. Taking differences in company size into
account (SME’s versus larger companies), we calculated a weighted average use
frequency of almost 60% for manufacturing companies in The Netherlands.
Based on literature study five groups of variables were found, to describe and explain the
differences between users and non-users of a portfolio approach: the company size, the
share of purchasing, the position of the purchasing, the professionalism of purchasing,
and the orientation of purchasing. These variables entered a use model. The results
indicated that the portfolio was relatively more used by larger companies with higher
purchasing shares. Statistical analysis proved that users contrast sharply and positively
with non-users of the portfolio approach, especially on their:
- purchasing’s professionalism (skills), and
- their contributions to the competitive position of the company.
The results underscored our main conclusion that portfolio use is positively related with
the maturity of purchasing within companies, expressed by the position and the
professionalism of the purchasing function. Reasons for not using the portfolio were
found to be a lack of knowledge, a lack of time as well as perceptions on the limitations of
the tool.
We have investigated and quantified the direct accountable impact of portfolio use,
operationalized as to what extent users:
- experience additional understanding of current purchasing problems and possibilities
(1), and
- develop differentiated strategies for their purchasing and supplier management (2).
These effects are considered as intermediary measures for overall purchasing
performance measures. Analysis confirmed that the portfolio technique ‘delivers’ what it
is supposed to: it significantly adds to the understanding of situations (diagnostic
purpose), and it assists the process of developing strategies (prescriptive purpose).
Research question 3
How are portfolio models employed by experienced purchasing professionals?
The research project began with the contention that little is known about the actual
process of using purchasing portfolio models. The literature study identified a number of
problems and unanswered questions, referring to the measurement of variables, the
disregard for the supplier’s side, the selection of strategies based on two dimensions, the
limited and deterministic character of the strategic recommendations, and the absence of
explicit movements within the matrix. The critique of the Kraljic approach however does
not include the experience of practitioners. How do experienced professionals handle
such issues in practice? What could we learn from their experience? The case studies
addressed the gap between conceptual problems and practical solutions, identifying and
describing advanced uses of a purchasing portfolio approach. Explorative case studies
revealed the way in which experienced practitioners handle the main measurement issues
and strategic issues in practice.
The investigated cases provided useful insights in the possibilities and actual use of
purchasing portfolio analysis. In the case studies we found a variety of approaches and
differences of scope and scale, which have to be viewed in the specific business context.
The cases studies revealed three distinctive methods of measuring variables and
weighting factors:
(1) consensus method
(2) one-by-one method
(3) weighted factor score method.
Each method satisfies the needs and expectations of the different users. The reason for
this can be found in the additional steps that have to be taken in the portfolio analysis.
Before strategic actions are determined, it is imperative to complete a further process of
interpreting and reflecting on the results. The positioning of items in the matrix should be
considered as the starting point of portfolio analysis, definitely not the finishing point. It
is imperative that users reflect on the results. If necessary, manual adjustments should be
made. In-depth discussions on the positions in the matrix are considered as the most
important phase of the analysis. Strategic discussions provide deeper insights and may
lead more easily to consensus-based decisions. It is felt by the users that the Kraljic
framework facilitates these important discussions to a large extent.
based on just two basic dimensions (e.g. Dubois and Pedersen, 2002: 40)? Actually, the
answer is simple: one cannot. In addition to the various factors that constitute the two
262
dimensions of any matrix, we have found that experienced portfolio users always
included additional information on:
- the specific situations on supply markets and
- the capacities and the intentions and competences of individual suppliers.
Unquestionably, the supplier’s side should be included in any strategic thinking on the
field of purchasing and supply management. Practitioners have found a reply to the
critique of the Kraljic approach which said that the supplier’s side is a disregarded
element in Kraljic’s model.
portfolio-based strategy intended movements in the matrix
bottleneck quadrant #1 keep safety stocks stay bottleneck
#2 decomplex the product and find a towards non-critical

new supplier
non-critical quadrant #3 pooling of requirements towards leverage
#4 individual ordering stay non-critical
leverage quadrant #5 maintain a partnership of stay leverage

convenience
#6 develop a strategic partnership towards strategic
strategic quadrant #7 maintain a strategic partnership stay strategic
#8 accept a locked-in partnership stay strategic
#9 terminate the partnership, find a towards leverage

new supplier
FIGURE 7.2 Strategic directions in the portfolio matrix: movements and impact
Comparative analysis of the case studies has resulted in a conceptual model of strategic
directions, providing an overview of the main strategic choices for the categories in the
matrix (see figure 7.2). These strategies are elaborated in 9 corresponding scenarios,
identifying different strategic directions (actions), and the circumstances that accompany
these purchasing strategies. The scenarios are used in the survey, in order to investigate
the conditions under which portfolio-based strategies are selected, and the role of power
and dependence in the Kraljic matrix. A dichotomy was identified between:
- strategies to hold a position (1) and
- strategies to move to another position (2).
At the right side of the matrix (in the bottleneck and the strategic areas) movements are
pursued in order to reduce a high level of supply risk. In terms of the matrix, this means
moving to the left. Non-critical items are preferably moved upwards and exceptionally
leverage positions are exchanged for strategic positions. These are the most common
movements within the matrix.
- partners of convenience, located in the leverage quadrant, where relationships are
dependence;
Research question 4
Under which conditions are the various portfolio-based strategies selected in
purchasing management?
The possibilities for selecting purchasing strategies are obviously limited by external
conditions. The questionnaire included an elaboration of the 9 purchase situations in
scenarios. Each scenario contains a description of:
- the selected portfolio-based purchasing strategy,
- Kraljic’s dimensions (profit impact and supply risk), and
- additional specific conditions for the selection of the purchasing strategy.
In this study we have introduced and used the concept of relationship-dependence profile, as
an instrument to measure the values and impact of conditions that accompany a scenario.
A relationship-dependence profile includes a limited number of key factors appropriate
to characterize different types of buyer-supplier relationships. Based on literature study
and based on the results of the case studies, four main groups are set to form a
relationship-dependence profile:
- determinants of buyer’s dependence (logistical indispensability, need for the supplier’s
technological expertise, availability of alternative suppliers, switching cost for the buyer);
- determinants of supplier’s dependence (financial magnitude, need for the buyer’s
technological expertise, availability of alternative buyers, switching cost for the supplier);
- relationship characteristics: trust and commitment (competence trust, goodwill trust,
affective commitment, calculative commitment)
- the nature of specifications (product customization, possibilities for standardization).
A combination of these factors is addressed to as a relationship-dependence profile.
264
By comparing the relationship dependence profiles of two scenarios in the same

quadrant, it was possible to test whether these profiles differ significantly. Pairwise
comparison of relationship profiles showed that all differences are significant at the
p<0.05 level. In other words: there are statistically significant differences between the
profiles that fit alternative strategic choices in the same portfolio quadrant. In addition,
an explorative investigation was conducted, in search for statistically differences between
profiles in each quadrant. Differences were traced back to the variables that constitute a
profile. This condition-seeking research strategy resulted in an overview of conditions under
which the various purchasing strategies are selected. The main findings are:
- in the bottleneck quadrant: a strategic move to the non-critical quadrant is only an option
if the conditions are satisfied, concerning the possibilities for standardization and the
availability of alternative suppliers.
- in the non-critical quadrant: the switch towards the leverage quadrant is only feasible if
there is a sufficient volume of trade, as perceived by the supplier (financial magnitude) and
an ample availability of alternative suppliers.
- in the leverage quadrant: the main conditions for engaging in a partnership with a
supplier are related to already higher levels of supplier’s dependence and especially of
buyer’s dependence. In addition, there is more commitment towards the future partner.
- in the strategic quadrant (1): satisfactory partnerships in the strategic quadrant are
distinguished from unsatisfactory ‘partnerships’ by higher levels of supplier’s dependence,
trust and commitment.
- in the strategic quadrant (2): if it is decided that an unsatisfactory ‘partnership’ should be

terminated and that a new supplier should be found, beforehand the buyer has a
favorable judgement on the availability of alternative suppliers. Accepting a locked-in
situation is accompanied by a higher level of calculative commitment; terminating a
partnership is associated with a lower level of competence trust.
Research question 5
What is the role of power and dependence in the Kraljic approach?
Literature study confirmed that power and dependence are very important in
understanding buyer/supplier relationships. There are indications that power and
dependence are important as well in the Kraljic approach, considering some of the
recommendations and the general idea of the portfolio approach: “to minimize supply
vulnerability and make the most of potential buying power” (Kraljic, 1983, 112). An in-
depth analysis of the Kraljic approach, the dimensions and the recommendations, has
provided a new perspective on the Kraljic matrix, making a reasonable case that the
resource dependence theory should be considered as the (implicitly applied) theoretical
foundation for the Kraljic portfolio approach.
The literature study made clear that a comprehensive view of the dyadic nature of buyer-
supplier relationships should include the assessment of:
(1) the difference between buyer’s and supplier’s dependence (net dependence) which
corresponds with the relative power between parties;
(2) the sum of buyer’s and supplier’s dependence (total interdependence) which indicates
the intensity and development phase of the relationship between parties.
We have applied these key concepts to the Kraljic matrix, comparing theoretical
expectations with the empirical findings (see figure 7.3). Most remarkable is the observed
supplier dominance in case of all of the 3 scenarios in the strategic quadrant. Especially
the unexpected dominance in the satisfactory partnerships, from the buyer’s perspective.
The results have shed a different light on the buyer’s view on issues of power and
dependence: even satisfactory partnerships are dominated by the supplier. Obviously, once a
buyer has entered a partnership this results in a disproportionate raise in the dependence
of the buyer on the supplying partner. We have explained this finding by extremely high
levels of logistical indispensability for the buyer. In addition, the buyer perceives that he:
- has more need for the supplier’s technological expertise than vice versa, and
- faces higher switching costs than the suppliers, and
- has fewer alternative trading partners than the supplier
The buyer dominance in the non-critical quadrant seemed to be caused by slightly lower
switching costs for the buyer and by a slightly higher number of alternative trading
partners, both in comparison with the supplier. The observed levels of total
interdependence were completely in accordance with prior expectations.
266
balance of power total interdependence
expected observed expected observed
bottleneck supplier dominance supplier dominance moderate moderate
non-critical balanced buyer dominance lowest lowest
leverage buyer dominance buyer dominance moderate moderate
strategic balanced supplier dominance highest highest
FIGURE 7.3 Comparison of power and interdependence in the Kraljic matrix: theory
and practice
Figures 7.4a and 7.4b visualize the differences between the expected, theoretical power
relations in the Kraljic matrix and the observed power relations in the survey data.
Obviously, the demarcation line between buyer and supplier dominance runs between
the left and the right side of the matrix. The level of supply risk determines the perceived
power balance between buyer and supplier, from the buyer’s perspective.
profit
impact
buyer
dominance
power
balance
supplier
dominance
supply risk
FIGURE 7.4a Theoretical power relations in the Kraljic matrix

Based on Kempeners and Van Weele (1997) and Gelderman(2000)
profit
impact
buyer power supplier

dominance balance dominance
supply risk
FIGURE 7.4b Observed power relations in the Kraljic matrix
In this research project we have elaborated on the determinants of organizational

dependence. Analysis of literature suggested that organizational dependence is a
function of the importance of the exchanged resources and the substitutability of the source.
Analysis of various conceptual and empirical studies to the determinants of dependence
has laid the foundation for a tentative model, which was adapted and improved by the
reactions and comments of purchasing practitioners (case studies). The literature review
revealed that most empirical studies, involving ‘dependence’-issues, selected
organizational dependence as an explanatory variable. In contrast a very limited number of
empirical studies was devoted to the explanation of organizational dependence. In
addition, no studies were found that made a distinction between different types of buyer-
supplier relationships, let alone that a distinction was made between the categories in a
purchasing portfolio approach. The results of our study therefore have tapped new
sources of knowledge, explaining the buyer’s dependence (and the supplier’s
dependence) in the different portfolio categories.
In line with the portfolio perspective in this study, we have found different relationships
and different determinants for different categories. This means there are conditional
relationships for most of the explanatory variables. Findings of interest are:
- there is a conditional relationship between the logistical indispensability of a product and
the buyer’s dependence: a positive relationship in cases of high supply risk (at the right
side of the matrix) and no relationship in cases of low levels of supply risk (at the left side
of the matrix);
268
- there is a conditional relationship between the need for the supplier’s expertise and the
buyer’s dependence: no relationship in case of bottleneck items and a positive
relationship in case of non-critical, leverage and strategic items;
- for practically all scenarios it was found that the availability of alternative suppliers and
the switching costs of the buyer have both a significant impact on the buyer’s dependence;
- a positive relationship between the financial magnitude and the supplier’s dependence in
all scenarios;
- there is a positive relationship between the need for the buyer’s technical expertise and the
supplier’s dependence, except for non-critical items for which no impact was established;
- for practically all scenarios it was found that both the availability of alternative buyers
and the switching costs of the supplier have a significant impact on the supplier’s
dependence.
7.2 Reflection on the methodology

In every study decisions have to be made regarding the methodology. In this section we
will reflect on some of these decisions and we will discuss the major limitations of this
research. In addition we will share some insights that we have acquired on some specific
research issues.
(1) Limitations of the study

All companies in this study are located in the Netherlands and they are manufacturing
companies. These limitations imply that the results may not be transferable to other
countries and other branches. A replication of our study could reveal similarities and
differences, indicating whether or not we can generalize from our findings.
Another significant limitation of this study is that we have focussed on the buyer’s
perspective. Obviously, dyadic data would present a richer picture, although there would
be almost insurmountable problems identifying the perceptions of corresponding
suppliers. More importantly, we have made a deliberate and argued choice for the
buyer’s side. The limitation to the buyer’s perspective is justified by the main object of
our research project, to describe and explain the selection of purchasing and supplier
strategies by means of a (purchasing) portfolio approach. This does not mean that the
supplier’s side is neglected in this study. On the contrary, it means that we have included
the buyer’s perceptions regarding the supplier’s side. Assessing buyer’s perceptions is the
most meaningful way to measure variables, since purchasing decisions are actually made
on buyer’s perceptions of the relevant conditions. To conclude, in line with the objectives
and the perspective of the overall study, it was most sensible to measure perceptions
from the buyer’s perspective.
In this study we have combined three methods: an extensive literature study, a series of
in-depth case studies and a large scale survey among purchasing professionals. We are
convinced that such a combination of methods is necessary for any in-depth study of
research topics in the field of business administration. Although the combination of

methods has proved to be very useful, there are limitations. The use of any research
method implies dealing with the usual limitations. Case studies are not suitable for
testing propositions and hypotheses, whereas survey studies do not allow for exploring
and identifying advanced business practices. These general limitations should not pose
problems since each research method is aimed at the specific objectives and research
questions at hand. Beyond the general limitations, there are some specific limiting
factors.
The findings of the case studies are based on a rather limited number of three cases. We
have captured the portfolio application in the context of three different companies: the
corporate level, the level of a multinational business unit and the level of a fairly small
company. Additional research in different types of companies could reveal other and
similar advanced purchasing portfolio practices. The comparison with other companies
could identify differences, to be explained by company specific-factors, such as company
size, organizational structure, technology, customers, network position, and
organizational culture.
The survey among purchasing practitioners yielded a satisfactory response rate of 20.6%
(n = 238). Statistical analysis indicated that the survey data did not suffer from non-
response bias. However, due to the composition of the NEVI-database, large firms were
over-represented in the sample in comparison with the proportion of SME’s.
The design of the questionnaire was characterized by the use of scenarios. The external
validity of experimental designs are usually problematic because of the artificial
manipulations of variables and situations. Although this was also true for our design, we
have made an effort to reduce the problems of validity, by basing the scenarios on the
empirical findings of the previous case studies and by testing and improving them in a
pilot study. However, there is no evidence in the study with which to test whether there
is respondent’s bias in the answers that can be ascribed to different interpretations of the
scenarios. In addition, respondents might be inclined to produce ‘socially desirable’
perceptions, building a logically consistent picture which may not coincide with reality.
This limitation however is usually associated with the use of a survey methodology,
assessing respondent’s perceptions.
(2) A scenario method versus a key supplier perspective

The unit of analysis in purchasing research, using a survey method, is usually the buying
company or a purchasing professional. In many studies respondents are asked to answer
questions that refer to their ‘suppliers’ in general. Obviously, it might not always be
appropriate to ask for the ‘average relations with all suppliers’. Alternatively
respondents are often asked to express their opinions and views on their relationship
with a single (type of) supplier. In our view, most empirical studies in purchasing are
limited to questions that relate to the major or the largest supplier. Instructions for
270
respondents are of the form ‘Select the most important supplier and answer the following
questions’. This means that the results of the study are restricted to the key supplier,
namely a supplier with ‘the largest share in our purchasing volume’ (turnover) with
whom the buying company has developed some kind of collaboration or partnership. In
some studies there is obviously one major supplier, especially in a channel context
(manufacturer- distributor). In other studies, relating to industrial relationships, the
limitation to the largest supplier is not a self-evident point of departure. After all, it is
clear beforehand that companies maintain different kinds of relationships with different
kinds of suppliers. A questionnaire that is restricted to the relationship with the major
supplier would not be appropriate for our study, which investigates a portfolio approach
to purchasing management. Some argue that the largest supplier can be used as an
approximation for the entire supply base, depending on the importance of this supplier
to the focal firm (e.g. Mol, 2001: 84). We have to disagree on this matter and believe that
the relationship with the major supplier will not be representative for other relationships
in the supply base. It seems that research reports are not always specific and explicit
about this sample bias and the consequences for the results and conclusions. Companies
are usually faced with vast differences in types of buyer-supplier relationships. In our
study we have developed a tailor-made solution: a scenario method. Based on the case
studies, 9 scenarios were developed that all focus on the selection of a purchasing
strategy, albeit under different conditions. The scenario method appeared to be an
appropriate methodology for measuring conditions (relationship-dependence profiles)
and for studying the role of power and dependence in buyer-supplier relationships.
(3) Calculative commitment and switching costs

There is a broad consensus that commitment is to be considered as a prerequisite for any
successful partnership. In our study we found a (very) strong statistical relationship
between calculative commitment and the buyer’s switching cost. We intended to measure
different things: calculative commitment should refer to a negative motivation for
maintaining a relationship with a supplier, in contrast with affective commitment
(‘pleasant working with a supplier’). Switching cost should refer to the costs that act as
economic barriers to the replacement of one supplier with another. However, for
respondents it was probably not clear how to distinguish between these factors. At closer
consideration, the operationalizations could have caused the confusion:
- switching cost: ‘you will incur high cost if you replace this supplier’
- calculative commitment: ‘we are doing business with this supplier, mainly because it
would take too much time, energy and expense to end the relationship’.
The corresponding questions could have easily been interpreted by respondents as
referring to the same basic construct of ‘switching cost’. A subsequent factor analysis has
confirmed this assumption. In all 9 scenarios a single component was found, including
the following three factors: competence trust, goodwill trust and affective commitment.
Consistently, calculative commitment showed low factor loadings on this ‘commitment

and trust’-component and high factor loadings on the components that include ‘switching
cost buyer’.
We propose to use a continuum that runs from ‘compliance’ to ‘commitment’. The values
of this relationship-motivation variable reflect the willingness to be engaged in a trading
relationship. Compliance captures the unwilling dimension of behavioral shifts, while
commitment is viewed as the willing dimension (affective commitment).
(4) Professionalism and orientation of purchasing

We have identified the maturity of the purchasing function as a key variable for gaining
insights in the utilization and effectiveness of a portfolio approach in purchasing. To a
large extent we have used the operationalization of Rozemeijer (2000) who identified
‘position of purchasing’, ‘purchasing’s professionalism’, and ‘purchasing orientation’ as
the main factors, constituting the construct variable ‘purchasing maturity’. In our study
purchasing’s professionalism has been measured by two items, referring to the skills of
purchasing personnel for working in cross-functional teams and for developing
purchasing and supplier strategies. Three dimensions of ‘orientation’ have been
distinguished: collaboration with suppliers (1), tough negotiations in pursuit of the lowest
prices (2), and the engagement of purchasers in clerical duties and operational activities (3).
To gain more insight in the structure of the explaining variables, an additional factor
analysis was conducted. The results are shown in figure 7.5, based on a principal
component analysis (extraction method) and a varimax rotation (rotation method). The
shaded cells show relatively high factor loadings (> .60). Three components remained,
explaining 66.8% of total variance. The first component highly corresponds to the two
variables that are selected to operationalize the ‘position of purchasing’-construct. The
second component corresponds to two elements of the ‘orientation’-construct: orientation
on collaboration and orientation on negotiation and lowest prices. As expected, they are
negatively correlated. These findings are very much in line with our prior expectations.
For the third component however, the matter is different. This component clearly
includes the two ‘skills’-variables, but in addition there is a strong negative relationship
with the last orientation-variable, the variable that refers to clerical and operational
activities. Obviously, especially an orientation on these inferior purchasing activities is
negatively associated with the extent to which purchasing professionals in a company
dispose of substantial skills for working in cross-functional teams and for developing
purchasing and supplier strategies. To conclude, the findings of the factor analysis
largely confirmed prior expectations, but an exception has to be made for the clerical
orientation-variable which is not connected to the other two orientation-items. Instead, a
clerical orientation is (negatively) connected to the professional skills of purchasers
(professionalism). This means that an orientation on these inferior purchasing activities is
negatively associated with the extent to the level of professionalism of purchasers. In our
272
view, the professionalism of purchasing is reflected by the skills of purchasers and their
(negative) orientation towards and engagement in clerical activities. Future research
could profit from this suggestion.
Components
1 2 3
contributes to competitive position .80 .00 .14
reports to top management .81 .00 .11
skills for cross functional teams .22 .00 .86
skills for developing strategies .15 .00 .83
orientation on clerical and operational activities .00 .20 -.66

orientation on collaboration .33 -.71 .15
orientation on negotiation and lowest prices .25 .82 .00
FIGURE 7.5 Factor analysis for 7 explanatory variables: the rotated component matrix
7.3 Recommendations for business

In this section we will present recommendations for (potential) users of a purchasing
portfolio approach. Practitioners might benefit from a 5-step approach that is developed
mainly with insights from observations and interviews of the case studies. We have
observed that in the course of time the Kraljic approach has entered many textbooks on
purchasing and supply management. Strikingly, (intended) practitioners will search in
vain for practical guidelines that might help them to introduce and apply this purchasing
portfolio approach. Most publications do not address the main issues and how the
portfolio analysis could or should be used in practice. This section aims to address this
gap, identifying issues and providing answers and solutions.
7.3.1 Adopting a portfolio approach
Undoubtedly non-users will have their reasons for not using a portfolio approach. Our
research indicated that the main reasons for not using the portfolio refer to a lack of time
and/or knowledge and to the perceived limitations of the tool. A lack of time is a matter of
priority, a lack of knowledge can be set to rights, while the perceived limitations should
be weighed against the attributed benefits. As we have concluded earlier the portfolio
technique ‘delivers’ what it is supposed to: it adds to the understanding of purchasing
situations (diagnostic purpose), and it assists the process of developing differentiated

purchasing strategies (prescriptive purpose). That is why we would break a lance for a
purchasing portfolio approach, especially considering the positive experiences from users
and the resulting impact.
In every case-company a champion was found who supported the portfolio analysis.
These ‘product champions’ were the highest purchasing professionals in the
organization. It should be noted that performing a portfolio analysis involves team work.
The views of colleagues from different fields of expertise should be added to the more
functional purchasing perspective. For a designer ‘replaceability’ might be important,
while the production manager might focus on ‘risk of failure’. For reasons of support and
implementation a cross functional team is required, with representatives from all relevant
departments and specialist fields. Considerable time and energy will have to be spent in
the preparation phase, getting the organization ‘ready for action’. Top management will
have to be convinced of the project, but also (line) managers from other departments.
Some practitioners get off on the wrong foot when they try to implement Kraljic’s
portfolio approach for the very first time. They underestimate the preparation phase and
the choices that have to be made, when implementing a purchasing portfolio approach.
When applying the portfolio approach for the first time, it is recommended to start with a
relatively simple design and a relatively small scope. For successful further adoption, the
first experience should be positive and should result in a (predictable) success. Within
easy reach are results in the non-critical category, where bundling of items is possible
and where a switch toward the leverage category is possible. In our experience,
companies without much purchasing sophistication are usually faced with a non-critical
quadrant ‘packed’ with low value items that are ordered individually from a large
number of suppliers. Another option would be to identify the largest uncovered supply
risks and try to find a solution. It is not unusual for companies not to have contracts for
some of their main strategic products. Based on the successes of a pilot project,
management may be excited about the further possibilities. In the long term, a successful
application of the portfolio approach is only possible when it is actually supported by the
highest executive officer that is responsible for purchasing.
Anyone who wishes to get going with the portfolio analysis, is confronted with a number
of practical problems and decisions. A first issue is the level of aggregation: whether
individual items, or smaller or broader product groups will be positioned in the matrix.
What is the unit of analysis? A decision has to be made at the very start of any portfolio
project. The purchasing portfolio analysis allows for very different modes of application.
Depending on the objectives, conditions and available time and resources, the following
considerations might be helpful, when deciding on aggregation issues:
274
- Will the analysis still be practicable? The positioning of individual items at the lowest
aggregation level means that data are needed on thousands of items. Some of them will
be very similar and many of them will be of very low value. Obviously, often this will not
be possible nor will it contribute to the required insights. A higher level might be
preferable, for instance at the level of item groups.
- Will the analysis still be feasible, considering issues of time and costs? It can not be
denied that a portfolio analysis could cost a lot of time and money, sometimes more than
estimated in advance. In general, the lower the level of aggregation and/or the broader
the scope of the analysis (including more organizational units/product groups), the
higher the costs of portfolio.
- Will the insights still be relevant? To control costs and complexity, it might be necessary
to work with item groups or larger categories, or to restrict the analysis to a single
department or organizational unit. However, an aggregation level that should be
considered as ‘too high’, will amount to nothing. For instance, if all components would be
grouped into a single category, then it is expected that no general assessment of ‘the’
supply risk is possible and that no general recommendations can be provided.
A useful advice would be to start with a classification of products that is known within
the company, for instance the accounting department or industry standards. Industrial
companies could start with the bill of materials that provide a full list of all needed items.
Subsequently the decision can be made to change or adapt the level of aggregation.
A second issue concerns measurement. Portfolio users should decide on the measurement
method, including decisions on factors, scores, and weights. Our research has shown that
experienced professionals handle these issues with respect to the measurement in
different ways. Three different measurement methods were identified:
- the ‘consensus’-based method,
- the ‘one-by-one’ method, and
- the ‘weighted factor score’ method.
The ‘consensus-based’ method is predominantly based on a process of reasoning and
discussing. The reaching of consensus is very important when choices are made with
respect to the measurement of variables and factors, and ultimately for the positioning of
items/products in the matrix. Users are content with the flexibility and possibilities of this
consensus based approach. Quite a different approach is the ‘one-by-one’ method. Just one
key variable is selected per dimension. Profit impact is usually measured as the financial
value of items, the supply risk is mostly operationalized by the number of available
suppliers (not contracted suppliers). As a result, positions in the matrix can be
determined in a rather quick and unambiguous way. A related benefit is that the method
allows for the comparison of different matrices that all use the same variables. The
‘weighted factor score’ method is well-known and allows for a completely customized
approach, when deciding on factors, weights, and (usually) scores.
The decision on the measurement method can be based on the following selection criteria,
that are derived from the specific advantages and disadvantages of the methods:
- the required objectiveness (high?, then 1-to-1)
- number of key factors (high?, then consensus or weighted factors)
- available time (‘no’ time?, then consensus or 1-to-1)
- needed customization and flexibility (high?, then weighted factors)
The positioning of items in the matrix should be considered as the starting point of
portfolio analysis, definitely not the finishing point. It is imperative that users reflect on
the results. If necessary, manual adjustments should be made. We consider in-depth
discussions on the positions in the matrix as the most important aspect of the analysis. To
our view the Kraljic framework facilitates these important discussions to a large extent.
To conclude, adopting a portfolio approach could work as a catalyst for change within the
company. The Kraljic matrix provides a practical framework for the non-purchasing
specialist, when analyzing and discussing purchasing issues. We have found that the
utilization of a portfolio approach is associated with higher levels of purchasing’s
professionalism (development of skills and competences) and in a better position of
purchasing within the company (recognition, status, contribution to the overall
competitive position). The portfolio project could put purchasing higher on the
company’s strategic agenda, clarifying the problems and possibilities of purchasing and
supplier management.
7.3.2 A 5-step approach to the application of purchasing portfolio analysis
In this section we will introduce and describe an outline for the application of the
portfolio analysis. It consists of 5 successive steps and the main questions and issues that
have to be addressed (see figure 7.6).
276
1. Preparation
2. Design and filling in the matrix
3. Interpretation of results
4. Strategic actions
5. Evaluation and follow-up
FIGURE 7.6 A 5-step approach to portfolio analysis
Ad 1 The preparation
The complete process of the portfolio analysis could take quite some time. Therefore it is
imperative to start with an adequate preparation. For reasons of support and
implementation a cross-functional team is required, with representatives from all
relevant departments and specialist fields. Considerable time and energy will have to be
spent in the preparation phase, getting the organization ‘ready for action’. Top
management will have to be convinced of the project, but also (line) managers from other
departments. The first step requires the formation of a team and a clear description of its
mission. The preparation phase is completed when the following questions are answered:
- What is the objective of the analysis?
- What information is available and/or needed?
- What are the limiting conditions in terms of time and money?
- Who will participate in the team?
Ad 2 Design and filling in the matrix

The next step is designing and filling in the matrix. In designing the matrix the main
issues are the level of aggregation, the selection of dimensions, variables and their
operationalizations, and the measurement method. In determining the level of
aggregation, there is a wide range of possibilities. This can vary from individual
components to complete assemblies or even suppliers. The first option might result in the
positioning of thousands of components, which will lead to an unworkable situation. As
a rule, the grouping of items will be necessary, but without an unacceptable loss of
information and relevance. The level of aggregation should be linked to the level on
which it is useful to select (differentiated) purchasing strategies. This means that major or
unique purchases should always be classified individually, while other purchases should
be classified by commodity groups or type of buy. If there are important individual items
within a commodity group with significant other values on the two dimensions of the
matrix, those should be taken out and analyzed separately. To conclude, there should be
a balance between relevance and practicability, when deciding on the level of
aggregation.
The dimensions ‘profit impact’ and ‘supply risk’ can be interpreted in many different
ways. What are the constituting factors? Kraljic stated the following factors:
For profit impact:
- Purchase volume
- Percentage of total purchase value
- Impact on product quality
- Impact on market growth
For supply risk:

- Availability/scarcity of the product
- Number of available suppliers
- Substitution possibilities
- Possibilities to switch between make or buy
- Risk of transport and storage.
However, as recognized by Kraljic, these lists are not exhaustive. Other factors could be
relevant for both dimensions. Within the project team consensus should be found
concerning the used variables and the way these variables are measured. To avoid
misunderstanding and unproductive discussions later on, the team has to come to an
agreement on the factors that constitute profit impact and supply risk. To get a first
impression of the supply risk, one could always consider the number of (alternative)
suppliers and the ease of switching to another supplier. To get a first impression of profit
impact, one could look at the purchase value of the products (price x volume). From these
points of departure the team should come to a tailor-made solution. In addition the team
should decide on the measurement method, as discussed in section 7.3.1.
Ad 3 Interpretation of the results

Non-experienced portfolio users might underestimate the interpretation of a filled-in
matrix. They might be in the supposition that the strategic actions can be derived from
the matrix in a rather deterministic way: ‘if bottleneck, than strategy X’. We have
278
emphasized earlier that the positioning of items in the matrix is the starting point of
portfolio analysis, not the finishing point. Portfolio users should reflect on:
- the positions of the items;
- the consequences of these positions.
Leading questions are ‘can we explain the found results?’ and ‘what do they actually
mean? The team is likely to benefit form a critical comparison between prior, sometimes
implicit, expectations and the actually found positions in the matrix. There might be some
disturbing discoveries, challenging the team to a more profound understanding of the
state of affairs. It might be revealed that, in contrast with prior beliefs, there are no
leverage items at all, or that all strategic items are positioned at the extreme right side of
the matrix, indicating that the company is facing extreme supply risks. The team should
want to know the reasons for such findings and the team should want to reflect on the
gravity of the situation. What is acceptable and what is not? In addition, the team should
take into account detailed information on with the overall business strategy, their
relationship with this strategy, the situations on supply markets, and the capacities and
intentions of individual suppliers. The preceding considerations form the necessary input
for the next step of the portfolio analysis.
Ad 4 Defining strategic actions

The fourth step is dedicated to the selection of strategic responses to the positions in the
portfolio matrix. We have stressed that one should always be aware of the fact that there
are various options to consider for each product category in the matrix. The leading
question is: “Which improvements are desirable, necessary, and/or feasible?” As a rough
guideline one should recognize that there are strategies to hold a position (1) and
strategies to move to another position (2). Obviously, changing positions will not always
be possible. Therefore, we have introduced the principle of ‘conditional dynamics’ (see
section 8.2.3). Whenever specific conditions are met, it might be possible to switch to
other, more favorable matrix positions. This study has identified the most common
switches within the matrix and the conditions that allow for these movements.
Ad 5 Evaluation and follow-up

The selected strategies should be translated into day-to-day actions. Without a specific
action plan nothing will probably happen. That would mean that all the effort of the
analysis was in vain. Obviously, defining the strategy is not the end of the 5-step
approach.
In fulfilling these action plans the positions of the items in the matrix might change. The
positions in a portfolio matrix are a reflection of the current situation, providing a
snapshot in time. Purchasing practitioners will benefit most if they can compare the
matrix with the matrix from an earlier period. After a reasonable amount of time the
positions of the items should be assessed again, allowing for control and intervention.
Shifts might require adjustment of the strategic actions.
Portfolio analysis should be included in the overall process of planning and control. This
means that the portfolio analysis should be conducted at least once a year. And it means
that portfolio-based strategies should be aimed at specific portfolio-based objectives.
These objectives should be concrete, measurable and they should be included in the
purchasing plan.
To conclude, practitioners might benefit from this 5-step approach. The approach is based
on the observations and the interviews during the case studies. Distinctive features are:
- it identifies the sequence and the substance of the main activities;
- the approach is based on insights form experienced portfolio users;
- it recognizes different possibilities of designing the matrix and of handling
measurement problems;
- it emphasizes that reflection on a (filled) matrix is more important than the matrix as
such (‘the matrix is the starting point of the analysis, not the finishing point’);
- it elaborates on the possibilities to formulate differentiated purchasing and supplier
strategies, including a focus on moving items to better positions in the matrix, whenever
the conditions allow for these strategic switches.
7.3.3 Conditional dynamics in the matrix: managing power and dependence
In this section we will address the issue of strategic switch in the portfolio matrix: how
are changes pursued and under which conditions? For these problems we have found the
principle of conditional dynamics: switching to another position in the matrix will
depend on the conditions that make it desirable and feasible. As we shall see, there is a
close connection with the problems of how to manage power and dependence in supplier
relations. We will provide guidelines for each category in the matrix, elaborating on
practical questions: ‘what do we need to know?’ and ‘what can we do?’
(1) Bottleneck items

Guiding principle: The bottleneck quadrant should be as empty as possible
Potential switch: Towards the non-critical quadrant
Conditions: Possibilities for standardization and availability of alternative
suppliers.
Dependence: Aimed at reducing the buyer’s dependence
What do we need to know?

For the assessment of bottleneck situation we need to have detailed information
regarding suppliers and supply markets on the one hand and product specifications on
the other hand. Some items have become a bottleneck as a result of over-specification. In
many industrial companies there is a natural desire to want “the best”. It is imperative to
find out whether limiting specifications are really necessary or merely desirable. The
280
leading principle should be: ‘fit-for-use’ (value analysis). Supply markets can be explored
differently with new, less restrictive specifications in search for appropriate suppliers.
Sometimes it will be possible to snare a supplier who is currently supplying a leverage or
strategic items: cross sourcing. Purchasers should be alert to the possibilities within the
current supply base. Insourcing might be another option, resulting from a make-or-buy
analysis.
What can we do?

Our study revealed that the logistical indispensability of bottleneck items has a
significant impact on the level of buyer’s dependence. If we can reduce our dependence
on the supplier, then the bottleneck item will be moved towards the non-critical
quadrant. Should that not be possible, then we will have to accept and deal with the
bottleneck position: ‘stay in the corner and make the best of it’. Without removing the
causes for our dependence on the supplier, we should reduce any detrimental
consequence of being dependent. The responsibility of the buyer might be:
- to close long-term contracts, if necessary under unfavorable conditions;
- to include a clause in the contract which says that the supplier should inform the buyer
betimes whenever he is considering to terminate the production of the item;
- to conduct a further risk analysis and make a contingency plan;
- to keep extra safety stocks if no alternatives are feasible.
(2) Non-critical items

Guiding principle: The non-critical quadrant should only contain items that have
to be ordered individually, otherwise they should be pooled.
Potential switch: Towards the leverage quadrant.
Conditions: The pooling of requirements is possible, leading to a volume of
trade which is of sufficient financial magnitude from a
supplier’s perspective.
Dependence: Aimed at increasing the supplier’s dependence.

We have to find out whether the pooling of requirements is possible, which will probably
imply the standardization of specifications. The benefits of pooling can only be gained
when implemented on a large scale, contracting a single, preferred supplier. A catalogues
system (e-procurement) might be feasible. In terms of power and dependence, we are
aiming at a more powerful position by increasing the supplier’s dependence (financial
magnitude).
What can we do?

A switch toward the leverage quadrant has significant consequence: bundling of items
into larger packages, supply base reduction, tendering and negotiating a beneficial
contract with a single supplier (systems contracting). Whenever bundling is not an
option, the items will have to be ordered individually at relatively high cost. Purchasers
may have to deal with internal political arguments rather than technical or economical
arguments against pooling. Companies with an underdeveloped purchasing function are
likely to have a non-critical quadrant which is overflowed by a mass of ‘special’ items.
They obviously are not making appropriate use of their potential buying power.
(3) Leverage items

Guiding principle: The leverage quadrant is allowed to be as full as possible.
Potential switch: In exceptional cases, towards the strategic quadrant.
Conditions: The supplier is able and willing to cooperate and contribute
more significantly to the company’s competitive position.
Dependence: Aimed at an increase of mutual dependence.

It should be emphasized that a leverage position in the Kraljic matrix does not always
imply buyer dominance. The vertical axis of the matrix measures the financial value of
items (‘profit impact’) for the buyer. It should be recognized however, that the distinction
between low and high values is not determined from the supplier’s perspective. A certain
amount of money could be assessed as ‘ high’ by the buyer, but at the same time as ‘low’
by the supplier. This might be the case for commodities that are purchased in
oligopolistic supply markets. It is therefore imperative that the buyer knows how
important the volume of trade actually is for the supplier, largely indicating the level of
supplier’s dependence. The purchaser is usually in a comfortable position, making
advantage of his buying power.
Occasionally there might be a possibility for a further development of the buyer-supplier
relationship into a more balanced and mutually committed partnership. An in-depth
assessment of the supplier, including his capabilities and intentions, should provide the
required information. Manufacturing companies will only want key suppliers to enter the
strategic quadrant. Such main suppliers are capable of co-design and co-makership in
combination with a positive attitude toward the partnership.
What can we do?

In most cases leverage items will not be moved to the strategic quadrant. The buyer
usually wants to make advantage of the many possibilities of a leverage position, such as:
- obtaining favorable contractual conditions;
- making use of the buying power in negotiations;
282
- applying competitive bidding;

- closing short term contracts, aimed at maximum supplier performance;
- periodical supplier evaluation in combination with a periodical decision process
whether to switch or not to another supplier.
In this thesis we have introduced the term ‘partner of convenience’ for suppliers in the
leverage quadrant. The implication of the concept should not be misunderstood. Such
suppliers are in fact usually very important for companies and not just for the delivery of
office supplies (facility buying). Companies need to contract smaller technologically
advanced firms with an excellent record on quality, delivery times, flexibility and prices.
These firms are asked to contribute to the optimization of the logistical coordination in
the supply chain.
Alternatively, there is much more to the scope and the impact of a ‘real’ strategic
partnership, compared to a partnership of convenience. The development of a strategic
partnership entails mutual trust and commitment, which results in a higher level of
mutual dependence. Obviously, from the perspective of the buying company a voluntary
rise of the buyer’s dependence will have to be matched by a proportional increase of the
supplier’s dependence. The partners will have a mutual need for each other’s
technological expertise and will both face higher switching costs, due to relationship
specific investments. The chosen supplier will have to enter a partnership program,
based on a mutual agreement on the intention, direction and expectations of the
partnership. Early supplier involvement is a logical consequence of the technological
collaboration that will be important in any strategic supplier partnership.
(4) Strategic items

Guiding principle: The strategic quadrant should preferably only contain items
that are purchased from high-performing suppliers on the basis
of mutual trust, commitment and partnership.
Potential switch: In case of unsatisfactory ‘partnerships’ towards the leverage
quadrant.
Conditions: Availability of alternative suppliers (not locked-in) and a
limited need for the supplier’s technological expertise.
Dependence: Aimed at reducing the buyer’s dependence.

Mutual dependence is generally considered as a necessary condition for any long-lasting
collaboration with suppliers that is to the satisfaction of both parties. From the
perspective of the buyer it is important to guard against a disproportional increase of the
supply risk and the vulnerability of the company. The dependence of the buyer could
result in insurmountable switching costs. A critical point of interest however, is the actual
supplier’s dependence. The buyer will have to gain in-depth knowledge of the position of
the supplier. What is the level of the supplier’s dependence on the buying company? Are
there any alternative buyers? What are the relationship-specific investments, resulting in
what kind of switching costs? Etc. The survey data of our study indicated that buyers
perceive lower levels of supplier’s dependence, compared to the levels of buyer’s
dependence in the strategic quadrant. They perceive that they encounter higher
switching costs and have fewer alternative trading partners. The buyer should make sure
whether these data are facts, or whether they are unsubstantiated assumptions. Another
point of interest is to know what produces this supplier’s dependence. The buyer should
make sure whether the strategic partners are committed to the partnership and are
equally dependent on their trading partner. Obviously, the excellent performance of the
supplier should be beyond discussion. In the course of time partnerships may become
unsatisfactory. The relationship might be disintegrating (‘falling asleep together’). There
is always a chance that a partnership evolves into an indolent relaxed relationship. The
buyer should be alert and assess periodically whether the performance expectations are
up-to-standard and whether they are met in practice. External benchmarks will be
valuable for this purpose.
So far, we have discussed the relationships with suppliers that resulted from a conscious
and voluntary choice. Our study has identified the possibilities of involuntary stays at the
strategic quadrant. Some relationships get off on the wrong foot. A company might be
forced by its customer into a locked-in relationship with a supplier. It is a known fact that
large companies in many occasions specify which suppliers are to be used by their first-
tier suppliers, because the various components have to fit into the end product.
Alternatively, buyers have to deal with suppliers who possess a patent or a concession,
resulting in a monopolistic market situation. Such involuntary relationships with
important ‘strategic’ suppliers result in positions in the strategic quadrant in the Kraljic
matrix: high profit impact and a high supply risk. However, they are to be distinguished
sharply from the more satisfactory relationships that are founded on voluntary
collaboration. A position in the ‘strategic’ category does not mean a high mutual
dependence. In case of a locked-in ‘partnership’ the supplier’s dependence is rather low,
just like the values of key relationship variables such as trust and commitment.
To conclude, the user of the portfolio approach should know for each strategic item the
nature of the relationship from a power and dependence perspective resulting in a
classification into satisfactory partnerships (1), unsatisfactory partnerships on a voluntary
basis (2) and locked-in relationships on an involuntary basis (3).
284
What can we do?

Obviously, the managerial problems with satisfactory partnerships are completely
different, compared to the handling of unsatisfactory partnerships and locked-in
relationships. In most cases there is not much one can do, when faced with a locked-in
relationship, whatever the reason might be. The buying company will have to deal with
the consequences of disadvantageous circumstances or decisions from past. It will be
very hard to convince major customers that alternative suppliers have to be admitted,
considering the unsatisfactory performance of the ‘locked-in partner’. In other cases, the
buyer has no real alternatives whatsoever, because the supplier has an indispensable
patent position. Whenever locked-in into the strategic quadrant, there is a close
resemblance to the locked-in positions in the bottleneck quadrant. In line with the
recommendations for dealing with bottleneck items, the buyer should try to reduce the
negative consequences and impact of an unfavorable position. The buyer could try to
urge the supplier to a better performance. The buyer could identify specific risks,
renegotiate the contract, and try to cover these risks as well as possible. These efforts
however, will not remove the causes for the high level of buyer’s dependence.
Alternatively, there will be cases of unsatisfactory partnerships based on voluntary

collaboration. The buyer has more possibilities for improvement, although it should be
recognized that a switch to the leverage quadrant will not be an easy one to accomplish.
The buyer’s dependence is high, just like the cost of terminating the old relationship and
developing a new relationship with another supplier. Are there possibilities to a re-
education? Is the magic over, of does the supplier deserve a second chance? Should we
invest in the relationship with this supplier and its performance, or not? Very important,
what are the switching costs that will be endured, when putting an end to the
relationship? The development of a new supplier costs time and money, but it could solve
the current problem in the strategic quadrant.
In a general sense, it is wise to maintain good relationships with potential suppliers. That is
to say, companies that are currently not contracted. As we have found in the Akzo Nobel
Coatings-case, potential suppliers may provide alternative arrangements in cases of
emergency or problems with the current suppliers (flexibility, fall back option).
Unfortunately, in case of strategic items such alternative suppliers will not always be
there.
7.4 Recommendations for future research

In this closing section we will give some recommendations for further study, beyond the
usual recommendations to replicate the study to other industries and/or other countries.
(1) Impact and implementation

Future research could include an empirical study to the actual, direct impact of the
application of a portfolio approach. The required research methodology should begin
with the development of performance measures. A distinction should be made between
the impact of the introduction of the tool in companies (first-time use) and the impact of a
long-standing application (repeated use). Adopting a portfolio approach could work as a
catalyst for change, leading the way to a more professional, mature and sophisticated
purchasing function. Immediate success is warranted, considering the new insights that
are to be attributed to portfolio use. To gain a deeper understanding of the adoption of
the portfolio tool, we recommend a series of action research studies aimed at identifying
normative guidelines for the implementation and the assessment of the full impact of the
portfolio approach. These action research studies should include ‘before’ and ‘after’
measurement of key variables in order to determine accountable changes. Another
possibility would be the use of critical incident techniques to shed more light on complex
managerial problems relating to the development and implementation of portfolio-based
Perhaps even more challenging would be some research focussed on the impact of a
repeated use, in terms of performance measures that count to top management. Only
longitudinal studies in companies could provide information about the long-term impact
and usefulness of the portfolio approach. Such research requires a complex design. The
researcher should overcome the difficulties of attributing results of portfolio use and of
comparing the use of the tool in different companies, because several company-specific
factors are likely to influence the impact and implementation of portfolio use. In addition
the personality of individual purchasers could be included as well, describing and
explaining the use and effectiveness of the portfolio approach.
(2) A network perspective

We have investigated the portfolio approach to the development and selection of
purchasing and supplier strategies. A portfolio approach views the supply base of the
company as a portfolio of items and/or supplier relationships that have to be managed in
a differentiated as well as integrated way. A matrix serves as the framework and point of
departure for the development of appropriate purchasing strategies, differentiated of the
various types of dyadic relationships with suppliers. However, there is support for a
research tradition that does not agree with the focus on dyadic relationships as the unit of
analysis. We are referring to the basic principles of the Industrial Network approach, also
known as the Markets-as-Networks tradition. In their view research should be focussed
on the behavioral actions within the wider network. McLoughin and Horan (2000: 289)
for instance claim that “if one wanted to understand the process of exchange in one
relationship, this could only be done by understanding the wider network of
relationships within which the exchange takes place”. As Gadde and Snehota (2000:315)
286
put it: “… if we are to understand the interactive nature of customer-supplier

relationships (…) the scope of the analysis needs to be broadened. Each relationship is
dependent with a number of other relationships, together forming a network.” For
example, when sourcing a new supplier, a buyer should consider the relationships that
that supplier has with competing firms and perhaps with major customers. Developing
the relationship with a certain supplier might give access to the supplier’s other
relationships and their capabilities and resources. Many large companies specify which
suppliers are to be used by their first-tier suppliers, mainly because particular critical
components have to fit with other critical components (Johnsen et al., 2000). In the TE
STRAKE case study we have acknowledged that companies can be obliged to enter into
forced ‘partnerships’, mainly because the purchasing and supply strategies have to
support the overall business strategy that focuses on the demands and requirements of
the major customers.
Without adding to an alleged controversy between a dyadic approach and a network

approach to buyer-supplier relationships, we feel that there might be a case of false
dichotomy. A network perspective does not exclude research which seeks to understand
the nature of single buyer-supplier relationships (e.g. Brennan and Turnbull, 1999;
McLoughlin and Horan, 2000). After all, it is recognized that dyadic relationships are the
building blocks of networks (e.g. Harland, 1996). Obviously, companies will have to
manage their dyadic exchange relationships with suppliers and customers. There are
academics who believe that industrial networks cannot be managed, actors within them
merely cope (Håkansson and Snehota, 1995). Lamming et al. (2000) agreed that networks
cannot be managed, as it is impossible to control the activities and directions of other
companies, which led Johnsen (2000) to the conclusion that “A better focus is the
relationship between firms as the manageable entity”. However, it would be worthwhile
to investigate the possibilities of transferring ideas and concepts (such as
interconnectedness) into a purchasing portfolio approach. Is it possible to conceive a
network as a manageable portfolio of interconnected relationships between actors? From
the network aggregation level several questions have to be addressed. What are the
possibilities to develop a matrix with a classification of relationships that reflects the
resource ties between companies and the contributions to the overall network
performance? What strategic recommendations could be given on how to manage
network positions, and how to defend or develop them? Alternatively, we could return to
the buyer’s perspective and investigate the impact and adaptations when including a
network perspective. Perhaps we could use the ‘old’ Decision Making Unit-concept,
borrowed from the marketing discipline, identifying the different types of actors and
their functions and influence in a network.
(3) Measuring power and dependence

We have acknowledged that the relative power of an organization over another is the
result of the net dependence of one on the other (cf. Pfeffer, 1981; Dickson, 1983; Kumar
et al., 1995; Geyskens et al., 1996). Our study has used this conceptualization for the
measurement and quantification of ‘power’. We have actually determined numerical
values for the relative power in buyer-supplier relationships, calculated as the difference
between the supplier’s dependence and the buyer’s dependence. In addition we have
specified and quantified regression models to explain the buyer’s dependence and the
supplier’s dependence in various portfolio-based scenarios. We have used corresponding
variables to explain buyer’s and supplier’s dependence (see figure 7.7).
The data that we have gathered allow for the numerical specification of a new type of
model that aims to explain ‘power’ in buyer-supplier relationships by corresponding
determinants of buyer’s and supplier’s dependence.
If:
P = balance of power = supplier’s dependence – buyer’s dependence
supplier’s dependence = f (financial magnitude, need for buyer’s expertise, availability of

alternative buyers, switching costs for supplier) = f (X1s, X2s, X3s, X4s)
buyer’s dependence = f (logistical indispensability, need for supplier’s expertise,

availability of alternative suppliers, switching costs for buyer) = f (X1b, X2b, X3b, X4b)
then we could specify:
P = a + ß 1 (X1s – X1b) + ß 2 (X2s – X2b) + ß 3 (X3s – X3b) + ß 4 (X4s – X4b)
In words: the balance of power is to be explained by variables that are calculated as the
differences between corresponding determinants of buyer’s and supplier’s dependence.
Of course, other variables could be used as well. The proposed operationalization and
model specification might be a promising point of departure for quantitative research to the
issues of power and dependence in buyer-supplier relationships. The value of the balance
of power might be contingent to the sizes of the buying and the supplying companies.
Alternatively, network positions or the positions in the supply chain could be included as
a contingent factor.
288
Perspective
dependency supplier’s dependence buyer’s dependence

component
dependency for the fulfillment of the financial magnitude logistical indispensability

primary function X1s X1b
technological dependency need for buyer’s expertise need for supplier’s expertise
X2s X2b
sources dependency availability of availability of

alternative buyers alternative suppliers
X3s X3b
economic dependency switching cost supplier switching cost buyer

X4s X4b
FIGURE 7.7 Corresponding determinants of buyer’s and supplier’s dependence
Cox (2001) argued in favor of a power perspective on extended networks of buyer-

supplier relationships. An analytical framework was presented, emphasizing the value
appropriation in complex power regimes. Value appropriation refers to the net operating
profits earned by companies participating in a supply chain. The framework predicts
which parties will benefit more than others, based on the power relations in the various
dyadic buyer-supplier relationships. It would be a challenge to operationalize the
concepts and control the predictions of the model. It is worthwhile to investigate the
impact of power and dependence on the distribution of profits. More in general, further
research should be directed towards the importance, impact and determinants of power
relations in chains of interdependent companies (supply chain). We would like to break a
lance for qualitative research which allows for in-depth listening to key informants, as
opposed to ‘completing questionnaires’. Faria and Wensley (2002) investigated the ways
in which first-tier suppliers responded to changes in customer requirements. Their
research is based on the analysis of managerial narratives which revealed the
discrepancies with mainstream management theories on networks and supply chain
management. Surprising discoveries were for example the substantial expressions of
power and conflict, the concerns with money and power, and the complete absence of the
end customer in many of the narratives. To conclude, we would welcome research that is
aimed at closing the gap between the mainstream management literature and the daily
reality of power and dependence in buyer-supplier relations.
(4) E-Procurement
Business world has observed an unprecedented hype regarding the Internet and e-
business, which ended with the dotcom crisis of 2001. In the late nineties e-procurement
was considered as the ‘revolution through electronic purchasing’. For instance Essig and
Arnold (2001) predicted that “without doubt, e-procurement will dramatically change the
way purchasing is done in the near future”. By now, we know that the predictions
regarding revolutionary changes, have not proven correct. Leonard and Spring (2002)
posited that the emphasis so far in applications of e-procurement has mainly been to
release buyers from administrative chores, although such systems should support the
strategic aspects of purchasing. Nevertheless, progress has been made, although at a
much slower pace than expected. Companies are keeping distance from sky-high
investments in e-procurement systems, whereas most e-market places are (mid 2002)
faced with serious financial problems. Alternatively, more positive experiences can be
observed with the implementation of catalogue systems, e-reverse auctions, e-tendering
and e-sourcing.
E-procurement applications have produced completely new subjects of research. In
contrast, a rather limited number of empirical studies have reported on the actual
implementations and impact of the various e-procurement solutions. In line with the
issues raised in this study, we would recommend research to the relationship between e-
procurement and the purchasing portfolio approach. Leonard and Spring (2002) have
outlined some hypothesized effects of e-procurement in each of the quadrants of the
Kraljic matrix:
- Non-critical items: e-procurement can assist internally in building a catalogue of items
against which decentralized units can place routine orders, consolidate these to facilitate
logistics and the measurement of delivery and quality performance.
- Leverage items: the role of e-procurement is in reducing the cost of requesting, receiving
and analyzing tenders, so that it will be no more expensive (in time) to have fifty firms
tendering than to have only three; ICT may also help interaction that precludes the need
for visits and meetings (email, video conferencing, drawing exchange by CAD file, etc.).
- Bottleneck items: e-procurement should offer some help, widening the search on the
whole world at minimal expense, and could support iterative interaction with potential
suppliers, for example in determining the design of a suitable replacement involving off-
the-shelf parts.
- Strategic items: the main active use of ICT within partnerships is anticipated to be in the
reduction of cost and improvement of effectiveness and interaction and information
transfer. This could range from simple messages transferred by e-mail, to virtual
meetings of multi-site and multi-firm product-development teams.
Future research could be of an explorative and descriptive nature, identifying actual e-
procurement practice in relation to a portfolio approach. Conceivable research questions
are: Which e-procurement solutions are actually deployed in the item categories of the
290
Kraljic matrix? Which strategies are selected and under which conditions? What is their
impact on buyer-supplier relationships? What is their impact on performance measures
like cost savings, reduced cycle times, and return on investments? As Van Weele (2001)
pointed out, powerful suppliers will not always be willing to participate in the e-
procurement solutions of buying companies. Virtual auctions seem appropriate for
leverage items, while catalogue systems could be applied for the handling of non-critical
items. Again, research should shed light on the relationship between e-procurement
solutions and power and dependence in buyer-supplier relationships.
291
References
Achrol, R.S. and G.T. Gundlach (1999), ‘Legal and Social Safeguards Against Opportunism in
Exchange’, Journal of Retailing, Vol. 75, No. 1, 107-124.
Åhman, S. (2002), ‘Strategic Sourcing of Suppliers in a Supply Network’, Proceedings of the 11th
International IPSERA Conference, Enschede, The Netherlands, 1-10.
Aldrich, H. and S. Mindlin (1978), ‘Uncertainty and Dependence: Two Perspectives on

Environment’, in: L. Karpik (ed.), Organization and Environment, Sage Publications Ltd, London, 149-
170.
Anderson, J.C. and J.A. Narus (1984), ‘A Model of the Distributor’s Perspective of Distributor-
Manufacturer Working Relationships’, Journal of Marketing, Vol. 48, Fall, 62-74.
Anderson, J.C. and J.A. Narus (1990), ‘A Model of Distributor Firm and Manufacturer Firm Working
Partnerships’, Journal of Marketing, Vol. 54, January, 42-58.
Anderson, J.C. and B.A. Weitz (1989), ‘Determinants of Continuity in Conventional Industrial
Channel Dyads’, Marketing Science, Vol. 8, Fall, 310-323.
Anderson, J.C. and B.A. Weitz (1992), ‘The Use of Pledges to Build and Sustain Commitment in
Distribution Channels’, Journal of Marketing Research, Vol. 29, February, 18-34.
Anderson, J.C., H. Håkansson and J. Johanson (1994), ‘Dyadic Business Relationships Within a
Business Network Context’, Journal of Marketing, Vol. 58, October, 1-15.
Ansoff, H.I. (1984), Implanting Strategic Management, Prentice Hall, New York.
Araujo, L., A. Dubois, L.-E. Gadde (1999), ‘Dyadic Business Relationships within a Business
Network Context, Journal of Marketing, Vol. 58, 1-15.
Archer, N.P. and F. Ghasemzadeh (1999), ‘An Integrated Framework for Project Portfolio Selection’,
International Journal of Project Management, Vol. 17, No. 4, 207-216.
Armstrong, J.S. and T.S. Overton (1977), ‘Estimating Nonresponse Bias in Mail Surveys’, Journal of
Marketing Research, Vol. 14, No. 8, August, 396-402.
Armstrong, J.S., R.J. Brodie and J. Ringbeck (1988), ‘The Value of Portfolio Planning Methods for
Strategic Marketing Decisions: Experimental Results’, Working Paper, August.
Armstrong, J.S. and R.J. Brodie (1984), ‘Effects of Portfolio Planning Methods on Decision Making:
Experimental Results’, International Journal of Research in Marketing, Vol. 11, 73-84.
Arndt, J. (1983), ‘The Political Economy Paradigm: Foundation for Theory Building in Marketing’,
Journal of Marketing, Vol. 47, October, 44-54.
292
Asanuma, B. (1989), ‘Manufacturing Supplier Relationships in Japan and the Concept of Relation-
Specific Skill’, Journal of Japanese and International Economies, Vol. 3, 1-30.
Axelsson, B., K. Frisk and A. Najib (2000), ‘From Buying to Supply Management – A Multifaceted
Change Venture’, Proceedings of the 9th International IPSERA Conference, London, Canada, 55-66.
Axelsson, B., J. Laage-Hellman and U. Nisson (2002), ‘Modern Management Accounting for Modern
Purchasing’, European Journal of Purchasing & Supply Management, Vol. 8, No. 1, 53-62.
Baatz, E. (1999), ‘How Purchasing Handles Intense Cost Pressure’, Purchasing, Vol. 126, No. 5, April
8, 61-66.
Bacharach, S. and E. Lawler (1975), ‘Perception of Power: An Empirical Test of Exchange Theory’,
Paper delivered at the Meetings of the Midwest Sociological Association, Chicago.
Bacharach, S. and E. Lawler (1981), Power and Politics in Organizations, Jossey-Bass, San Fransisco.
Barney, J.B. (1991), ‘Firm Resources and Sustained Competitive Advantage’, Journal of Management,
Vol. 17, 99-120.
Barney, J.B. and M.H. Hansen (1994), ‘Trustworthiness as a Source of Competitive Advantage’,
Strategic Management Journal, Vol. 15, 175-190.
Barney, J.B. (1996), Gaining and Sustaining Competitive Advantage, Addison-Wesley

Publishing Company,
Barringer, B.R. and J.S. Harrison (2000), ‘Walking a Tightrope: Creating Value Through
Interorganizational Relationships’, Journal of Management, Vol. 26, No. 3, 367-403.
Bauer, F. (1977), ‘Better Purchasing: High Rewards and Low Risks’, The Mac Kinsey Quarterly,
Winter, 75-86.
Bello, D.C., R. Lohtia and S.P. Dant (1999), ‘Collaborative Relationship for Component
Development: The Role of Strategic Issues, Production Costs, and Transaction Costs’, Journal of
Business Research,
Vol. 45, 15-31.
Benito, G.R.G., T. Pedersen and B. Petersen (1999), ‘Foreign Operation Methods and Switching
Costs: Conceptual Issues and Possible Effects’, Scandinavian Journal of Management, Vol. 15, No. 2,
213-229.
Bensaou, M. and N. Venkatraman (1995), ‘Configurations of Interorganizational Relationships: A

Comparison Between U.S. and Japanese Automakers’, Management Science, Vol. 41, No. 9, 1471-
1492.
Bensaou, M. (1999), ‘Portfolios of Buyer-Supplier Relationships’, Sloan Management Review, Vol. 40,
No. 4, Summer, 35-44.
Berger, H., N.G. Noorderhaven and B. Nooteboom (1995), ‘Determinants of Supplier Dependence:
An Empirical Study’, in: On Economic Institutions: Theory and Applications, J. Groenewegen, C. Pitelis
and S. Sjöstrand (ed.), Edward Elgar Publishing Limited, Hants, England.
Berry, L.L. and A. Parasuraman (1991), Marketing Services, The Free Press, New York.
Blau, P.M. (1964), Exchange and Power in Social Life, John Wiley, New York.
Blenkhorn, D.L. and H.F. Mackenzie (1994), ‘The importance of buyer-seller dependence in
relationship marketing’, in: Sheth, J.N. and A. Parvatiyar (ed.), Relationship Marketing: Theory,
Methods and Applications, 1994 Research conference proceedings, Emory University, Atlanta.
Blois, K.J. (1999), ‘Trust in Business to Business Relationships: An Evaluation of its Status’, Journal of
Management Studies, Vol. 36, No. 2, March, 197-215.
Blomqvist, K. (1997), ‘The Many Faces of Trust’, Scandinavian Journal of Management, Vol. 13, No. 2,
271-286.
Boodie, M.L.J. (1997), World Class Purchasing in Nederland is fictie en helaas nog geen werkelijkheid.
Berenschot Inkoopenquête 1997, Berenschot Inkoopmanagement, Utrecht (Dutch text).
Bourantas, D. (1989), ‘Avoiding Dependence on Suppliers and Distributors’, Long Range Planning,
Vol. 22, No. 3, 140-149.
Bowman, E.H. (1974), ‘Epistemology, Corporate Strategy and Academe’, Sloan Management Review,
Winter, 35-50.
Boyer, K.K., J.R. Olson, R.J. Calantone and E.C. Jackson (2002), ‘Print versus Electronic Surveys: A
Comparison of Two Data Collection Methodologies’, Journal of Operations Management, Vol. 20, No.
4, 357-373.
Brealy, R.A. and S.C. Myers (1996), Principles of Corporate Finance, McGraw-Hill, New York,
international edition.
Brennan, R. and P.W. Turnbull (1999), ‘The Managerial (Ir)relevance of IMP’, The 16th International
Marketing and Purchasing Conference, University College Dublin, Ireland.
Brown, J.R. and G.L. Frazier (1978), ‘The Application of Channel Power: Its Effects and
Connotations’, in Research Frontiers in Marketing: Dialogues and Directions, S.C. Jain, ed., Chicago:
American Marketing,
266-270.
Brown, J.R., R.F. Lusch and D.D. Muehling (1983), ‘Conflict and Power-Dependence Relations in
Retailer-Supplier Channels’, Journal of Retailing, Vol. 59, No. 4, Winter, 53-81.
Brown, R. (1991), ‘Making the Product Portfolio a Basis for Action’, Long Range Planning, Vol. 24,
No. 1, 102-110.
294
Buchanan, L. (1992), ‘Vertical Trade Relationships: The Role of Dependence and Symmetry in
Attaining Organizational Goals’, Journal of Marketing Research, Vol. 29, February, 65-75.
Burt, D.N. (1989), ‘Managing Suppliers Up to Speed’, Harvard Business Review, July-August, 127-135.
Buvik, A. and T. Reve (2001), ‘A Symmetrical Deployment of Specific Assets and Contractual
Safeguarding in Industrial Purchasing Relationships’, Journal of Business Research, Vol. 51, No. 2,
101-113.
Buvik, A. and Ø. Halskau (2001), ‘Relationship Duration and Buyer Influence in Just-In-Time
Relationships’, European Journal of Purchasing & Supply Management, Vol. 6, No. 7, 111-119.
Buzzell, R.D., B.T. Gale and R.G.M. Sultan (1975), ‘Market Share - A Key to Profitability’, Harvard
Business Review, 53, January/February.
Buzzel, R.D. (1983), ‘Is Vertical Integration Profitable?’, Harvard Business Review, January-February,
99-102.
Cagliano, R., F. Caniato and G. Spina (2002), ‘Supply Strategy Configurations: Linking Operational
Integration with Supply Management Practices’, Paper Presented at the 11th International IPSERA
Conference, Enschede, The Netherlands.
Calabrese, G. (2000), ‘Small-medium Supplier-Buyer Relationships in the Car Industry: Evidence

from Italy’, European Journal of Purchasing & Supply Management, Vol. 6, No. 1, 59-65.
Caldwell, N.D., C.M. Harland, Ph. Dowell, M. Woerndl and J. Zheng (2002), ‘The Role of E-business
in Large Firm Supply Chain Interactions with Manufacturing SMES: A Conceptual Framework’,
Proceedings of the 11th International IPSERA Conference, Enschede, The Netherlands, 123-134.
Campbell, N.C.G. and M.T. Cunningham (1983), ‘Customer Analysis for Strategy Development in
Industrial Markets’, Strategic Management Journal, Vol. 4, No. 4, 369-380.
Campbell, N.C.G. (1985), ‘An Interaction Approach to Organizational Buying Behavior’, Journal of
Business Research, Vol. 13, No. 1, February, 35-50.
Capon, N., J.U. Farley and J.M. Hulbert (1987), Corporate Strategic Planning, Colombia University
Press, New York.
Carter, J.R. and R. Narasimhan (1996), ‘Is Purchasing Really Strategic?’, International Journal of
Purchasing and Materials Management, Vol. 32, No. 1, Winter, 20-28.
Carter, J.R. (1997), ‘Supply positioning at SGX Corporation’, Best Practices in Purchasing & Supply
Chain Management, December, 5-8.
Carter, S. (1995), Successful Purchasing in a Week, Hodder & Stoughton, England.
Catry, B. and M. Chevalier (1974), ‘Market Share Strategy and the Product Life Cycle’, Journal of
Marketing, Vol. 38, No. 4, October, 29-34.
Christopher, M. (1992; 1998), Logistics and Supply Chain Management, Pitman Publishing, London, 1st
and 2nd ed.
Christipher, M. and U. Jüttner (2000), ‘Developing Strategic Partnerships in the Supply Chain: A
Practitioner Perspective’, European Journal of Purchasing & Supply Management, Vol. 6, No. 2, 117-127.
Clauwaert, F. (1993), ‘Portfolioanalyse’, Handboek Inkoopmanagement, Samsom, Alphen aan den Rijn,
C3500, 1-20 (Dutch text).
Coate, M.B. (1983), ‘Pitfalls in Portfolio Planning’, Long Range Planning, Vol. 16, No. 3, 47-56.
Cook, V. (1985), ‘The Net Present Value of Market Share’, Journal of Marketing, Vol. 49, Summer,
49-63.
Cooper, R. and R.S. Kaplan (1991), ‘Profit Priorities from Activity-Based Costing’, Harvard Business
Review, Vol. 69, May-June, 130-135.
Cooper, R.G., S.J. Edgett and E.J. Kleinschmidt (1998), Portfolio Management for New Products,
Addison-Wesley Longman, Inc.
Cooper, R.G., S.J. Edgett and E.J. Kleinschmidt (1999), ‘New Product Portfolio Management:
Practices and Performance’, The Journal of Product Innovation Management, Vol. 16, No. 4, 333-351.
Cooper, R.G., S.J. Edgett and E.J. Kleinschmidt (2000), ‘New Products, New Solutions: Making
Portfolio Management More Effective’, Research Technology Management, Vol. 16, No. 4, 18-33.
Corsten, H. (1996), ‘Beschaffung’, in: Betriesbwirtschaftslehre, Corsten, H. and M. Reiss, München,

609-736.
Cousins, P.D. (1999), ‘Supply Base Rationalisation: Myth or Reality?’, European Journal of Purchasing
& Supply Management, Vol. 5, No. 3-4, 143-155.
Cousins, P.D. (2002), ‘A Conceptual Model for Managing Long-term Inter-organisational

Relationships’, European Journal of Purchasing & Supply Management, Vol. 8, No. 2, 71-82.
Cox, A. (1995), ‘Strategic Procurement Management in the Public and Private Sectors’, in: The
Relative Benefits of Competitive and Collaborative Approaches, eds. Lamming, R. and A. Cox.
Cox, A. (1996), ‘Relational Competence and Strategic Procurement Management: Towards an

Entrepreneurial and Contractual Theory of the Firm’, European Journal of Purchasing and Supply
Management, Vol. 2, No. 1, 57-70.
Cox, A. (1997), Business Success - a way of thinking about strategic, critical supply chain assets and
operational best practice, Earlsgate Press, Great Britain.
Cox, A. (2001), ‘Understanding Buyer and Supplier Power: A Framework for Procurement and
Supply Competence’, The Journal of Supply Chain Management, Vol. 37, No. 2, Spring, 8-15.
Cox, W.E. (1967), ‘Product Life Cycles as Marketing Models’, The Journal of Business, Vol. 40,
October, 375-384.
296
Croom, S.R. (2000), ‘The Impact of Web-Based Procurement on the Management of Operating
Resources Supply’, The Journal of Supply Chain Management, Vol. 36, No. 1, Winter, 4-13.
Cunningham, M.T. and E. Homse (1982), ‘An Interaction Approach to Marketing Strategy’, in
Håkansson, H. (ed.), (1982), International Marketing and Purchasing of Industrial Goods: an Interaction
Approach, Wiley, Chichester, 328-345.
Cunningham, M.T. (1982), ‘An Interaction Approach to Purchasing Strategy’, in Håkansson, H.

(ed.), International Marketing and Purchasing of Industrial Goods: an Interaction Approach, Wiley,
Chichester, 345-357.
Cusumano, M. and A. Tekeshi (1991), ‘Supplier Relations and Management: A Survey of Japanse
Auto Plants and U.S. Auto Plants’, Strategic Management Journal, Vol. 12, November, 563-588.
Dacin, P.A. and D.C. Smith (1994), ‘The Effect of Brand Portfolio Characteristics on Consumer
Evaluations of Brand Extensions’, Journal of Marketing Research, Vol. 31, No. 2, 229-242.
Dahl, R.A. (1957), ‘The concept of power’, Behavioral Science, Vol. 2, July, 201-215.
Dant, R.P. and G.T. Gundlach (1998), ‘The challenge of autonomy and dependence in franchised
channels of distribution’, Journal of Business Venturing, Vol. 14, No. 1, 35-67.
Day, G.S. (1981), ‘Strategic Market Analysis and Definition: An Integrated Approach’, Strategic
Management Journal, Vol. 2, July-September, 29-38.
Day, G.S. (1986), Analysis for strategic market decisions, St. Paul, MN: West Publishing.
De Boer, L. (1997), ‘Het inkoopportfoliomodel van Kraljic nader beschouwd’, Praktijkboek

Professioneel Inkoopmanagement, Kluwer Bedrijfsinformatie, Deventer, 2.9.B.1-01 - 2.9.B.1-10
(Dutch text).
De Boer, L. (1998), Operations research in support of purchasing – design of a toolbox for supplier selection,
Doctoral thesis, University of Twente, Enschede.
De Jong, G. and B. Nooteboom (2000), The Causal Structure of Long-Term Supply Relationships - An
Empirical Test of a Generalized Transaction Cost Theory, Kluwer Academic Publishers, Boston.
DelVecchio, D. (2000), ‘Moving Beyond Fit: The Role of Brand Portfolio Characteristics in Consumer
Evaluations of Brand Reliability’, Journal of Product and Brand Management, Vol. 9, No. 7, 457-471.
Dhalla, N.K. and S. Yuspeh (1976), ‘Forget the Product Life Cycle Concept’, Harvard Business Review,
Vol. 54, January/February, 102-112.
Dick, A.S. and K. Basu (1994), ‘Customer Loyalty: Toward an Integrated Conceptual Framework’,
Journal of the Academy of Marketing Science, Vol. 22, No. 2, 99-113.
Dickson, P.R. (1983), ‘Distributor Portfolio Analysis and the Channel Dependence Matrix: New
Techniques for Understanding and Managing the Channel’, Journal of Marketing, Vol. 47, Summer,
35-44.
Dierickx, I. and K. Cool (1989), ‘Asset Stock Accumulation and Sustainability of Competitive
Advantage’, Management Science, Vol. 35, No. 12, 1504-1513.
Dobler, D.W. and D.N. Burt (1996), Purchasing and Supply Management, McGraw-Hill, New York, 6th
ed.
Donaldson, L. (2000), ‘Organizational Portfolio Theory: Performance-Driven Organizational

Change’, Contemporary Economic Policy, Vol. 18, No. 4, October, 386-396.
Doney, P.M. and J.P Cannon (1997), ‘An Examination of the Nature of Trust in Buyer-Seller
Relationships’, Journal of Marketing, Vol. 61, April, 35-51.
Dubinsky, A.J. and T.N. Ingram (1984), ‘A Portfolio Approach to Account Profitability’, Industrial
Marketing Management, Vol. 13, 33-41.
Dubois, A. and A.-C. Pedersen (2002), ‘Why Relationships do Not Fit into Purchasing Portfolio
Models – A Comparison Between the Portfolio and Industrial Network Approaches’, European
Journal of Purchasing & Supply Management, Vol. 8, No. 1, 35-42.
Dwyer, F.R., P.H. Schurr, and S. Oh (1987), ‘Developing Buyer-Seller Relationships’, Journal of
Marketing, Vol. 51, April, 11-27.
Dyer, J.H. and W.G. Ouchi (1993), ‘Japanese-Style Business Partnerships: Giving Companies a
Competitive Edge’, Sloan Management Review, Vol. 35, Fall, 51-63.
Dyer, J.H. (1996), ‘Specialized Supplier Networks as a Source of Competitive Advantage: Evidence
from the Auto Industry’, Strategic Management Journal, Vol. 17, April, 271-291.
Dyer, J.J. and H. Singh (1998), ‘The Relational View: Cooperative Strategy and Resources of
Interorganizational Competitive Advantage’, Academy of Management Review, Vol. 23, No. 4, 660-
679.
Dyer, J.H. and W. Chu (2000), ‘The Determinants of Trust in Supplier-automaker Relationships in
the U.S., Japan and Korea’, Journal of International Business Studies, Vol. 31, No. 2, Second Quarter,
259-285.
El-Ansary, A.I. and L.W. Stern (1972), ‘Power Measurement in the Distribution Channel’, Journal of
Marketing Research, Vol. 9, February, 47-52.
Elliott-Shircore, T.I. and P.T. Steele (1985), ‘Procurement Positioning Overview’, Purchasing and
Supply Management, December, 23/26.
Ellram, L.M. and A. Carr (1994), ‘Strategic Purchasing: A History and Review of the Literature’,
International Journal of Purchasing and Materials Management, Vol. 30, No. 2, Spring, 10-18.
Ellram, L.M. (1996), ‘A Structured Method for Applying Purchasing Cost Management Tools’,
International Journal of Purchasing and Materials Management, Vol. 32, No. 1, Winter, 20-28.
298
Ellram, L.M. (1998), ‘Purchasing’, in Lambert, D.M., J.R. Stock and L.M. Ellram (eds.), Fundamentals
of Logistics Management, Irwin Mc. Graw-Hill, 345-380.
Emerson, R.M. (1962), ‘Power-Dependence Relations’, American Sociological Review, Vol. 27,
February, 31-41.
Erdogan, B.Z. and M.J. Baker (2002), ‘Increasing Mail Survey Response Rates from an Industrial
Population’, Industrial Marketing Management, Vol. 31, No. 1, 65-73.
Essig, M. and U. Arnold (2001), ‘Electronic Procurement in Supply Chain Management: An

Information Economic-Based Analysis of Electronic Markets’, The Journal of Supply Chain
Management, Vol. 37, No. 4, Fall, 43-49.
Etgar, M. (1976), ‘Channel Domination and Countervailing Power in Distribution Channels’, Journal
of Marketing Research, Vol. 13, August, 254-262.
Faria, A. and R. Wensley (2002), ‘In Search of ‘Interfirm Management’ in Supply Chains:
Recognizing Contradictions of Language and Power by Listening’, Journal of Business Research, Vol.
35, 603-610.
Faulkner, T. (1996), ‘Applying “options thinking” to R&D Valuation’, Research Technology

Field, A. (2000), Discovering Statistics Using SPSS for Windows, SAGE Publications Ltd, London.
Fiocca, R. (1982), ‘Account Portfolio Analysis for Strategy Development’, Industrial Marketing
Management, Vol. 11, 53-62.
Ford, D. (ed.) (1997), Understanding Business Markets: Interaction, Relationships and Networks, The
Dryden Press, 2nd edition.
Forker, L.B. and P. Stannack (2000), ‘Cooperation versus Competition: Do Buyers and Suppliers
Really See Eye-To-Eye?’, European Journal of Purchasing & Supply Management, Vol. 6, No. 1, 31-40.
Foster, G. and M. Gupta (1994), ‘Marketing, Cost Management and Management Accounting’,
Journal of Management Accounting Research, Vol. 6, Fall, 43-54.
Foster, G., M. Gupta and L. Sjoblom (1996), ‘Customer Profitability Analysis: Challenges and New
Directions’, Cost Management, Vol. 10, Spring, 5-17.
Frazier, G.L., J.D. Gill and S.H. Kale (1989), ‘Dealer Dependence and Reciprocal Actions in a
Channel of Distribution in a Developing Country’, Journal of Marketing, Vol. 53, January, 50-69.
Frazier, G.L. and R.C. Rody (1991), ‘The Use of Influence Strategies in Interfirm Relationships in
Industrial Product Channels’, Journal of Marketing, Vol. 60, October, 19-38.
Frazier, G.L. and K.D. Antia (1995), ‘Exchange Relationships and Interfirm Power in Channels of
Distribution’, Journal of the Academy of Marketing Science, Vol. 23, Fall, 321-326.
Frederick, J. (1934), Industrial Marketing, Prentice Hal, New York.
Frohlich, M.T. and R. Westbrook (2001), ‘Arcs of Integration: An International Study of Supply
Chain Strategies’, Journal of Operations Management, Vol. 19, No. 2, 185-200.
Frohlich, M.T. (2002), ‘Techniques for Improving Response Rates in OM Survey Research’, Journal of
Operations Management, Vol. 20, No. 1, 53-62.
Gadde, L.E. and I. Snehota (2000), ‘Making the Most of Supplier Relationships’, Industrial Marketing
Ganesan, S. (1994), ‘Determinants of Long-Term Orientation in Buyer-Seller Relationships’, Journal

of Marketing, Vol. 58, No. 2, April, 1-19.
Garbarino, E. and M.S. Johnson (1999), ‘The Different Roles of Satisfaction, Trust, and Commitment
in Customer Relationships’, Journal of Marketing, Vol. 63, April, 70-87.
Gaski, J.F. (1984), ‘The Theory of Power and Conflict in Channel of Distribution’, Journal of
Marketing, Vol. 48, Summer, 9-29.
Gaski, J.F. (1986), ‘Interrelations Among a Channel Entity’s Power Sources: Impact of the Exercise of
Reward and Coercion on Expert, Referent, and Legitimate Power Sources’, Journal of Marketing
Research, Vol. 23, February, 62-77.
Gassenheimer, J.B. and R. Ramsey (1994), ‘The Impact of Dependence on Dealer Satisfaction: A
Comparison of Reseller-Supplier Relationships’, Journal of Retailing, Vol. 70, No. 3, 253-266.
Gelderman, C.J. and R.S.J. Tuninga (1998), ‘Paradigmawisseling in marketing?’, Bedrijfskunde, Vol.
70, No. 1, 89-97 (Dutch text).
Gelderman, C.J. and A.J. van Weele (2000), ‘New Perspectives on Kraljic’s Purchasing Portfolio
Approach’, Proceedings of the 9th International IPSERA Conference, London, Canada, 291-298.
Gelderman, C.J. (2000), ‘Rethinking Kraljic: Towards a Purchasing Portfolio Model, Based on
Mutual Buyer-Supplier Dependence’, Danish Purchasing & Logistics Forum (Dansk Indkøbs- og
Logistikforum), Vol. 37, No. 10, October, 9-15.
Gelderman, C.J. and A.J. van Weele (2001), ‘Advancements in the Use of a Purchasing Portfolio
Approach: A Case Study’, Proceedings of the 10th International IPSERA Conference, Jönköping,
Sweden, 403-415.
Gelderman, C.J. and L. de Boer (2001), ‘Stappenplan toepassing Kraljic inkoopportfolio analyse’,
Handboek Inkoopmanagement: Stappenplannen, checklisten en modellen, Samsom, Alphen aan den Rijn,
SE 2.1, 1-25 (Dutch text).
Gelderman, C.J. and W.F. van Oppen (2001), ‘DSM brengt dynamiek in de inkoopportfolio’,
Tijdschrift voor Inkoop & Logistiek, Vol. 17, No. 3, 31-34 resp. No. 4, 7 (Dutch text).
300
Gelderman, C.J. and A.J. van Weele (2002a), ‘Purchasing Portfolio Analysis: Towards a
Comprehensive Model of Conditions, Goals and Strategies Based on Explorative Case Studies’,
Proceedings of the 11th International IPSERA Conference, Enschede, The Netherlands, 212-226.
Gelderman, C.J. and A.J. van Weele (2002b), ‘Strategic Direction Through Purchasing Portfolio
Management: A Case Study’, International Journal of Supply Chain Management, Vol. 38, No. 2,
Spring, 30-37.
Gelderman, C.J. and P. van Haaster (2002), ‘De portfolio is veel werk en gezond verstand’,
interview met Peter Kraljic, Tijdschrift voor Inkoop & Logistiek, Vol. 18, No. 9, September (Dutch text).
Gelderman, C.J. (2002), ‘Handling the Main Issues and Initial Problems, Adopting a Purchasing
Portfolio Approach’, Danish Purchasing & Logistics Forum (Dansk Indkøbs- og Logistikforum), Vol. 39,
No. 12, December, 32-38.
Gelderman, C.J. and A.J. van Weele (2003), ‘Explaining the Utilisation of Purchasing Portfolio
Models in Business – Results of a Large Scale Survey’, Proceedings of the 12th International IPSERA
Conference, Budapest, Hungary.
Gellner, D.M. (1978), ‘Involvement in Role-Playing Simulations: A Demonstration with Studies on

Obedience’, Journal of Personality and Social Psychology, Vol. 36, 219-235.
Geyskens, I., J.E.M. Steenkamp, L.K. Scheer and N. Kumar (1996), ‘The Effects of Trust and
Interdependence on Relationship Commitment: A Trans-Atlantic Study’, International Journal of
Research in Marketing, Vol. 13, 303-317.
Granovetter, M. (1985), ‘Economic Action and Social Structure: A Theory of Embeddedness’,

American Journal of Sociology, Vol. 91, 481-510.
Graves, S.B., J.L. Ringuest and R. H Case (2000), ‘Formulating Optimal R&D Portfolios’, Research
Technology Management, Vol. 43, No. 3, 47-51.
Greenwald, A.G., A.R. Pratknis, M.R. Leippe and M.H. Baumgardner (1986), ‘Under What
Conditions Does Theory Obstruct Research Progress?’, Psychological Review, Vol. 93. , No. 2, 216-229.
Greer, T.V., N. Chuchinprakarn, and S. Seshadri (2000), ‘Likehood of Participating in Mail Survey
Research - Business Respondents’ Perspectives’, Industrial Marketing Management, Vol. 29, No. 2, 97-
109.
Grönroos, C. (1990), Service Management and Marketing: Managing the Moments of Truth in Service
Competition, Lexington Books, Massachusetts/Toronto.
Grönroos, Chr. (1994), ‘Quo Vadis Marketing? Toward a Relationship Marketing Paradigm’, Journal
of Marketing Management, Vol. 10, No. 4, 347-360.
Gummesson, E. (1987), ‘The New Marketing - Developing Long-term Interactive Relationships’,

Long Range Planning, Vol. 20, No. 4, August, 10-20.
Gundlach, G.T. and E.R. Cadotte (1994), ‘Exchange Interdependence and Interfirm Reaction:
Research in a Simulated Channel Setting’, Journal of Marketing Research, Vol. 31, 516-532.
Gundlach, G.T., R.S. Achrol and J.T. Mentzer (1995), ‘The Structure of Commitment in Exchange’,
Journal of Marketing, Vol. 59, January, 78-92.
Hadeler, B.J. and J.R. Evans (1994), ‘Supply Strategy: Capturing the Value’, Industrial Management,
Vol. 36, No. 4, July/August, 3-4.
Hahn, C.K., C.A. Watts, and K.Y. Kim (1990), ‘The Supplier Development Program: A Conceptual
Model’, International Journal of Purchasing and Materials Management, Vol. 26, No. 2, Spring, 2-7.
Håkansson, H. (ed.) (1982), International Marketing and Purchasing of Industrial Goods: an Interaction
Approach, Wiley, Chichester.
Håkansson, H, (1989), Corporate Technological Behaviour: Co-operation and Networks, Routledge,

London.
Håkansson, H. and A.K. Erikkson (1993), ‘Getting Innovations out of Supplier Networks’, Journal of
Business-to-Business Marketing, Vol. 1, No. 3, 30-37.
Håkansson, H. and I. Snehota (1995), Developing Business Relationships in Industrial Networks,

Routledge, London.
Hall, D.L. and A. Nauda (1990), ‘An Interactive Approach for Selecting IR&D Projects’, IEEE
Transactions on Engineering Management, Vol. 37, No. 2, 126-135.
Hallén, L., J. Johanson and N. Seyed-Mohamed (1991), ‘Interfirm Adaptions in Business
Relationships’, Journal of Marketing, Vol. 55, April, 29-37.
Hambrick, D.C., I.C. MacMillan and D.L. Day (1982), ‘Strategic attributes and performances in the
BCG matrix - a PIMS based analysis of industrial product businesses’, Academy of Management
Journal, Vol. 25, No. 3, 510-531.
Hamermesh, R.G. (1986), ‘Making Planning Strategic’, Harvard Business Review, Vol. 64,
July/August, 115-120.
Handfield, R.B. (1993), ‘A Resource dependence perspective of Just-In-Time purchasing’, Journal of

Handfield, R.B, D. Krause, Th. V. Scannell and R. Monczka (2000), ‘Avoid the Pitfalls in Supplier
Development’, Sloan Management Review, Vol. 41, No. 2, 37-49.
Handfield, R.B. and C. Bechtel (2002), ‘The Role of Trust and Relationship Structure in Improving
Supply Chain Responsiveness’, Industrial Marketing Management, Vol. 31, No. 4, 367-382.
Harland, C.J. (1996), ‘Supply Chain Management: Relationships, Chain and Networks’, British
Journal of Management, Vol. 7, Special Issue, March, S63-S80.
302
Harryson, S. (1995), Japanse R&D Management: A Holistic Network Approach, doctoral thesis,
University of St. Gallen.
Hartmann, E., Th. Ritter and H.G. Gemünden (2001), ‘Determining the Purchase Situation:
Cornerstone of Supplier Relationship Management’, http://web.bi.no/imp2001, The 17th Annual IMP
Conference, Oslo, Norway, 1-29,
Haspeslagh, P. (1982), ‘Portfolio Planning: Uses and Limits’, Harvard Business Review, Vol. 60, No. 1,
January/February, 58-73.
Heege, F. (1981), ‘Portfolio-Management in der Beschaffung’, Der Betriebswirtschaft, February,

Sonderdruck, 17-23 (German text).
Heide, J.B. and G. John (1988), ‘The Role of Dependence Balancing in Safeguarding Transaction-
Specific Assets in Conventional Channels’, Journal of Marketing, Vol. 52, No. 1, January, 20-35.
Heide, J.B. and G. John (1990), ‘Alliances in Industrial Purchasing: The Determinants of Joint Action
in Buyer-Seller Relationships’, Journal of Marketing Research, Vol. 27, February, 24-36.
Heide, J.B. and G. John (1992), ‘Do Norms Matter in Marketing Relationships?’, Journal of Marketing,
Vol. 56, April, 32-44.
Heide, J.B. and A.S. Minor (1992), ‘The Shadow of the Future: Effects of Anticipated Interaction and
Frequency of Contract on Buyer-Seller Cooperation’, Academy of Management Journal, Vol. 35,
No. 2, 265-291.
Heide, J.B. (1994), ‘Interorganizational Governance in Marketing Channels’, Journal of Marketing,

Vol. 58, No. 1, 71-85.
Helper, S. and M. Sako (1995), ‘Supplier Relations in Japan and the United States: Are They
Converging?’, Sloan Management Review, Vol. 35, Spring, 77-84.
Henderson, B.D. (1970), ‘The Product Portfolio’, Boston Consulting Group (Perspectives), Boston.
Henderson, B.D. (1972), ‘Cash Traps’, Boston Consulting Group (Perspectives), Boston.
Henderson, B.D. (1973), ‘The Experience Curve Reviewed: IV. The Growth Share Matrix of the
Product Portfolio, Perspectives’, No. 135, The Boston Consulting Group, Boston, M.A.
Hines, P., R. Lamming, D. Jones, P. Cousins and N. Rich (2000), Value Stream Management, Prentice-
Hall, London, UK.
Hofer, C.W. and D. Schendel (1978), ‘General Electric’s Nine-Cell Strategic Business Screen’,
Strategy Formulation and Analytical Concepts.
Homburg, Chr. (1995), ‘Single sourcing, double sourcing, multiple sourcing…? Ein ökonomischer
Erklärungsansatz’, Zeitschrift für Betriebswirtschaft, Vol. 65, No. 8, 813-834 (German text).
Homburg, Chr. and D. Daum (1997), ‘Die Kundenstruktur als Controlling Herausforderung’,
Controlling, Heft 6, November/Dezember, 394-404 (German text).
Hooley, G.J. and J. Saunders (1993), Competitive positioning: the key to marketing strategy, Prentice Hall
International (UK) Ltd.
C:\Documents and Settings\gel\Local Settings\Temporary Internet Files\OLK385\Multiple

Comparisons with Repeated Measures.htm
Howell, D.C. (2002), Statistical Methods for Psychology, Belmont, CA: Duxbury Press and
http://www.uvm.edu/~dhowell/StatPages/More_Stuff/RepMeasMultComp/RepMeasMultComp.htm
l#Return3
Hughes, M. (1981), ‘Portfolio Analysis’, Longe Range Planning, Vol. 14, February, 101-103.
Hunt, S.D. and R.M. Morgan (1996), ‘The Resource-Advantage Theory of Competition: Dynamics,
Path Dependencies, and Evolutionary Dimensions’, Journal of Marketing, Vol. 60, October, 107-114.
Jackson, B.B. (1985), Winning and Keeping Industrial Customers, Lexington Books, Lexington MA.
Jacobs, J. (1974), ‘Dependency and Vulnerability: An Exchange Approach to the Control of

Organizations, Administrative Sciences Quarterly, Vol. 19, No. 1, 45-59.
Jacobson, R. and D. Aaker (1985), ‘Is market share all that it’s cracked up to be?’, Journal of
Marketing, Vol. 49, Fall, 11-21.
Jacobson, R. (1988), ‘Distinguishing Among Competing Theories of the Market Share Effect’, Journal
of Marketing, Vol. 52, October, 68-80.
Jap, S.D. and S. Ganesan (2000), ‘Control Mechanisms and the Relationship Life Cycle: Implications
for Safeguarding Specific Investments and Developing Commitment’, Journal of Marketing Research,
Vol. 37, No. 2, May, 227-245.
Jarillo, J.C. (1988), ‘On Strategic Networks’, Strategic Management Journal, Vol. 9, No. 1, 31-41.
Jobber, D. and D. O’Reilly (1996), ‘Industrial Mail Surveys: Techniques for Inducing Response’,
Marketing & Intelligence Planning, Vol. 14, No. 1, 29-34.
Johansson, U.E. (1997), ‘Decision Making in Inter-Firm Networks as a Political Process’, Organization
Studies, Vol. 18, No. 3, 361-384.
John, G. (1984), ‘An Empirical Investigation of Some Antecedents of Opportunism in a Marketing

Channel’, Journal of Marketing Research, Vol. 21, August, 278-289.
Johnsen, T. (2000), ‘How Can Companies Manage Supply Networks?’, Danish Purchasing & Logistics
Forum (Dansk Indkøbs- og Logistikforum), Vol. 37, December.
Johnsen, R.E. and D. Ford (2001), ‘Asymmetrical and Symmetrical Customer-Supplier Relationships:
Contrasts, Evolution and Strategy’, http://web.bi.no/imp2001, The 17th Annual IMP Conference, Oslo,
Norway, 1-20,
304
Johnson, J.L. (1999), ‘Strategic Integration in Industrial Distribution Channels: Managing the
Interfirm Relationship as a Strategic Asset’, Journal of the Academy of Marketing Science, Vol. 27, No.1,
4-18.
Joshi, A.W. and S.J. Arnold (1997), ‘The Impact of Buyer Dependence on Buyer Opportunism in
Buyer-Supplier Relationships: The Moderating Role of Relational Norms’, Psychology & Marketing,
Vol. 14, No. 8, December, 823-845.
Kamann, D.J.F. (2000a), ‘Kraljic krijgt extra dimensie’, Tijdschrift voor Inkoop & Logistiek, Vol. 16,
April, No. 4, 8-12 (Dutch text).
Kamann, D.J.F. (2000b), ‘Matrices, Cubes and Triangles in Purchasing’, Poster Presentation at the the
9th International IPSERA Conference, London, Canada, 1-6.
Kamath, R.R. and J.K. Liker (1990), ‘Supplier dependence and innovation: a contingency model of
suppliers’ innovative activities’, Journal of Engineering and Technology Management, Vol. 7, No. 2, 111-
127.
Keep, W.W., S.C. Hollander and R. Dickinson (1998), ‘Forces Impinging on Long-Term Business-to-
Business Relationships in the United States: An Historical Perspective’, Journal of Marketing, Vol. 62,
April, 31-45.
Kempeners, M. and A.J. van Weele (1997), ‘Inkoopportfolio: basis voor inkoop- en
marketingstrategie’, in: Van der Hart, H.W.C. and A.J. van Weele (eds), Dynamiek in commerciële
relaties, F&G Publishing, Bunnik (Dutch text).
Kirchhoff, B.A., M.J. Merges and J. Morabito (2001), ‘A Value Creation Model for Measuring and
Managing the R&D Portfolio’, Engineering Management Journal, Vol. 13, No. 1, 19-22.
Kim, K. (2001), ‘On the Effects of Customer Conditions on Distributor Commitment and Supplier
Commitment in Industrial Channels of Distribution’, Journal of Business Research, Vol. 52, No. 2, 87-
99.
Knudsen, D. (2002), ‘Uncovering the Strategic Domain of E-Procurement’, Proceedings of the 11th
International IPSERA Conference, Enschede, The Netherlands, 397-407.
Kotler, Ph. (1971), Marketing Decision Making: A Model Building Approach, Holt, Rinehart and
Winston, Inc., New York.
Kotler, Ph. and G. Armstrong (1991), Principles of marketing, Prentice Hall, Englewood Cliffs, N.J., 5th
ed.
Kraljic, P. (1977), ‘Neue Wege im Beschaffungsmarketing’, Beschaffung Aktuell, December, 20-26.
Kraljic, P. (1983), ‘Purchasing Must Become Supply Management’, Harvard Business Review, Vol. 61,
No. 5, September-October, 109-117.
Krapfel, R.E., D. Salmond and R.E. Spekman (1991), ‘A Strategic Approach to Managing Buyer-
Seller Relationships’, European Journal of Marketing, Vol. 25, No. 9, 22-37.
Krause, D.R. and L.M. Ellram (1997), ‘Critical Elements of Suppleir Development: The Buying Firm
Perspective’, European Journal of Purchasing and Supply Management, Vol. 3, No. 1, 21-31.
Kumar, N., J.D. Hibbard and L.W. Stern (1994), The Nature and Consequences of Marketing
Channel Intermediary Commitment, Marketing Science Institute, No. 94-115, Cambridge MA.
Kumar, N., L.K. Sheer and J.E.M. Steenkamp (1995), ‘The Effects of Perceived Interdependence on
Dealer Attitudes’, Journal of Marketing Research, Vol. 32, August, 348-356.
Kumar, N. (1996), ‘The Power of Trust in Manufacturer Relationships’, Harvard Business Review, Vol.
74, November-December, 93-106.
Lado, A.A., N.G. Boyd and S.C. Hanlon (1997), ‘Competition, Cooperation, and the Search for
Economic Rents: A Syncretic Model’, Academy of Management, Vol. 22, No. 1, 110-141.
Lambert, D.M., M.A. Emmelhainz, and J.T. Gardner (1996), ‘Developing and Implementing Supply
Chain Partnerships’, The International Journal of Logistics Management, Vol. 7, No. 2, 1-17.
Lambert, D.M. and M.C. Cooper (2000), ‘Issues in Supply Chain Management’, Industrial Marketing
Management, Vol. 29, 65-83.
Lamming, R.C. (1993), Beyond Partnership, Prentice-Hall, UK.
Lamming, R.C., Johnsen, T., C.M. Harland and J. Zheng (2000), ‘Managing in Supply Networks:
Cascade and Intervention’, Proceedings of the 9th International IPSERA Conference, London, Canada,
393-404.
Lamming, R.C. and D. Harrison (2001), ‘Smaller Customers and Larger Suppliers: The Potential for
Strategic Purchasing Approach: A Case Study’, Proceedings of the 10th International IPSERA
Conference, Jönköping, Sweden, 595-610.
Laseter, T. (1998), Balanced Sourcing: Cooperation and Competition in Supplier Relationships, Jossey-Bass
Publishers, San Francisco.
Lazzarini, S.G., F.R. Chaddad and M.L. Cook (2001), ‘Integrating Supply Chain and Network
Analyses: The Study of Netchains’, Chain and Network Science, Vol. 1, No. 1, 7-22.
Leenders, M.R. and H.E. Fearon (1993), Purchasing and Materials Management, R.R. Irwin Inc.,
Homewood Illinois, 10th ed.
Leonard, A. and M. Spring (2002), ‘ “E-Procurement Must Become Supply Management”: Portfolio
Approaches and Electronic Procurement’, Proceedings of the 11th International IPSERA Conference,
Enschede, The Netherlands, 456-470.
Lilliecreutz, J. and L. Ydreskog (1999), ‘Supplier Classification as an Enabler for a Differentiated

Purchasing Strategy’, Global Purchasing & Supply Chain Management, November, 66-74.
Lockridge, R.K. (1981), ‘Strategy in the 1980’s’, BCG Perspectives, 241.

306
Lorenzoni, G. and A. Lipparini (1999), ‘The Leveraging of Interfirm Relationships as a Distinctive

Organizational Capability: a Longitudinal Study, Strategic Management Journal, Vol. 20, 317-338.
Louwers, D., W. van Oppen and G. Walravens (1999), ‘DSM werkt succesvol met multidisciplinaire
inkoopteams’, Tijdschrift voor Inkoop & Logistiek, Vol. 15, January-February, 8-12 (Dutch text)..
Lusch, R. and J.R. Brown (1996), ‘Interdependency, Contracting, and Relational Behavior in
Marketing Channels’, Journal of Marketing, Vol. 60, October, 19-38.
MacBeth, D.K. and N. Ferguson (1994), Partnership Sourcing – An Integrated Supply Chain
Management Approach, Longman Group UK.
Macneil, I.R. (1980), The New Social Contract: An Inquiry into Modern Contractual Relations, New
Haven: Yale University Press.
Maltz, A. and L. Ellram (1999), ‘Outsourcing Supply Management’, The Journal of Supply Chain
Management, Vol 35, No. 2, 4-17.
Mangione, Th. W. (1995), Mail Survey. Improving the Quality, Thousand Oaks, Sage.
Markowitz, H. (1952), ‘Portfolio Selection’, Journal of Finance, 7, March, 77-91.
Marshall, A. (1890), Principles of Economics, Macmillan, London.
Marvin, Ph. (1972), Product Planning Simplified, American Management Association, New York.
Matheson, D., J.E. Matheson and M.M. Menke (1994), ‘Making Excellent R&D Decisions’, Research
Technology Management, Vol. 37, No. 6, November-December, 21-24.
Matthyssens, P. and C. van den Bulte (1994), ‘Getting Closer and Nicer: Partnerships in the Supply
Chain’, Longe Range Planning, Vol. 27, No. 1, 72-83.
Mattieu, J.E. and D.M. Zajac (1990), ‘A Review and Meta-Analysis of the Antecedents, Correlates
and Consequences of Organizational Commitment’, Psycological Bulletin, Vol. 108, No. 2, 171-194.
McAllister, D.I. (1995), ‘Affect-and Cognition-Based Trust as Foundations for Interpersonal

Cooperation in Organizations’, Academy of Management Journal, Vol. 38, February, 24-59.
McCutcheon and F.I. Stuart (2000), ‘Issues in the Choice of Supplier Alliance Partners’, Journal of
McDonald, M.H.B. (1990), ‘Some Methodological Comments on The Directional Policy Matrix’,
Journal of Marketing Management, Vol. 6, No. 1, 59-68.
McIvor, R. and M. McHugh (2000), ‘Partnership Sourcing: An Organization Change Management

Perspective’, The Journal of Supply Chain Management, Vol. 36, No. 3, Summer, 12-20.
McKiernan, P. (1997), ‘Strategy Past; Strategy Futures’, Long Range Planning, Vol. 30, No. 5, 790-798.
McLagan, P.A. (2000), ‘Portfolio Thinking’, Training & Development, Vol. 54, No. 2, 44-51.
McLoughlin, D. and C. Horan (2000), ‘Business Marketing: Perspectives from the Markets-as-
Networks Approach’, Industrial Marketing Management, Vol. 29, 285-292.
Mickwitz, G. (1959), Marketing and Competition, Helsingfors, Finland: Centraltrykeriet.
Miles, G., S.B. Preece and M.C. Baetz (1999), ‘Dangers of Dependence: The Impact of Strategic
Alliance Use by Small Technology-Based Firms’, Journal of Small Business Management, Vol. 37, No.
2, April, 20-29.
Miles, M.B. and A.M. Huberman (1994), Qualitative Data Analysis, Sage Publications, London, 2nd ed.
Mintzberg, H. (1983), Power in and around organizations, Prentice Hall, Englewood Cliffs, N.J.
Mol, M.J., J. Schreuder and R. Goedegebuure (2001), ‘Het Belang van Inkoop in de Nederlandse
Industrie en de Veranderende Rol van de Inkoper’, Tijdschrift voor Inkoop & Logistiek, Vol. 17, No. 11,
November, 69-72 (Dutch text).
Mol, M.J. (2002), Outsourcing, Supplier Relations and Internationalisation: Global Sourcing Strategy as a
Chinese Puzzle, Doctoral thesis, Erasmus Research Institute of Management, Erasmus University
Rotterdam.
Møller, M. and J. Momme (2000), ‘Supplier segmentation in theory and practice – towards a
competence perspective’, Proceedings of the 9th International IPSERA Conference, London, Canada,
508-519.
Monckza, R.M., R. Trent and R. Handfield (1998), Purchasing and supply chain management,
International Thomsom Publishing, Cincinnati, Ohio.
Moorman, C., G. Zaltman and R. Dehpandé (1992), ‘Relationships Between Providers and Users of
Market Research: The Dynamics of Trust Within and Between Organizations’, Journal of Marketing
Research, Vol. 29, August, 314-328.
Moorman, C., R. Dehpandé and G. Zaltman (1993), ‘Factors Affecting Trust in Market Research
Relationships’, Journal of Marketing, Vol. 57, January, 81-101.
Morgan R.M. and S.D. Hunt (1994), ‘The Commitment-Trust Theory of Relationship Marketing’,
Journal of Marketing, Vol. 58, July, blz. 20-38.
Morgan, R.D. and E.W. Garnsey (1994), ‘The Integration of Suppliers inot the Product Development
Process’, R&D Management, Vol. 24, No. 2, 28-35.
Morrison, A. and R. Wensley (1991), ‘Boxing up or boxed in? A short history of the Boston
Consulting Group share/growth matrix’, Journal of Marketing Management, 7, 105-129.
Neef, D. (2001), E-Procurement – From Strategy to Implementation, Prentice Hall PTR, Upper Saddle
River, NJ.
308
Nellore, R. and K. Söderquist (2000), ‘Portfolio Approaches to Procurement – Analysing the Missing
Link to Specifications’, Long Range Planning, Vol. 33, No. 2, 245-267.
Nellore, R. and J.E. Taylor (2000), ‘Using Portfolio Approaches to Manage Engineering-Purchasing-
Supplier Interaction’, Production and Inventory Management, Vol. 41, No. 1, First Quarter, 6-12.
Nicholson, A. (1993), ‘Strategic Management for the Purchasing Professional (Part 2)’, Purchasing &
Supply Management, Vol. 27, October, 32-36.
Neuman, W.L. (1997), Social Research Methods, Allyn and Bacon, London, 2nd ed.
Noordewier, T.G., G. John and J.R. Nevin (1990), ‘Performance Outcomes of Purchasing
Arrangements in Industrial Buyer-Vendor Relationships’, Journal of Marketing, Vol. 54, October, 80-
93.
Nooteboom, B. (1996), ‘Trust, Opportunism and Governance: A Process and Control Model’,
Organization Studies, Vol. 17, No. 6, 985-1010.
Nooteboom, B., H. Berger and N.G. Noorderhaven (1997), ‘Effects of Trust and Governance on
Relational Risk’, Academy of Management Journal, Vol. 40, No. 2, 308-338.
Nooteboom, B., G. de Jong, R.W. Vossen, S. Helper and M. Sako (2000), ‘Network Interactions and
Mutual Dependence: A Test in the Car Industry’, Industry and Innovation, Vol. 7, No. 1, June,
117-144.
Oliver, A.L. and M. Ebers (1998), ‘Networking Network Studies: An Analysis of Conceptual
Configurations in the Study of Inter-organizational Relationships’, Organizational Studies, Vol. 19,
No. 4, 549-583.
Olsen, R.F. and L.M. Ellram (1997), ‘A Portfolio Approach to Supplier Relationships’, Industrial
Marketing Management, Vol. 26, No. 2, 101-113.
Ouchi, W.G. (1984), The M-Form Society, Avon Books, New York.
Parker, D. and K. Hartley (1997), ‘The Economics of Partnership Sourcing versus Adversarial
Competition: A Critique’, European Journal of Purchasing & Supply Management, Vol. 3,
No. 2, 115-125.
Partnership Sourcing, (1991), Partnership Sourcing Ltd, London.
Patel, P. and M. Younger (1978), ‘A Frame of Reference for Strategy Development’, Long Range
Planning, Vol. 11, April, 6-12.
Peppers, D. and M. Rogers (2001), Customer Development Strategies for the Business-to-Business World,
Doubleday – Random House Inc., New York.
Pfeffer, J. and G.R. Salancik (1978), The External Control of Organizations - A Resource Dependence
Perspective, Harper & Row Publishers, New York.
Pfeffer, J. (1981), Power in Organizations, Pitman Publishing Inc., Massachusetts.
Pine II, B.J. (1993), Mass Customization: The New Frontier in Business Competition, Harvard Business
School Press, Boston.
Polli, R. and V.J. Cook (1969), ‘Validity of the Product Life Cycle’, The Journal of Business, Vol. 42,
October, 385-400.
Pooley, J. and S.C. Dunn (1994), ‘A Longitudinal Study of Purchasing Positions: 1960-1989’, Journal
of Business Logistics, Vol. 15, No. 1, 193-214.
Porter, M.E. (1980), Competitive Strategy, The Free Press, New York.
Prahalad, C.K. and F. Hamel (1990), ‘The Core Competence of the Corporation’, Harvard Business
Review, May/June.
Provan, K.G. and J.B. Gassenheimer (1994), ‘Supplier Commitment in Relational Contract
Exchanges with Buyers: A Study of Interorganizational Dependence and Exercised Power’, Journal
of Management Studies, Vol. 31, No. 1, 55-68.
Provan, K.G. and S.J Skinner (1989), ‘Interorganizational dependence and control as predictors of
opportunism in dealer-supplier relations’, Academy of Management Journal, Vol. 32, No. 1, 202-212.
Quinn, J.B. and F.G. Hilmer (1994), ‘Strategic Outsourcing’, Sloan Management Review, Summer,
43-55.
Rangan, V.K. and G.T. Bowman (1995), ‘Beating the Commodity Magnet’, in: Rangan, V.K., B.P.
Shapiro and R.T. Moriarty (ed), Business Marketing Strategy: Concepts and Applications, Irwin Inc.,
Chicago.
Ramsay, J. (1996a), ‘The Case Against Purchasing Partnerships’, International Journal of Purchasing
and Materials Management, Vol. 32, No. 4, Fall, 13-19.
Ramsay, J. (1996b), ‘Power Measurement’, European Journal of Purchasing and Supply Management,
Vol. 2, No. 2/3, 129-143.
Ramsay, J. (2001), ‘Purchasing’s Strategic Irrelevance’, European Journal of Purchasing and Supply
Remenyi, D., B. Williams, A. Money and E. Swartz (1998), Doing Research in Business and
Management – An Introduction to Process and Method, Sage Publications, London.
Reilly, F.K and E.A. Norton (1999), Investments, The Dryden Press, Orlando, 5th ed.
Ries, A. and J. Trout (1981), Positioning: the battle for your mind, McGraw-Hill, London.
Rietveld, G.J. (1998), ‘De opmars van inkoop’, Internet publication, http://www.compendium.nl,
(Dutch text).
310
Rozemeijer, F.A. (2000), Creating Corporate Advantage in Purchasing, Doctoral thesis, Eindhoven
Centre for Innovation Studies, Eindhoven University of Technology.
Ruyter, J.C. de, L. Moorman and J.G.A.M. Lemmink (2001), ‘Antecedents of Commitment and Trust
in Customer-Supplier Relationships in High Technology Markets’, Industrial Marketing Management,
Vol. 30, No. 3, 271-286.
Sako, M. (1992), Prices, Quality and Trust – Inter-firm Relationships In Britain and Japan, Cambridge
UK: Cambridge University Press.
Saunders, M.(1997), Strategic Purchasing and Supply Chain Management, Pitman Publishing, London,
2nd ed.
Saunders, M.N.K., Ph. Lewis and A. Thornhill (2000), Research Methods for Business Students, Pearson
Educational Limited, Harlow, 2nd ed.
Schary, Ph.B. and T. Skjøtt-Larsen, (2001), Managing the Global Supply Chain, Copenhagen Business
School Press, Denmark, 2nd ed.
Schijns, J.M.C. (1998), Het meten en managen van klant-organisatie relaties, dissertation, the University
of Maastricht (Dutch text).
Scott, C. and R. Westbrook (1991), ‘New Strategic Tools for Supply Chain Management’,
International Journal of Physical Distribution, Vol. 21, No. 1, 23-33.
Shapiro, B.P. V.K. Rangan, R.T. Moriarty and E.B. Ross (1987), ‘Manage Customers for Profits (not
just sales)’, Harvard Business Review, September/October, 101-108.
Sharpe, W.F. (1963), ‘A Simplified Model for Portfolio Analysis’, Management Science, January,
277-293.
Slade, B.N. (1993), Compressing the Product Development Cycle from Research to Marketplace, American
Management Association, New York.
Smeltzer, L.R. and S.P. Siferd (1998), ‘Proactive Supply Management: The Management of Risk’,
International Journal of Purchasing and Materials Management, Vol. 34, No. 1, Winter, 38-45.
Smith, J.B. and D.W. Barclay (1997), ‘The Effects of Organizational Differences and Trust on the
Effectiveness of Selling Partner Relationships’, Journal of Marketing, Vol. 61, January, 3-21.
Spekman, R.E. and D. Strauss (1986), ‘An Exploratory Investigation of a Buyer’s Concern for Factors
Affecting more Co-operative Buyer-Seller Relationships’, Industrial Marketing and Purchasing, Vol. 1,
No. 3, 26-43.
Sriram, V., R. Krapfel and R. Spekman (1992), ‘Antecedents to Buyer-Seller Collaboration: An

Analysis From the Buyers’ Perspective’, Journal of Business Research, Vol. 25, No. 4, 303-320.
Stannack, P. (1996), ‘Purchasing Power and Supply Chain Management Power - Two Different
Paradigms? - A Response to Ramsay’s ‘Purchasing power’ (1995)’, European Journal of Purchasing and
Supply Management, Vol. 2, No. 1, 47-56.
Steele, P.T. and B.H. Court (1996), Profitable Purchasing Strategies: A Manager’s Guide for Improving
Organizational Competitiveness through the Skills of Purchasing, McGraw-Hill Book Company, London.
Stevens, J.P. (2001), Applied Multivariate Statistics for the Social Sciences, Lawrence, Erlbaum
Associates, Inc. Mahwah, NJ, 4th ed.
Stinchcombe, A.L. (1985), ‘Contracts as Hierarchical Documents’, in A.L. Stinchcombe and C.

Heimer (eds.), Organizational Theory and Project Management, Norwegian University Press, Oslo,
121-171.
Storbacka, K. (1994), ‘Customer Relationship Profitability - Propositions for Analysis of

Relationships and Customer Bases’, in: Sheth, J.N. and A. Parvatiyar (ed.), Relationship Marketing:
Theory, Methods and Applications, 1994 Research conference proceedings, Emory University, Atlanta.
Strandvik, T. and V. Liljander (1994), ‘Relationship Strength in Bank Services’, in: Sheth, J.N. and A.
Parvatiyar (ed.), Relationship Marketing: Theory, Methods and Applications, 1994 Research conference
proceedings, Emory University, Atlanta.
Swift, C.O. (1995), ‘Preference for Single Sourcing and Supplier Selection Criteria’, Journal of
Business Research, Vol. 32, No. 2, 105-111.
Syson, R. (1992), Improve Purchase Performance, Pitman Publishing, London.
Szymanski, D.M., S.G. Bharadwaj and P.R. Varadarajan (1993),’An Analysis of the Market Share-
Profitability Relationship, Journal of Marketing, Vol. 57, July, 1-18.
Tabacknick, B.G. and L.S. Fidell (2001), Using Multivariate Statistics, Allyn & Bacon, Needham
Heights, MA, 4th ed.
Thibaut, J.W. and H.H. Kelley (1959), The Social Psychology of Groups, John Wiley & Sons, New York.
Thompson, J.D., (1967), Organizations in Action, McGraw -Hill, New York.
Thorelli, H.B. (1986), ‘Networks: Between Markets and Hierarchies’, Strategic Management Journal,
Vol. 7, 37-51.
Treacy, M. and F. Wiersema (1993), ‘Customer Intimacy and Other Value Disciplines’, Harvard
Business Review, Vol. 71, January-February, 499-504.
Trent, R.J. and R.M. Monczka (1998), ‘Purchasing and Supply Management: Trends and Changes
Throughout the 1990s’, International Journal of Purchasing and Materials Management, Vol. 34, No. 4,
Fall, 2-11.
Turnbull, P.W. (1990), ‘A Review of Portfolio Planning Models for Industrial Marketing and
Purchasing Management’, European Journal of Marketing, Vol. 24, 7-22.
312
Turnbull, P.W. and J. Valla (ed.) (1986), Strategies for International and Industrial Marketing: The
Management of Customer Relationships in European Industrial Markets, Croom Helm, London.
Turnbull, P.W. and S. Topcu (1994), ‘Customer’ Profitability in Relationship Life Cycles’, Proceedings
of the 10th IMP Conference, Groningen, The Netherlands.
Turnbull, P.W. and J. Zolkiewski (1997), ‘Profitability in Customer Portfolio Planning’, in: Ford, D.
(ed), Understanding Business Markets: Interactions, Relationships and Networks, The Dryden Press,
London, 2nd edition.
Tully, S. (1995), ‘Purchasing’s New Muscle’, Fortune, Vol. 65, February 20, 45-51.
Tuten, T.L. and D.J. Urban (2001), ‘An Expanded Model of Business-to-Business Partnership
Formation and Success’, Industrial Marketing Management, Vol. 30, 149-164.
Ulrich, D. and J.B. Barney (1984), ‘Perspectives in Organizations: Resource Dependence, Efficiency,
and Population’, Academy of Management Review, Vol. 9, No. 3, 471-481.
Vammen, S. (2000), ‘Purchasing Should be Blamed for its Low Status and Invisibility’, Danish
Purchasing & Logistics Forum (Dansk Indkøbs- og Logistikforum), Vol. 37, No. 7, July, 22-27.
Van der Hart, H.W.C. and A.J. van Weele (ed.) (1997), Dynamiek in commerciële relaties. F&G
Publishing, Bunnik (Dutch text).
Van de Ven, A. (1976), ‘On the Nature, Formation, and Maintenance of Relations among
Organizations, Administrative Science Quarterly, Vol. 21, December, 598-621.
Van Weele, A.J. (1984), ‘Purchasing Performance Measurement and Evaluation’, Journal of
Purchasing and Materials Management, Vol. 20, No. 3, Fall, 16-24.
Van Weele, A.J. (1988, 1992, 1997, 2001), Inkoop in strategisch perspectief. Analyse, planning en praktijk,
Samsom Bedrijfsinformatie, Alphen aan den Rijn, 1st, 2nd 3rd and 4th ed (Dutch text).
Van Weele, A.J. and F.A. Rozemeijer (1996), Revolution in purchasing: building competitive power
through pro-active purchasing, Eindhoven University of Technology, Eindhoven.
Van Weele, A.J. and F.A. Rozemeijer and G. Rietveld (1998), ‘Professionalising Purchasing in
Organizations: Toward a Purchasing Development Model’, Holland Consulting Group, Amsterdam,
The Netherlands.
Van Weele, A.J. (2000), Purchasing Management: Analysis, Planning and Practice, Chapman & Hall,
London, 2nd edition.
Van Weele, A.J. (2002), ‘Measuring Purchasing Performance’, Presentation at the Nordisk Konference
‘Strategic Challenges for Leading Edge Purchasing Executives’, October 2, Copenhagen, Denmark.
Van Weele, A.J. and F.A. Rozemeijer (1999), Ondernemend samenwerken: ontwikkelen van
concurrentiekracht in netwerken’, Samsom, Alphen aan den Rijn, The Netherlands (Dutch text).
Venkatesan, R (1992), ‘Strategic Sourcing: To Make or Not To Make’, Harvard Business Review,
November/December, blz. 98-107.
Venkatraman, N. and V. Ramanujam (1986), ‘Measurement of Business Performance in Strategy

Research: A Comparison of Approaches’, Academy of Management Review, Vol. 11, No. 4, 901-813.
Verhage, B.J. and E. Waarts (1988), ‘Marketing Planning for Improved Performance: A Comparative
Analysis’, International Marketing Review, Vol. 5, 20-30.
Verheul, Th. A.M. and S.C. Santema (1998), ‘Relatie tussen inkoper en verkoper gedoemd tot
mislukken?’, Tijdschrift voor Inkoop & Logistiek, Vol. 14, No. 9, 60-64 (Dutch text).
Webster, F.E., Jr. (1992), "The Changing Role of Marketing in the Corporation", Journal of Marketing,
Vol. 56, October, blz. 1-17.
Webster’s Third New International Dictionary (1993), Merriam-Webster, Incorporated, Chicago.
Weinfurt, K.P. (2000), ‘Repeated Measures Analysis: ANOVA, MANOVA, and HLM’, in: Grimm,
L.G. and P.R. Yarnold (eds), Reading and Understanding More Multivariate Statistics, Washington, DC:
American Psychological Association, 317-361.
Wensley, R. (1981), ‘Strategic marketing: Betas, boxes, or basics?’, Journal of Marketing, Vol. 45,
173-181.
Wensley, R. (1994), ‘Making better decisions: The challenge of marketing strategy techniques’,
International Journal of Research in Marketing, Vol. 11, 85-90.
Wesselink, R., L.J. Pagrach, M. Mulder, J.Chr.J. Bruijstens and H. Miltenburg (2001), De
Inkoopprofessie in Beeld – Een Onderzoek naar de Beroeps- en Opleidingsprofielen voor Inkoopprofessionals,
Stichting NEVI Examens, Zoetermeer, The Netherlands (Dutch text).
Wilemon, D.L. (1971), ‘Power and Negotiation Strategies in Marketing Channels’, The Southern
Journal of Business, 71-82.
Wilkinson, I. and L.C. Young (2002), ‘On Cooperating Firms, Relations and Networks’, Journal of
Business Research, Vol. 55, 123-132
Williamson, O.E. (1975), Markets and Hierarchies: Analysis and Antitrust Implications, Free Press,
New York.
Williamson, O.E. (1985), The Economic Institutions of Capitalism, Free Press, New York.
Williamson, O.E. (1991), ‘Comparative Economic Organization: An Analysis of Discrete Structural
Alternatives’, Administrative Science Quarterly, Vol. 36, June, 269-296.
Wind, Y. (1980), ‘New twists for some old tricks’, The Wharton Magazine, Spring, 34-39.
Wind, Y. and V. Mahajan (1981), ‘Designing Product and Business Portfolios’, Harvard Business
Review, Vol. 59, January/February, 156-165.
314
Wind, Y., V. Mahajan and D.J. Swire (1983), ‘An Empirical Comparison of Standardized Portfolio
Models’, Journal of Marketing, Vol. 47 (Spring), 89-99.
Wiseman, F. and M. Billington (1984), ‘Comment on a Standard Definition of Response Rates’,

Journal of Marketing Research, Vol. 21, 336-338.
Womack, J.L, D.T. Jones and D. Roos (1990), The Machine That Changed The World, Rawson
Associates, New York.
Wood, L. (1990), ‘The End of the Product Life Cycle’, Journal of Marketing Management, Vol. 6,
145-155.
Wynstra, J.Y.F. (1998), Purchasing Involvement in Product Development, Doctoral thesis, Eindhoven
Centre for Innovation Studies, Eindhoven University of Technology.
Wynstra, J.Y.F. and E. ten Pierick (2000), ‘Managing Supplier Involvement in New Product
Development: A Portfolio Approach’, European Journal of Purchasing & Supply Management, Vol. 6,
No. 1, 49-57.
Yin, R.K. (1994), Case Study Research: Design and Methods, Sage Publications, Thousand Oaks,
CA, 2nd ed.
Yorke, D.A. (1984), ‘An Interaction Approach to the Management of a Portfolio of Customer
Opportunities’, Market Intelligence & Planning, Vol. 2, no. 3, 5-22.
Yorke, D.A. and G. Droussiotis (1994), ‘The Use of Customer Portfolio Theory: An Empirical Study’,
Journal of Business & Industrial Marketing, Vol. 9, No. 3, 6-18.
Young, L.C. and I.F. Wilkinson (1997), ‘The Space Between: Towards a Typology of Interfirm
Relations’, Journal of Business-to-Business Marketing, Vol. 4, No. 2, 53-97.
Yu, J. and H. Cooper (1983), ‘A Quantitative Review of Research Design Mail Survey Response
Behavior’, Public Opinion Quarterly, Vol. 55, 613-639.
Zaheer, A., B. McEvily, and V. Perrone (1998), ‘The Strategic Value of Buyer-Supplier
Relationships’, International Journal of Purchasing and Materials Management, Vol. 34, No. 3, Summer,
20-26.
Zallocco, R.L., D.W. Scotton and D.A. Jeresko (1983), ‘Strategic market planning in the commercial
airline industry’, Journal of Academy of Marketing Science, Vol. 11, 404-416.
Zand, D.E. (1972), ‘Trust and Managerial Problem Solving’, Administrative Science Quarterly, Vol.
17, June, 229-239.
Zolkiewski, J. and P.W. Turnbull (2001), ‘Relationships Portfolios – Past, Present and Future’,
http://web.bi.no/imp2001, The 17th Annual IMP Conference, Oslo, Norway.
Appendix A
Interview guide for the case studies
1 Background information on the company

- main activities
- principal products and markets
- organizational structure (business units)
- company’s strategic focus
- annual sales
- number of employees
2 Background information on purchasing

- annual purchase spend
- main purchasing categories
- purchase spend, divided according to purchasing categories
- centralized and/or decentralized purchasing
- development of the purchasing function
- position and perceived importance of purchasing
3 Purchasing and supply strategies

- main objectives of purchasing
- basic principles for purchasing practice
- buyer-supplier relationship strategies
- connection between purchasing and business strategy
- development of purchasing plans and strategies
4 Purchasing portfolio analysis

- main purposes and use
- occasions, frequency, situations and timing
- Kraljic-based or otherwise
- choice of dimensions
- factors per dimension
- labels for quadrants
316
- generic strategic recommendations per category

- measurement of dimensions/variables (establishing positions in the matrix)
- attributed results
5 Portfolio based purchasing strategies

- basic principles for portfolio based purchasing strategies
- procedures and additional tools for strategy development
- specific purchasing strategies for specific categories
- objectives of specific, portfolio based strategies
- conditions and contingencies for strategies and categories
Appendix B
Sources of the case studies
1. DSM
Guide to Purchasing Marketing Planning (1998), DSM N.V.
Internet http://www.dsm.nl/, updated 00 02 09.
Interviews with:
- Willem F. van Oppen, Director Purchasing Services, DSM Services, 00 02 18, 00 03 20, 00
04 21, and 00 05 22
- Wim A. B. Donners, President of DSM Elastomers Europe/Asia, 00 03 31
- Jan H. Kruit, Business Group Director DSM Hydrocarbons, 00 08 02 and 00 07 23
- Andriëtte Dobbelsteen, Business Process Analyst, DSM Services, 00 10 26
- Kees Aartsen, Marketing Manager Aromatics, DSM Hydrocarbons, 00 10 26
- Malcolm Saggers, Director Chemicals & Raw Materials, DSM Services, 00 10 26
- Hans Timmermans, Global Program Manager E-Procurement, DSM, 00 11 03.
Louwers, D, W. van Oppen and G. Walravens (1999), ‘DSM werkt succesvol met
multidisciplinaire inkoopteams’, Tijdschrift voor Inkoop & Logistiek, Vol. 15, January-
February, 8-12 (Dutch text).
State of the Union between BG/BU’s & Purchasing Services (1999, 2000), DSM Purchasing
Services.
318
2. Akzo Nobel Coatings

Internet, http://www.akzonobel.com/ updated 00 09 11
Interviews with:
- Dick L. Bartelse, Purchasing Vice President Decorative Coatings International, Akzo
Nobel Decorative Coatings, 00 09 22 and 00 11 01
- Bert van den Heuvel, Purchasing Manager Raw Materials/Packaging, Akzo Nobel
Decorative Coatings, 00 10 25
- Ruben Manniën, Purchaser Raw Materials, Akzo Nobel, Car Refinishes, 00 11 09
- Wynanda de Vries, Purchaser Packaging, Akzo Nobel, Car Refinishes, 00 11 09.
Bartelse, D.L. (2000), ‘Het inkopen van grondstoffen’, Handboek Internationaal Inkopen,
April, B 1300-1 – B 1300-20 (Dutch text).
Annual Report 1999, Februay 2000, Akzo Nobel.
3. TE STRAKE – Engineering & Production

TE STRAKE – Jaarverslag 1999 (annual report).
Internet, http://www.testrake.nl/, updated 01 02 05
Interviews with:
- Wilfred van der Made, Strategic Buyer, Business Unit Engineering & Production, 01 02
09, 01 04 17,
and 01 06 11.
- Hans Goedhart, Demand Chain Manager, Business Unit Engineering & Production, 01
05 02 and 01 06 11.
- Jos Gunsing, Technology Coordinator, Business Unit Engineering & Production, 01 05
02
- Peter Hamers, Initial Purchasing Electronics, Business Unit Engineering & Production,
01 05 02 and 01 06 11.
- Charmaine Kuijpers, Technical Purchasing/Marketing, Business Unit Engineering &
Production,
01 05 02 and 01 06 11.
Appendix C
Reactions to the mutual dependence

model
1. Reactions DSM
Buyer’s dependence
1. Financial magnitude is not considered as a determinant of DSM’s dependence on
suppliers.
2. There are different forms of criticality to buyer’s dependence. In some cases there is a
logistic dependence, because products are indispensable in business operations
(production processes). In other cases, there is a technological dependence, because
products and services are technologically very complex. DSM has important service level
agreements for the maintenance and repair of critical machinery.
3. The number of alternative suppliers is an important determinant to DSM’s dependence
on suppliers. More specific, in several cases DSM has no actual alternatives, because
suppliers hold critical patent positions.
4. Finally, switching cost is considered as an important determinant too.
Supplier’s dependence
1. Financial magnitude is generally considered as the most important determinant of
supplier’s dependence. Financial magnitude, to be more specific, refers to sales and
profit, based on current business and on potential business.
2. Criticality is a determinant of supplier’s dependence, in the sense that DSM provided
unique know how. Suppliers are dependent of DSM, because they enhance their
competences, capabilities and possibilities for development.
3. The number of alternative buyers is, in a general way, important to the supplier’s
dependence.
4. The same goes for switching cost.
320
2. Reactions Akzo Nobel Coatings
1. Financial magnitude is not considered as a determinant of dependence on suppliers.
2. Criticality can raise problems of dependence, in case of dedicated products. To be more
specific, technical dependence is usually caused by specifications that limit alternative
possibilities.
3. The number of alternative suppliers is the most important determinant to Akzo Nobel’s
dependence on suppliers.
4. In case of leverage and non-critical items switching costs are usually low, while in case
of strategic and bottleneck items switching costs can be high.
supplier’s dependence.
2. Criticality can be a determinant of supplier’s dependence, if the supplier considers
Akzo Nobel as an important customer for reasons of image and reference. Suppliers
might want to use Akzo Nobel as a referent customer in their communication with other
buying companies.
3. The number of alternative buyers is not an important determinant to the supplier’s
dependence.
4. The same goes for switching cost.
3. Reactions TE STRAKE - Engineering & Production
1. Financial magnitude is not considered as a determinant of dependence on suppliers.
2. There are different forms of criticality to buyer’s dependence. There is a logistic
dependence, because a large number of items have to be delivered on a tight schedule for
the assembly of the end product. The business unit needs reliable and short lead. In case
of the key preferred suppliers, there is a technological dependence, because of the
technological expertise of suppliers.
3. The number of alternative suppliers is a very important determinant to TE STRAKE’s
dependence on suppliers. On occasions, a major customer lists a number of second tier
suppliers that have to be contracted. The interference and the demands of major customer
can create dependence on certain suppliers.
4. To conclude, switching costs are usually high, especially in case of dedicated suppliers
with whom long-term relationships are maintained.
supplier’s dependence.
2. Criticality as such is not considered as a determinant of supplier’s dependence,
although some key suppliers depend to a certain extent on the technological input of the
buyer.
3. In a general sense, the number of alternative buyers is a determinant to the supplier’s
dependence.
4. Some suppliers are very dependent on TE STRAKE, partly due to the relative financial
value of purchases, partly due to the switching costs that have to be made, when
partnering with another buyer.
322
Appendix D
A 5-step approach to the application

of purchasing portfolio analysis
1. Preparation
- determine objective and purpose of the analysis
- assess sources of available information
- decide on time and budget
- form a group of participants
2. Design and filling in the matrix

- determine the scope of the analysis (items, level of detail and aggregation)
- select a measurement methods (‘one-by-one’, ‘consensus’ or ‘weighted factor scores’)
- determine the variables and the weights
- determine the scores of the factors
- determine the total scores/positions of all the items
3. Interpretation of the results

- understand and review the positions in the matrix
- if necessary, re-adjust positions
- reflect on the consequences
4. Determination of strategic actions

- consider each time: moving or keeping positions
- select strategies for bottleneck items
- select strategies for non-critical items
- select strategies for leverage items
- select strategies for strategic items
5. Evaluation and follow-up

Appendix E
Action research: using the outline in

practice
E Action research: using the outline in practice

This appendix summarizes the results of an action research project that was carried out
within Emtec High Tech Industrial Solutions, a Dutch company that provides industrial
production equipment for the high-tech industry. The project was aimed at developing
differentiated purchasing and supplier strategies for a new product, based on a portfolio
analysis. In the Spring of 2002 the research was carried out by Edwin Noordman
(Manager Logistics at Emtec) for his end-term project of the Executive MBA Programme
of TSM Business School. It was supervised by Cees J. Gelderman (Open University of the
Netherlands).
Background information
The semiconductor industry, with 70% of Emtec’s sales, is the company’s most important
market. Emtec is positioned in four market segments, the largest of which is the supply of
(sub) assemblies to the OEM market. Emtec is both a main supplier and a second tier
supplier. The other segments include the MRO market, engineering services, and the
design and manufacturing of customer-specific equipment. In 2001 approximately 400
employees realized a total turnover of about 45 million Euro. The semiconductor industry
is very cyclical. The overall growth rate is at least 15% per year, varying from 30%
growth in good years to a 20% decline in a normal downturn. The latest downturn
started in the first half of 2001 and lasted for more than a year. With a 30% to 40%
declining market it was one of the severest downturns ever. For Emtec this downturn had
serious consequences. The order intake dropped to almost zero and in a time frame of a
couple of months the workload dropped 30% to 40%. Especially in the production unit
this led to a big reorganization, including a head count reduction of over 25%. This
downturn showed in a painful way the weak position in the supply chain. Customers
simply cancelled or postponed orders without any financial compensation. An attempt to
pass this risk further down to the suppliers did not succeed. The power position of Emtec
was too weak and its dependency on a limited number of customers was too high. In
order to reduce this dependency, the strategy of Emtec was to start up a new OEM
324
business alongside the current activities. One of the main initiatives was the introduction
of a new product, the ASIGS (Automated Silicon on Insulator Gluing System). The ASIGS
is a tool for Substrate Transfer Technology (STT) in the semiconductor industry, jointly
developed with Philips Research. The ASIGS was designed specifically for semiconductor
manufacturers to develop Semiconductor On Anything process capability. In the STT
process silicon wafers, including the Integrated Circuits, are glued on insulating
substrates like glass. Removing the performance limiting silicon substrate results in
reduced power consumption and higher performance of Radio Frequency components.
Research questions
This research was aimed at developing differentiated purchasing strategies for the
ASIGS, based on a portfolio analysis in order to contribute to Emtec’s competitive
advantage. Within the boundaries of the research a reduction of costs and of risk, and a
better use of suppliers should increase the company’s competitive advantage.
In accordance with this objective three research questions were formulated:

1. How should the components and materials of the ASIGS be positioned in the
purchasing portfolio matrix?
2. How should these positions be assessed, in view of the problems and developments in
the semiconductor industry?
3. What strategies could be recommended in order to improve the positions that are
found in the matrix?
In the semiconductor industry the pace of new developments is very high. The
investments and capacity required for this continuous development are also very high.
Involving suppliers in this development can reduce costs and the time to market
(Wynstra, 2000). Involving suppliers in an early stage might increase the dependency and
can therefore be contradictory to reducing risk. The key issue is to find the balance
between involving suppliers and reducing the risk. An analysis of the current situation
made it clear that Emtec had been very dependent on a limited number of suppliers.
Obviously, this resulted in a high supply risk. Non-performance of a single supplier
would have a great influence on Emtec’s own performance. In the semiconductor
industry reliability and short delivery times are key issues. A low price is an important
factor in the competitive advantage but certainly not the only one. Reliability of the
organization (and not only the product) is very important. The portfolio analysis should
therefore include a critical review of the positions in the matrix that are likely to be
achieved. Are they acceptable? Are certain positions likely to cause problems? Before
strategic actions are determined, it is imperative to complete a further process of
interpreting and reflecting on the results. The filling in of a matrix should be considered
as the starting point of portfolio analysis, not the finishing point (Gelderman and Van
Weele, 2002a).
Design and results

To address the research questions it was concluded that the Kraljic matrix would be the
most suitable model, as it aims at minimizing supply risk and making the most of buying
power. The Kraljic model is generally accepted and used, although there is little literature
on the actual use and application of the portfolio analysis. Gelderman and De Boer (2001)
suggested a 5-step approach for the portfolio analysis, based on empirical research. The 5
steps are: 1) Preparation. 2) Design and filling in the matrix. 3) Interpretation of the
results. 4) Defining strategic actions. 5) Evaluations and follow-up. In this research
project, this stepwise approach was used, because it closely fitted the research objectives
and the research questions.
Step 1 Preparation
In this preparation phase a cross-functional team was formed. Most of the preparation
time was spent on getting the organization ready for this research. Although the top
management was interested in the research, other parts of the organization had to be
convinced as well, and not just the people that were needed on the team, but their line
managers as well. At the end of the day they were the ones to decide how much time
their people were allowed to spend on the project. In this preparation phase close-
working relationships were developed with the project manager responsible for the
product design and the manager of the purchasing department.
What was the objective?

The objective of the study was to define differentiated purchasing strategies. During the
preparation some complementary objectives seemed to be important as well. In the
second half of 2002 a new model of the ASIGS was to be designed, based on the current
system. The engineers needed the outcome of this analysis to improve the design.
Furthermore, the purchasing manager was interested in using the Kraljic method more
frequently.
Additional objectives were:
- to give input for the design of the next generation tool,
- to introduce the Kraljic portfolio analysis to improve the professionalism of the
purchasing department and to improve the co-operation between the engineering and
purchasing departments.
What information was available?

For the portfolio analysis the team needed the complete Bill Of Materials (BOM) of the
ASIGS and specific purchasing information of all these components. The PDM (Product
326
Data Management) system contained the BOM. The purchasing information, if available,
should be in the ERP system. The ERP system also contained a duplicate of the BOM. But
it turned out that the information in the DPM application was the most reliable. As a
starting point the BOM from the DPM system was used.
What was the available time and budget?

The available time was limited by the start of the next generation design. The time
available for the portfolio analysis was about 2 months. This relatively short time frame
meant that in some of the steps the team would be unable to go through every detail. In
some cases assumptions had to be made. As far as the budget was concerned, there were
no constraints, although the number of hours the team members could spend on the
project was not unlimited.
Who were to participate in the team?

Taking the objectives of the research into account representatives of various departments
had to form part of the team. The team should include a purchaser, an engineer and
someone from the production department. The role of the last one was to translate the
engineering data into production and purchasing information. He also decided which
custom-made parts should be made inhouse and which should be outsourced.
Since there is a distinction between a mechanical part and an electro-technical and
software part it was imperative to include representatives from both disciplines. In
addition, the head of the purchasing department and project manager in charge of the
design had to be on the team as well. The role of the manager of the research project was
mostly to provide the theoretical framework and to direct the process by chairing
meetings and processing the collected data.
With a total of 9 individuals, the team was quite large. In collecting and processing data
this was likely to cause some coordination problems. Therefore, it was decided to split up
the team in a mechanical and an electro-technical group. After selecting the factors with
the whole group, the groups filled in the two matrices. When the matrices were
completed, they were combined and strategic actions were planned by the group as a
whole. As all departments were involved, a sense of awareness and acceptance was
created, which helped to keep the situation workable.
Step 2 Designing and filling in the Matrix
To answer the first research question, the project team had to design the matrix by
defining the level of aggregation, the variables and their weight. During the first meeting
with the team these issues were discussed.
Level of aggregation or detail:

In discussing the level of aggregation the purchasers at first wanted to group items as
much as possible and reduce the number of suppliers. In their opinion one group per
supplier was sufficient. The engineers on the other hand wanted to see as much detail as
possible. If they got detailed information out of the portfolio analysis it would be easier to
improve the design. By excessive grouping of individual items the outcome would not be
specific enough to be used as input for a redesign.
For the grouping of items the right balance had to be found between the level of
aggregation and detail. Too much detail meant that the analysis would take too long,
which would cause serious problems as time was limited. Special attention had to be paid
to specific components (e.g. machined parts), since it was unknown at the time how these
should be purchased. Were all these parts going to be purchased as individual
components or as complete sub-assemblies?
As a starting point the criteria for grouping were:
- Simple standard components like fasteners, pneumatics, bearings etc. were grouped.
- Standard components critical for the function of the ASIGS were not grouped with other
non-critical components.
- The type of manufacturing method combined specific components.
- Exceptions to these were specific components that have to be processed together. These
were combined into one group.
Determining the variables and their weight:

In setting the variables the team came to the conclusion that the analysis still had to be
manageable. This meant that only a limited number of variables could be taken into
account. However, with only a small number of groups, more factors could be taken into
account.
As a starting point the team set the following factors:
- Profit impact:
- Price of the item / group as a % of the total price
- Supply risk:
- Number of suppliers.
- Replaceability in the current design.
- Risk of failure. What is the risk when an item/group does not meet the required
specification or is failing?
In discussing the weights of the three factors on the supply risk axis the team could not
determine which factor was the most important. From the purchaser point of view the
number of suppliers was the most important factor. For the designer the replaceability
carried more weight. But after some discussion the team concluded that the strategy and
the customer point of view were also very important. At this point in the analysis it was
decided that (for the moment) all factors were to be equally weighted.
328
Filling in the matrix: grouping

As described previously, the team was divided into two sub teams: one team for the
mechanical part and one for the electro-technical part. The sub teams started to split up
the BOM of the complete machine into mechanical and electro-technical parts. Based on
this list a group for each purchase component was defined. It turned out that for a lot of
parts it was quite easy to define these groups. For some items it was more complex.
Generally speaking, in defining the groups one has to always keep in mind which
components are critical and which are not, otherwise one runs the risk of including a
crucial item in a group of non-critical items. The result could be that, for example,
bottleneck components are overlooked. One should also bear in mind which factors
determine the two axes of the matrix.
After the first session most item groups were defined. There were some items left
ungrouped and undefined. After processing the data in the second session the remaining
items were filled in and some corrections were made. The team found 37 groups
altogether.
Filling in the matrix: variables

Finding the values for the profit impact axis was rather straightforward, as the purchase
value was the only factor. These data were partly available in the systems. For the
missing items the purchasers contacted the supplier to request a quotation. When all the
prices were available, the total per group was summed up.
Determining the supply risk was more complex. There were more variables and these
variables were not directly linked to hard and straightforward data. To fill in the data the
consensus-based method was used. For the selected factors (number of suppliers,
replaceability, risk of failure) one cannot easily set a hard number. The team used a 1 to 5
scale for each factor. Each member filled in these factors for each group. In a meeting the
team discussed each item and tried to come to a common understanding of the reasons
why the risk would be high or low. Although at times this led to some heated
discussions, consensus could be found on all of the standard components.
In discussing these items it turned out that the definitions of replaceability and supply
risk were not completely clear. When was something difficult to replace and when was it
not? The project team came to the conclusion that one measure for replaceability could be
the number of hours needed to adapt the design if the specific component were no longer
available.
In defining the risk of failure there was a similarity with the so-called FMEA (Failure
Mode Effect Analysis) method. This method is used to find possible failures in the design
phase of equipment. How does a failure in a component or not meeting the specifications
affect the function of the end product?
In determining the supply risk for the special components (custom made) the team faced
an unforeseen problem. The first factor (number of suppliers) was easy to determine. The
purchasers had a good view on the supply market. But as far as the replaceability and the
risk of failure was concerned, it turned out to be impossible to make a judgement. At first
the team combined all the outsource parts into three groups, one group for each process
category (machine parts, sheet metal & welding, and plastics). It was concluded that the
level of aggregation was too high. It was not possible to determine the risk of failure. The
replaceability was not applicable for the outsourced parts. Since those parts were
specifically designed for the machine they could not be replaced anyhow.
Step 3 Interpretation of the results
At this stage, all data were collected and the supply risk of the items was determined, so
that the portfolio could be filled in (see figure E1). The supply risk axis is the sum of the
scores of the three factors: number of suppliers, replaceability, and risk of failure. For
some items the replaceability was not applicable. To be able to compare these items with
the other items the sum of the two remaining factors was multiplied by 1,5 to fit in the
same scale. For the profit impact the factor was the total purchase value of the item in
that group. The values varied from less then 100 Euro to almost 10,000 Euro per group.
For reasons of interpretation, a logarithmic scale was used to represent the profit impact
axis. In the matrix 15 items were located in the leverage and the strategic quadrant,
representing 90% of total value.
Figure E1 Current ASIGS portfolio

330
In evaluating the individual items one item was placed in the wrong quadrant. One
bottleneck item seemed to be misinterpreted and was therefore moved to the routine
quadrant. For the remaining items some discussion arose about the position within the
quadrant. But these were all minor issues and in the project meeting the group decided
that it would not be worthwhile to make additional adjustments to the portfolio. The
over-all picture and the relative position of the items made sense. The project team found
this to be a good representation of the current situation. The first research question of the
project was answered by the matrix (figure E1) that was found for ASIGS.
It was concluded that the quality of the portfolio was sufficient to start with the
development of the strategic actions, addressing the second research question. “How
should these positions be assessed, in view of the problems and developments in the
semiconductor industry?”
Would Emtec encounter the same problems with the ASIGS as during the last
semiconductor downturn? What would the power position be if no improvement could
be made? Would Emtec be able to keep pace with the fast development of new
technology?
Assessing the matrix, in view of these problems, the project team concluded that
improvement was required to prevent future problems. With the current portfolio there
was a realistic risk of falling back into the same problems again. Looking at the over-all
matrix the team concluded that the number of items in the strategic and bottleneck
quadrant was too high. This resulted in a too high dependency on the suppliers. This had
been one of the main problems in the recent downturn. So a strong desire existed to
reduce this dependency. If dependency cannot be reduced it should be managed in a
proper way. Furthermore, the team concluded that the number of routine items was
relatively high. With an increasing market demand these time-consuming items could
hinder the efficiency of the purchasing department. Extra capacity might be needed, but
this would increase the cost price. There was a wish to improve the efficiency by
reducing the number of suppliers.
Step 4 Defining strategic actions
Kraljic (1983) defined a set of generic strategies. Gelderman and Van Weele (2002a)
elaborated these strategies into a larger number of strategies, taking into account
different conditions and circumstances. A distinction is made between strategic actions to
switch to another position and strategic actions that imply ‘staying in the same category’.
The first type of strategy is of a more pro-active, aggressive nature. The second type of
strategy is of a more re-active nature, accepting the status quo. Obviously, moving in the
matrix is not always possible or desirable. Figure E2 shows these different strategies, in
the ASIGS portfolio.
Figure E2 Generic strategies in the matrix

(Based on Gelderman and Van Weele , 2002a)
In this phase all the items in the ASIGS portfolio could be assessed with these generic
strategies. In assessing the position in the matrix and defining strategic actions, special
attention had to be paid to the problems that Emtec had faced in the recent
semiconductor downturn. In the downturn, Emtec’s dependency on suppliers resulted in
serious problems. As described by Gelderman (2000), the Kraljic matrix can be seen as a
buyer-supplier dependency matrix. Reducing the supply risk will lead to a lower
dependency. In some cases a strategic partnership, and therefore a higher dependency, is
desirable. This, however, can only be the case in a win-win situation. In terms of the
research objective, a partnership should contribute to the competitive advantage. When a
partnership is not contributing, the items in the strategic quadrant should be moved to
the leverage quadrant by reducing the dependency.
To translate these strategies into strategies applicable for Emtec, and to answer the third
and final research question, the project team was split up into three subgroups.
Discussing all the items in the whole group would not have been very efficient. After
completing the homework in the sub-groups the results were combined. The homework
for the three groups was to:
- analyze the design to find whether improvements in the position of the bottleneck and
the strategic items were possible. This group consisted of the two engineers and the
project manager for the redesign.
332
- define the strategic purchasing actions should the items hold their position deliberately
or if improvement was not possible. This group consisted of the two purchasers and the
manager of the purchasing department.
- analyze whether the routine products could be pooled. Parallel to this research a group
of engineers worked on a standardization project. The objective of this project was to
reduce the number of suppliers and standardize components in newly designed
equipment. The ASIGS had been designed before this standardization took place. The
task of the last group was to check whether the design of the ASIGS complied with these
standards. A member of the standardization group was invited for this task.
In a project team meeting the results were combined in one overview, with for each item
the relevant possibilities for improving the position, the purchasing strategy and the
estimated new position in the matrix.
Strategic items:
In the found portfolio there were 6 items in the strategic quadrant. In 4 cases a reduction
of
the supply risk would hardly be possible, whereas a strategic partnership could
contribute to the competitive advantage of the company. These 4 items could be pooled
into 2 groups, one group for the ‘control system’ of the ASIGS and one for the ‘motion
systems’. One of the leverage items was added to the ‘control system’ group to increase
the volume and make better use of the partnership advantages.
A partnership for the remaining 2 items would probably not contribute to the competitive
advantage. Extra suppliers should be qualified for these technical complex parts in order
to reduce the supply risk and to re-position the items in the leverage quadrant.
Furthermore, the project team advised to keep the technology to manufacture these parts
inhouse.
For one group, the first steps to a partnership had already been taken. In spring 2002
Emtec was in the middle of a qualification procedure to become a certified partner of the
‘control system’ supplier. This certified partnership included additional support,
guaranteed delivery of components till ten years after the end-of-life cycle, worldwide
spare parts delivery within 24 hours, and additional discounts. All these issues should
contribute to the competitive advantage of the company. Especially the spare part service
was considered to be very important, since continuity is a major issue in the
semiconductor industry.
In the field of the ‘motion systems’ there was no strategic partnership yet. Although the
current suppliers were actively involved in the design of the ASIGS, there was no formal
agreement and long-term advantage for Emtec. Preferably, a supplier should be selected
for all ‘motion systems’. The partnership should include aspects like spare part delivery,
guaranteed delivery after end of life and support in the design phase. Like the ‘control
system’ partnership, securing the deliveries should improve the continuity. Involving the
supplier or even completely outsourcing the design of the ‘motion system’ should reduce
the development time and costs. The project team doubted whether the current suppliers
were capable to meet these requirements. Until one or more new supplier were selected,
the current suppliers should be contracted to secure deliveries and spare parts
availability. Involving suppliers in product development, as described by Wynstra (2000),
requires a differentiated approach, depending on the development risk and the degree of
responsibility held by the suppliers. Further analysis was recommended of supplier
involvement in Emtec’s product development.
Leverage items:
As the leverage quadrant is considered the most favorable position in the matrix, the
project team tried to further improve the position in this quadrant. By reducing the
supply risk and by pooling individual groups, the buying power could be further
increased. In the ASIGS portfolio 13 items, both from the leverage and routine quadrant,
were pooled into 3 groups. Two of these groups, one for electro-technical and one for
mechanical, represented a wide range of standard components. For each group there was
a number of known suppliers. The strategy would be to bargain for the best deal and
switch between suppliers if a better deal could be achieved. The project team however
expected the ERP system to be a barrier for switching suppliers. The main problem was
the limited use of component and material encoding in combination with the linkage to
suppliers. It is very labor-intensive to change the supplier, since this often results in
different encoding of standard components. This could be a roadblock to selecting a
different supplier. The project team advised to make this a key issue in the selection of a
new ERP system.
The third pooled group consisted of the collection of machined parts with a low supply
risk. The buying power should be increased by grouping. The strategy would be to go for
the best deal, including not only the price, but also aspects such as quality, lead-time, and
delivery performance (all key issues). A possible strategy would be to select suppliers in
low wage countries. However, within Emtec there was no experience in this field. Further
investigation or gaining expertise in this field would be needed before this option could
be put into practice.
The machined parts in the strategic and bottleneck quadrant should not be pooled in this
group. These more complex parts would limit the number of suppliers capable of
manufacturing these parts and reduce the buying power.
Routine items:
The main strategy for the routine items would be to pool and to standardize. These items
were partly pooled, as described in the previous section. For the remaining items the
334
design should conform to the prescribed standards. The project team found several items
that did not conform to Emtec’s design standard. The project team advised to modify the
design, where possible, to meet these standards. Not only should the number of suppliers
be reduced, but the number of different components as well.
When items conform to the standard, the over-all buying power would increase. Looking
solely at the position in the ASIGS portfolio, the position would probably remain in the
routine quadrant. But due to standardization the number of suppliers would reduce,
which would lead to a more efficient processing. Other possibilities to reduce the
purchasing costs would be the use of electronic data interchange or putting items on
stock.
Bottleneck items:
The project team defined 3 strategic actions for the bottleneck items. The first was to
move the items to the routine quadrant by reducing the supply risk. In case of the
machined parts qualifying extra suppliers and keeping the manufacturing technology
inhouse could reduce the supply risk. For the standard component placed at the top of
the bottleneck quadrant, the project team advised to contract the supplier to secure
deliveries. The bottleneck item with very low profit impact should be put on stock.
Step 5 Evaluation and follow-up
Some of the strategic actions could be easily implemented in the short term. Others
would require more time and energy. In the implementation a distinction can be made
between the actions that require a modification of the design (e.g. standardization) and
actions without any modification of the design (e.g. pooling). The engineering
department should make these design modifications. It makes sense to combine this
action with the design of the new fully automated ASIGS. This new system is based on
the current ASIGS, combined with a fully automated robot system for use in mass
fabrication. By combining the modification of the current system and the design of the
new system an optimal effect should be achieved with relatively little effort. In the
redesign, a high level of standardization should be aimed for.
Most of the improvements without modifications of the design are the responsibility of
the purchasing department. Extra suppliers should be selected, bottleneck suppliers
should be contracted, routine items could be pooled. These are all issues within the
responsibility of the purchasing department. Completing the partnership for the ‘control
systems’ and selecting a strategic partner for the ‘motion systems’ should be a joint effort
from the purchasing manager and the managers in charge of the engineering and OEM
units. To make the best use of suppliers all departments should be committed and
involved in the selection process. The project team advised to complete these actions
before starting with the design of the next generation machine. This new design could be
the ideal test case for these newly formed partnerships.
The 2001-downturn and the subsequent liquidity problems resulted in a lot of pressure
on relations with suppliers. Many suppliers became very reserved in starting new
business with Emtec. This would be hindering the purchasing manager and could be a
serious threat in the implementation of the strategic actions. Therefore, it was seen as a
top priority to solve the liquidity problems and stabilize the relations with the suppliers.
Solving these problems was far beyond the influence of the members on the project team.
Top management would have to deal with these problems.
When all actions are implemented the portfolio will look differently, probably much like
figure E3. This future portfolio is based on estimated values. Of course, the actual
portfolio can only be filled in when the strategic actions are implemented and the effect
on the supply risk and profit impact can be determined.
Figure E3 Future, improved ASIGS portfolio
Comparing the current portfolio with the future portfolio, the improvements are:
- Reduction of groups from 37 to 25 by pooling.
- Reduced number of bottleneck items from 3 to 2 and secured the delivery of these
bottleneck products.
- Reduced number of strategic items from 6 to 3 and partnerships for these items, which
contributes to the competitive advantage.
- Increased value in the leverage quadrant from 47% to 61% of the total purchase value.
The objective of this research was to contribute to the competitive advantage of Emtec by
developing differentiated purchasing strategies. Within the boundaries of this research
the competitive advantage can be expressed in terms of reducing costs in the broad sense,
reducing risk, and making better use of suppliers (for instance, reducing development
336
time and costs). Translating the above mentioned improvements into these three aspects
produces the following results:
- The increased value in the leverage quadrant, from 47% to 61% of the total purchase
value, results in more buying power. With more buying power better deals can be made,
resulting in a lower cost price. Also, the reduced number of suppliers by pooling and
standardization of the routine items leads to a more efficient purchasing process.
- At the end of the research project, spring 2002, it was not possible to calculate the
achievable cost price reduction. The project team estimated that a cost price reduction of
at 10% to 20% would be realistic.
- Repositioning the bottleneck and strategic items by reducing the supply risk leads to
less dependency. The number of items with a high supply risk will be reduced from 9 to
5. For the remaining items the risk should be covered by strategic partnerships, contracts
securing delivery and extra stock. Completely eliminating all the risk is not possible. Also
the items in the leverage and routine quadrant are not 100% risk free. The project team
evaluated the remaining risk as acceptable.
- By starting partnerships for the items in the strategic quadrant a win-win situation
could be achieved. Emtec can make better use of these suppliers in two ways. Involving
the supplier in the design of new equipment will reduce development time and costs.
Especially for Emtec, a partnership in the field of ‘motion systems’ is seen as valuable.
The supplier is not only delivering the hardware but is responsible for the complete
motion system, including the performance and part of the implementation. By involving
the supplier specialized knowledge will be available. Once the equipment is installed in
the field, the supplier can be used to take care of the spare part supply and operational
support. In the semiconductor industry reliability and continuity are key issues. Making
the supplier responsible for part of these issues is evaluated as very helpful.
To conclude, the project team was convinced that implementing these strategic actions
would significantly contribute to the competitive advantage of Emtec. The 5-step
approach for the portfolio analysis provided sufficient and appropriate guidance to
successfully complete the purchasing portfolio analysis. The project team experienced
this project and the introduction of a purchasing portfolio approach as very useful. It
helped to get a better understanding and more insight into purchasing issues in general
and the ASIGS in particular. By close co-operation, the engineers and the purchasers
started to develop a better understanding of each other’s worlds and accompanying
problems. In the new strategy of Emtec, starting an OEM business, the purchasing
function should get a more prominent place and should significantly contribute to the
competitive advantage of the company. This project can be seen as a first step in the
professionalization of the purchasing function within Emtec.
Appendix F Elaboration and

overview of the 9 scenario’s
Scenario Purchasing Kraljic category: Additional information Strategic direction

strategy treatment variables
#1 safety stocks bottleneck: - stay in the same
- low profit impact category
- high supply risk
#2 decomplex bottleneck: - move to

and new supplier - low profit impact non-critical
- high supply risk
#3 pooling non-critical: - move to leverage
- low profit impact
- low supply risk
#4 individual ordering non-critical: - efficient ordering stay in the same
- low profit impact - e.g. Yellow Pages or category
- low supply risk purchase card
#5 partner of leverage: - price, quality, delivery stay in the same
convenience - high profit impact and flexibility category
- low supply risk - only short term
cooperation on buyer’s
terms
#6 towards a strategic leverage: - high expectations for the move to strategic
partnership - high profit impact supplier (value,
- low supply risk partnering, expertise)
#7 strategic partnership strategic: - good relationship with stay in the same
- high profit impact the supplier category
- high supply risk - excellent performance
#8 locked-in strategic: - poor relationship with stay in the same
‘partnership’ - high profit impact the supplier category
- high supply risk - forced relationship
#9 terminate strategic: - poor relationship with move to leverage
partnership, find new - high profit impact the supplier
supplier - high supply risk
338
Appendix G
Questionnaire (in Dutch)
Onderdeel A Inkoop - algemeen

A1 In welke bedrijfstak is uw bedrijf voornamelijk werkzaam? (één antwoord
aankruisen)
0 elektrotechnische industrie 0 transportmiddelen-industrie
0 chemische industrie 0 bouwmaterialen/glas- en
aardewerkindustrie
0 metaalproducten-industrie 0 aardolie-
/steenkoolverwerkende industrie
0 machine-industrie 0 instrument- of optische industrie
0 hout-, meubel- of papierindustrie 0 textiel-, kledingindustrie
0 metaal-basisindustrie 0 grafische industrie
0 anders, namelijk:
A2 Wat is uw functie?
0 inkoopdirecteur/hoofd inkoop 0 senior buyer
0 logistiek manager 0 inkoopmanager
0 inkoper/inkoopassistent 0 supply chain manager
0 anders, namelijk:
A3 Hebben uw inkoopactiviteiten betrekking op een business unit, divisie of op een

andere organisatorische eenheid? Bij alle komende vragen, kunt u uitgaan van deze
eenheid!
A4 Wat is het aantal medewerkers, in 2001 op basis van full time-aanstellingen (fte, excl.
uitzendkrachten)?
0 minder dan 100 medewerkers
0 100 – 200 medewerkers
0 201 – 500 medewerkers
0 501 – 1.000 medewerkers
0 meer dan 1.000 medewerkers
A5 Wat is de omzet, in miljoenen Euro’s, in 2001?

0 minder dan 5 miljoen Euro
0 5 tot 10 miljoen Euro
0 meer dan 500 miljoen Euro
A6 Wat was in 2001 het inkoopaandeel: waarde van ingekochte goederen en diensten t.o.v. de
omzetwaarde?
........%
A7 Beoordeel de volgende stellingen, op een schaal die loopt van 1 (volledig mee oneens)
tot 5 (volledig mee eens).
1. Binnen ons bedrijf heeft men in voldoende mate inzicht in de mogelijkheden van 1 2 3 4 5
en problemen met in te kopen producten.
2. In ons bedrijf werken wij in voldoende mate met gedifferentieerde 1 2 3 4 5

leveranciersstrategieën.
3. De directie vindt dat Inkoop een belangrijke bijdrage levert aan de 1 2 3 4 5

concurrentiepositie van het bedrijf.
4. Inkoop rapporteert direct aan de directie. 1 2 3 4 5
5. Inkopers in ons bedrijf beschikken over voldoende vaardigheden om in cross- 1 2 3 4 5

functionele teams te werken.
6. Inkopers in ons bedrijf beschikken over voldoende vaardigheden om inkoop- en 1 2 3 4 5

leveranciersstrategieën te ontwikkelen.
7. Ons inkoopbeleid is vooral gericht op samenwerking met leveranciers. 1 2 3 4 5
8. Ons inkoopbeleid is vooral gericht op scherp onderhandelen met leveranciers 1 2 3 4 5

om de laagste prijzen te realiseren.
9. Inkopers binnen ons bedrijf houden zich vooral bezig met administratieve 1 2 3 4 5
werkzaamheden en het oplossen van dagelijkse problemen met leveranciers.
340
Onderdeel B Inkoop- en
leveranciersstrategieën
Het doel van dit onderdeel van de vragenlijst is om te weten welke omstandigheden in
de praktijk leiden tot de keuze van bepaalde inkoop- en leveranciersstrategieën. Daartoe
vragen wij aan u om zich in te leven in specifieke inkoopsituaties die we telkens eerst
zullen schetsen. Iedere situatie beschrijft omstandigheden en een bepaalde oplossing, een
strategie die wordt gekozen. Aan u de vraag om een vergelijkbare, typerende situatie in
herinnering te brengen, zoals die zich binnen uw bedrijf heeft voorgedaan. Vervolgens
kunt u vragen beantwoorden die betrekking hebben op uw organisatie (als koper) en de
leverancier (als verkoper) van het betreffende product. In totaal gaat het om 9 situaties,
waarover telkens dezelfde serie vragen wordt gesteld.
Situatie (1)
Het gaat om een product met een relatief lage inkoopwaarde, maar met een hoog
inkooprisico. Uw organisatie is daarmee kwetsbaar voor wat betreft de toelevering van
één leverancier X. Getracht wordt de toelevering zeker te stellen door relatief grote
voorraden aan te houden. Probeert u zich een vergelijkbare situatie voor te stellen, zoals
u die in uw eigen bedrijf heeft meegemaakt.
Beantwoord nu de volgende vragen, op een schaal die loopt van 1 tot 5.
in in
beperkte hoge
mate mate
1. In hoeverre herkent u de beschreven situatie in uw praktijk? 1 2 3 4 5
B1 Beantwoord de volgende vragen, op een schaal die loopt van 1 Volledig Volledig
(volledig mee oneens) tot 5 (volledig mee eens). mee mee
oneens eens
2. U bent afhankelijk van leverancier X. 1 2 3 4 5
3. Betrouwbare levertijden van dit product zijn belangrijk voor een ongestoorde 1 2 3 4 5
voortgang van productieprocessen.
4. U heeft die technologische expertise nodig waarover 1 2 3 4 5

leverancier X beschikt.
5. Er zijn andere leveranciers die het product ook kunnen leveren. 1 2 3 4 5

6. U moet hoge kosten maken, als u leverancier X wilt vervangen. 1 2 3 4 5
7. Leverancier X is afhankelijk van u. 1 2 3 4 5
8. U bent, wat omzet betreft, een belangrijke klant voor 1 2 3 4 5

leverancier X.
9. Leverancier X heeft uw technologische expertise nodig. 1 2 3 4 5
10 Er zijn andere kopers waar leverancier X zijn producten 1 2 3 4 5

. kan verkopen.
11 Leverancier X moet hoge kosten maken als hij u wilt vervangen door een andere 1 2 3 4 5
. koper.
12 De specificaties van het product zijn in hoge mate toegesneden op de wensen 1 2 3 4 5

. en eisen binnen ons bedrijf.
13 Het is mogelijk om voor vergelijkbare artikelen standaard-specificaties te gaan 1 2 3 4 5

. hanteren, door te werken met minder specifieke eisen.
14 Wij kunnen erop vertrouwen dat leverancier X gemaakte beloften en afspraken 1 2 3 4 5

. daadwerkelijk nakomt.
15 Wij kunnen erop vertrouwen dat leverancier X geen misbruik maakt van zijn 1 2 3 4 5
. positie en daadwerkelijk rekening houdt met onze belangen.
16 We doen zaken met leverancier X, voornamelijk omdat we goed en plezierig 1 2 3 4 5

. kunnen samenwerken.
17 We doen zaken met leverancier X, voornamelijk omdat het te veel tijd, energie 1 2 3 4 5
. en geld zou kosten om de relatie te beëindigen.
Situatie (2)
Het gaat om een product met een relatief lage inkoopwaarde, maar met een hoog
inkooprisico. Uw organisatie is daarmee kwetsbaar voor wat betreft de toelevering van
leverancier X. Als antwoord op dit probleem wordt nu gekozen voor het gaan zoeken
naar andere oplossingen, met name door te werken met meer gangbare specificaties
(minder complex) en zo nodig met een andere leverancier.
Probeert u zich een vergelijkbare situatie voor te stellen, zoals u die in uw eigen bedrijf
heeft meegemaakt.
342
in in
beperkte hoge
mate mate
B2 Beantwoord de volgende vragen, op een schaal die loopt van Volledig Volledig
1 (volledig mee oneens) tot 5 (volledig mee eens). mee mee
oneens eens
3. Betrouwbare levertijden van dit product zijn belangrijk voor een 1 2 3 4 5

ongestoorde voortgang van productieprocessen.
4. U heeft die technologische expertise nodig waarover leverancier 1 2 3 4 5

X beschikt.
8. U bent, wat omzet betreft, een belangrijke klant voor leverancier X 1 2 3 4 5

.
10. Er zijn andere kopers waar leverancier X zijn producten kan verkopen. 1 2 3 4 5
11. Leverancier X moet hoge kosten maken als hij u wilt vervangen door een 1 2 3 4 5
andere koper.
12. De specificaties van het product zijn in hoge mate toegesneden op de 1 2 3 4 5

wensen en eisen binnen ons bedrijf.
13. Het is mogelijk om voor vergelijkbare artikelen standaard-specificaties te 1 2 3 4 5

gaan hanteren, door te werken met minder specifieke eisen.
14. Wij kunnen erop vertrouwen dat leverancier X gemaakte beloften en 1 2 3 4 5

afspraken daadwerkelijk nakomt.
15. Wij kunnen erop vertrouwen dat leverancier X geen misbruik maakt van zijn 1 2 3 4 5
positie en daadwerkelijk rekening houdt met onze belangen.
16. We doen zaken met leverancier X, voornamelijk omdat we goed en 1 2 3 4 5

plezierig kunnen samenwerken.
17. We doen zaken met leverancier X, voornamelijk omdat het te veel tijd, 1 2 3 4 5
energie en geld zou kosten om de relatie te beëindigen.
Situatie (3)
Hier gaat het om een relatief goedkoop product met een relatief laag inkooprisico. Het
product is daarom niet erg belangrijk voor de organisatie, maar moet toch worden
ingekocht. Het product wordt ingekocht bij leverancier X. Als benadering van deze
situatie wordt ervoor gekozen om het product in een pakket onder te gaan brengen met
vergelijkbare producten. Door het bundelen van producten kunnen inkoopbehoeften
worden geconcentreerd bij een enkele leverancier.
heeft meegemaakt.
Beantwoord de volgende vragen op basis van de situatie voordat deze bundeling heeft
plaatsgevonden.
in in
beperkte hoge
mate mate
1 (volledig mee oneens) tot 5 (volledig mee eens). mee Mee
oneens eens
4. U heeft die technologische expertise nodig waarover leverancier X beschikt. 1 2 3 4 5
8. U bent, wat omzet betreft, een belangrijke klant voor leverancier X. 1 2 3 4 5
andere koper.

344


16. We doen zaken met leverancier X, voornamelijk omdat we goed en plezierig 1 2 3 4 5

kunnen samenwerken.
Situatie (4)
Hier gaat het om een relatief goedkoop product met een relatief laag inkooprisico. Het
product is daarom niet erg belangrijk voor de organisatie, maar moet toch worden
ingekocht. Het product wordt afzonderlijk besteld, iedere keer als dat nodig is. Thans bij
leverancier X.
heeft meegemaakt.
in in
beperkte hoge
mate mate
oneens eens

andere koper.




kunnen samenwerken.
Situatie (5)
Met dit product bevindt u zich in een gunstige positie: het inkooprisico is laag, terwijl het
product een relatief groot bedrag vertegenwoordigt. U koopt het product thans bij
leverancier X. Er wordt scherp onderhandeld, teneinde de beste condities binnen te
halen. Kopen tegen de laagste prijs met behoud van kwaliteit en leveringszekerheid krijgt
prioriteit. Brede concurrentiestelling (‘competitive bidding’) behoort tot de
mogelijkheden. Contracten worden alleen aangegaan voor de korte termijn.
heeft meegemaakt.
346
in in
beperkte hoge
mate mate
oneens eens
andere koper.




kunnen samenwerken.
Situatie (6)
Met dit product bevindt u zich in een gunstige positie: het inkooprisico is laag, terwijl het
product een relatief groot bedrag vertegenwoordigt. U koopt het product thans bij
leverancier X. U ziet echter mogelijkheden om de relatie te verdiepen, zodat leverancier X
in de toekomst meer zal gaan bijdragen aan de concurrentiepositie van uw organisatie.
Van leverancier X wordt verwacht dat hij zich als partner gaat gedragen en diens
technologische expertise in dienst stelt van de koper, uw organisatie.
heeft meegemaakt.
Beantwoord de vragen op basis van de situatie voordat de samenwerking met leverancier
X is gestart.
in in
beperkte hoge
mate mate
oneens eens
andere koper.
348
12. De specificaties van het product zijn in hoge mate toegesneden op de wensen 1 2 3 4 5
en eisen binnen ons bedrijf.
13. Het is mogelijk om voor vergelijkbare artikelen standaard-specificaties te gaan 1 2 3 4 5

hanteren, door te werken met minder specifieke eisen.
14. Wij kunnen erop vertrouwen dat leverancier X gemaakte beloften en afspraken 1 2 3 4 5
daadwerkelijk nakomt.

kunnen samenwerken.
17. We doen zaken met leverancier X, voornamelijk omdat het te veel tijd, energie 1 2 3 4 5
en geld zou kosten om de relatie te beëindigen.
Situatie (7)
Dit product vertegenwoordigt een bijzondere positie: het inkooprisico is hoog, evenals de
financiële waarde van het product. Het gaat dus om een belangrijk product. U beschouwt
leverancier X als een belangrijke partner, waarmee een waardevolle relatie van
strategische samenwerking bestaat. Beide partijen hebben een zakelijk belang bij het
instandhouden van de relatie. De verhoudingen zijn goed en de prestaties van
leverancier X worden als uitstekend beoordeeld.
heeft meegemaakt.
in in
beperkte hoge
mate mate
oneens eens
11. Leverancier X moet hoge kosten maken als hij u wilt vervangen door een andere 1 2 3 4 5
koper.


kunnen samenwerken.
Situatie (8)
financiële waarde van het product. Het gaat dus om een belangrijk product. De
verhoudingen met leverancier X laten echter te wensen over. Uw organisatie (de koper)
was destijds in hoge mate gedwongen om bij leverancier X te kopen. De organisatie kan
thans niets anders doen, dan zich neerleggen bij de feiten en de onvrijwillige relatie met
leverancier X zo goed mogelijk houden.
heeft meegemaakt.
350
in in
beperkte hoge
mate mate
oneens eens
andere koper.


kunnen samenwerken.
Situatie (9)
financiële waarde van het product. Het gaat dus om een belangrijk product. Van
leverancier X wordt verwacht dat hij zich als partner opstelt. De samenwerking verloopt
echter niet naar wens en het blijkt niet mogelijk om leverancier X bij te sturen. In
antwoord op deze situatie wordt besloten om te gaan zoeken naar een andere leverancier,
waarmee vervolgens een relatie moet worden opgebouwd. Het is duidelijk dat dit geen
eenvoudige opgave is.
heeft meegemaakt.
Beantwoord de vragen op basis van de situatie voordat er naar een andere leverancier
wordt gezocht.
in in
beperkte hoge
mate mate
oneens eens
andere koper.
352


kunnen samenwerken.
Onderdeel C Gebruik van de

inkoopportfolio analyse
Bedrijven hebben al gauw te maken met 100’en leveranciers die 1.000’en artikelen
leveren. Als hulpmiddel wordt daarom wel gebruik gemaakt van de inkoopportfolio-
analyse, ontwikkeld door Kraljic. De benadering van Kraljic komt erop neer dat alle
producten in een 2x2 matrix worden geplaatst en dat voor iedere categorie een
bijpassende aanbeveling wordt gegeven.
C1 Hoe vaak wordt de inkoopportfolio analyse binnen uw organisatie toegepast?

0 nooit
0 minder dan 1x per jaar
0 ongeveer 1x per jaar
0 meer dan 1x per jaar
Als u hier ‘nooit’ aankruist, dan hoeft u in dit onderdeel alleen vraag C2 te
beantwoorden.
Als u een ander antwoord heeft aangekruist, dan hoeft u vraag C2 juist niet te
beantwoorden.
C2 Indien ‘nooit’, waarom wordt de portfolio analyse niet toegepast? (meer antwoorden
mogelijk)
0 omdat het te veel tijd kost
0 omdat de uitkomsten niet bruikbaar zijn
0 omdat er onvoldoende kennis aanwezig is binnen de organisatie
0 omdat het hoofd inkoop/de inkoopdirecteur de portfolio analyse niet wil gebruiken
0 omdat te weinig factoren worden meegenomen in de analyse
0 omdat complexe, strategische inkoopbeslissingen niet kunnen worden gebaseerd op de
te eenvoudige portfolio analyse
0 omdat, .............
C3 Indien de inkoopportfolio analyse wel wordt toegepast, wat zijn de belangrijkste

redenen?
(meer antwoorden mogelijk)
0 de analyse levert een belangrijke bijdrage aan het inzicht in producten en leveranciers
0 de analyse levert een belangrijke bijdrage aan het realiseren van inkoopbesparingen
0 de analyse identificeert eventuele problemen en mogelijkheden op inkoopgebied
354
0 de analyse is een belangrijke bron van informatie voor het ontwikkelen van
gedifferentieerde inkoop- en leveranciersstrategieën
0 anders, namelijk ..............................
C4 Indien de inkoopportfolio analyse wel wordt toegepast, bent u dan tevreden over dit
hulpmiddel?
0 zeer tevreden
0 tevreden
0 niet tevreden/niet ontevreden
0 ontevreden
0 zeer ontevreden
C5 Indien de inkoopportfolio analyse wel wordt toegepast, op welke manier worden de posities
van artikelen in de matrix bepaald?
0 met een gewogen factor score methode, waarbij de score op een variabele (dimensie) wordt
bepaald door deelscores op factoren te vermenigvuldigen met gewichten en deze bij elkaar te
tellen
0 op basis van overleg waarin consensus wordt gezocht naar de meest passende plaats in de
matrix
0 met een matrix waarbij op de ene as ‘de inkoopwaarde’ staat en op de andere as ‘het aantal
leveranciers’, zodat alle waarden objectief en eenduidig kunnen worden bepaald
0 op een andere manier, namelijk ................................................................. ................. .....................
We danken u hartelijk voor uw medewerking!
0 ja, ik wil een samenvatting van de onderzoeksresultaten

0 ja, ik wil het boek ‘Ondernemend Samenwerken: ontwikkelen van concurrentiekracht in
netwerken’
Sturen naar: Naam:

Adres:
Postcode/plaats:
Appendix H
Measurement and operationalization

of organizational dependence
source perspective: measurement of operationalization

dependence of dependence
a(n)
El-Ansary and Stern wholesaler indirect 3 items on a 5-point scale:

(1972) and dealer - percentage of a channel
wholesaler’s and member’s business
dealer’s perceptions - commitment to channel member
- difficulty and cost in replacing
a channel member
Etgar (1976) insurance indirect 4 measures of agent’s

agent dependence:
agent’s perceptions - number of insurers that the agent
represents
- percentage of agent’s premiums
with his leading insurer
- difficulty in replacing the insurer
- reliance on casualty income
Brown, Lusch and retailer indirect 5 items:

Muehling (1983) - number of brands
retailer’s perceptions - percentage of store’s sales
and estimates derived from the major brand
- percentage of store’s profit
derived from the major brand
- percentage of customers lost, if
the major supplier was changed
- difficulty in replacing the major
supplier (7-point Likert scale)
Frazier (1983) automobile direct, proxy measure 6 measures of role performance

dealer of the manufacturer on an
dealer’s perceptions 11-point scale
Heide and John (1988) manufactu- indirect 6 items on a 7-point scale

rers’ agency measuring consequences of
buyer’s perceptions the loss of a manufacturer:
- hurting sales of related products
- switching to other lines
- replacement with a similar line
from another firm
356

a(n)
- selling new products

- significant loss of income
- other principals would like to
have us as their agent
Provan and Skinner dealer indirect 3 variables:

(1989) - dependence for critical services
dealer’s perceptions - number of suppliers
- availability of alternate suppliers
Frazier, Gill and Kale dealer indirect 4 items:

(1989) - current level of sales from the
dealer perceptions manufacturer’s products
- current level of profit
- anticipated level of sales
- anticipated level of profit
Bourantas (1989) buyer direct, proxy measure: relative magnitude as an estimate

index of dependence = for resource importance =
‘resource importance’ purchases from a supplier divided
x by total purchases
‘source substitutability’
substitutability =
1/(n + 1)
with:
n = number of alternative sources
Kamath and Liker supplier direct, proxy measure proportion of the supplier’s dollar
(1990) revenues from business with the
supplier’s perceptions specific OEM in question
Noordewier, John and buyer direct 5-point scale, from

Nevin (1990) and ‘totally dependent’ to
supplier buyer’s perceptions ‘not at all dependent’
Hallén, Johanson and customer and indirect supplier’s dependence:

Seyed-Mohamed supplier - customer importance, estimated
(1991) perceptions and
estimates by supplier’s sales
- buyer concentration, estimated
by supplier’s sales to its three
largest customers
customer’s dependence:
- supplier importance, estimated
by the customer’s purchases
- supplier’s market share
- product complexity

a(n)
Sriram, Krapfel and buyer indirect multiple items on a 5-point scale,

Spekman (1992) relating to:
buyer’s perceptions - availability of other suppliers,
- importance of the supplier,
- buyer’s switching costs,
- availability of alternative buyers
for the supplier’s products,
- buyer’s ability to manufacture
the procured component
Buchanan (1992) store direct 1 item for mutual dependence:

(buyer) - ease/difficulty for both to replace
buyer’s perceptions one another
Handfield (1993) buyer direct, proxy measure number of alternate suppliers, the
firm was currently using in
obtaining the critical resource
Provan and dealer direct, proxy measure percentage of total annual

Gassenheimer (1994) purchases of a product provided
by the primary supplier
Gassenheimer and dealer indirect 11 items on a 7-point Likert scale,

Ramsey (1994) referring to:
dealer’s perceptions - logistical support services
- sales support services
- product support services
Heide (1994) buyer and indirect 4 items on a 7-point Likert scale:

supplier - easily replace the volume
buyer and supplier - alternative buyers/suppliers
dependence on the - easily adapt to a new supplier/
same scale format new buyer
- redesign effort when dealing with
buyer’s perceptions a new supplier/ impact on
supplier’s operations when dealing
with new buyers
Ganesan (1994) retailer and indirect 2 measures:

vendor - percentage of volume in this
buyer’s perceptions category is accounted for by this
vendor/this retailer
- a number of items focusing on
the extent to which the vendor/the
retailer is important to the other
Berger, Noorderhaven supplier and direct 5-point Likert-type scales

and Nooteboom (1995) buyer
supplier’s perceptions
358

a(n)
Lusch and Brown wholesaler indirect items on a Likert scale

(1996) and major
supplier wholesaler’s wholesaler’s dependence:
perceptions - we are dependent on our major
supplier
- our major supplier would be
difficult to replace
- our major supplier would be
costly to replace
supplier’s dependence:
- our supplier is dependent on us
- our major supplier would find it
difficult to replace us
- our major supplier would find it
costly to lose us
Geyskens, Steenkamp, dealer and indirect items on a 7-point Likert scale

Scheer and Kumar supplier
(1996) dealer’s perceptions dealer dependence:
- expectations on profit and sales,
accounted for by this supplier
- other suppliers with comparable
product lines
supplier dependence:
- relationship with us is very
important to the supplier
- other firms who could provide the
supplier with comparable
distribution
Young and Wilkinson firm indirect 2 items on a 6-point Likert-scale:

(1997) - ease of replacing the other firm
seller’s and buyer’s - impact of being let down by the
perceptions other firm
Joshi and Arnold (1997) buyer indirect a high dependence and a low
dependence scenario, based on:
buyer’s perceptions - number of alternative suppliers
- switching costs
Dant and Gundlach franchisee indirect 4 items on a 5-point Likert scale:

(1998) - easily replace franchisor
franchisee’s - loss in income, if the relationship
perceptions with franchisor was terminated
- difficult to generate income from
other sources
- easily adapt to selling a different
product line

a(n)
Miles, Preece and technology indirect dependence on alliances;

Beatz based firm 5 items on a 5-point Likert scale:
(1999) manager’s perceptions - take advantage of opportunities
without an alliance partner
- need a partner for customers
- need a partner for investors
- need a partner to achieve critical
mass in production
- need to develop strategic
alliances to reach potential
Johnson (1999) distributor indirect 3 items on a 7-point Likert scale:

- easily switch sales to other lines
distributor’s - easily replace supplier’s product
perceptions line with a similar line from
another company
- significant loss in income, if the
relationship was terminated
De Jong and supplier indirect supplier’s dependence:

Nooteboom (2000) 4 items on a 5-point Likert scale:
supplier’s perceptions - lost in investments
- manufacturing complexity
- average piece price
- location specificity
customer’s dependence:
2 items on a 6-point scale:
- number of months to replace
the supplier
- supplier’s share in customer’s
product
Kim (2001) distributor and indirect distributor’s dependence:

supplier 4 items on a 7-point Likert scale:
distributor’s - effort to find a good alternative
perceptions and - effort to compensate for the loss
supplier’s perceptions by switching to other lines
- effort do develop a profitable
relationship with another supplier
- effort to diversify into selling new
products, dropping the supplier
supplier’s dependence:
4 items on a 7-point Likert scale:
- effort to find a good alternative
- effort to compensate for the loss
360

a(n)
by switching to other distributors

- effort do develop a profitable
relationship with another
distributor
- effort to use our own salesforce
to
sell this product line
Buvik and Holskau buyer indirect supplier’s dependence:

(2001) 2 items on a 7-point Likert scale:
buyer’s perceptions - the difficulty for the supplier,
replacing our company
- economic problems for the
supplier, should the sales to our
company cease
Handfield and Bechtel buyer indirect buyer’s dependence:

(2002) 3 items on a 5-point scale:
buyer’s perceptions - supplier is only source of input
- key-input material is not
produced/available in the U.S.
- many suppliers
Summary
In the last 30 years the role of purchasing has changed dramatically in many companies,
from a clerical, administrative function into a strategic function that contributes to the
competitive advantage. This ‘revolution in purchasing’ has lead to tremendous changes
in the scope, the impact and the responsibilities of purchasing management (e.g. Van
Weele and Rozemeijer, 1996). In accordance with an increased focus on the management
of supplier relationships, there is a growing acceptance and use of purchasing portfolio
approaches, aimed at developing appropriate and differentiated purchasing and supplier
strategies. However, hardly any empirical research has been conducted into this subject,
which is why we initiated this study.
Chapter 1 clarifies the background of the research project, starting with an overview of
recent developments in purchasing management. Major developments all point at the
importance of the management of supplier relationships. Companies need a variety of
relationships, where no general ‘best’ type of relationship exists (e.g. Young and
Wilkinson, 1997; Gadde and Snehota, 2000). For long, the ABC-analysis was the only tool
for differentiating between important and less important purchases. However, the ABC-
analysis is limited to a single dimension (the financial value of items), and it does not
provide differentiated strategies. In a seminal paper Kraljic (1983) introduced the first
comprehensive portfolio approach for the use in purchasing and supply management.
His approach includes the construction of a 2x2 matrix that classifies products in four
categories (bottleneck, non-critical, leverage and strategic items) on the basis of two
dimensions: profit impact and supply risk (‘low’ and ‘high’). Each of the four categories
requires a distinctive approach towards suppliers. Non-critical items require efficient
processing, leverage items allow the buying company to exploit its full purchasing
power, for instance through tendering. Bottleneck items cause significant problems and
risks which should be handled by volume insurance, vendor control, security of
inventories, and backup plans. Three general strategies (diversify, balance, or exploit) are
recommended for items in the strategic quadrant, according to the relative power
position of the company in the corresponding supply markets.
362
In the course of time purchasing portfolio models, and especially the Kraljic matrix, have
gained ground in purchasing practice. However, in contrast to the increased adoption,
there is a lack of (academic) research into the actual practice, the possibilities, and the
theoretical foundations of purchasing portfolio models. This research project has
addressed this gap, in its pursuit of gaining a better understanding of:
- how they could be used in order to pursue differentiated purchasing strategies.
The research project includes three successive research methods: literature study, case
studies, and a survey. The literature study covers three main areas: portfolio models in
purchasing management, portfolio models in related disciplines, and a discussion on
power and dependence in buyer-supplier relationships. We have started the literature
study with a review of portfolio models in related business areas, namely investment
theory, strategic management, and marketing management (chapter 2). The main reason
for starting with other areas than purchasing management was that we wanted to learn
from disciplines with a longer tradition and experience in the use of such models.
Chapter 3 provides a review of the main portfolio approaches in purchasing and supply
management. The models are discussed and evaluated on their (1) dimensions, (2)
categories, (3) strategic recommendations, and (4) use issues (acceptance and adoption).
The analysis of literature made clear that there are more similarities than differences
between the various models in purchasing. Most models use similar dimensions to those
of the Kraljic model, they use the same categories and they provide identical
recommendations. Based upon our literature analysis we have concluded that the Kraljic
matrix is the dominant approach in the profession. In addition, in chapter 3 we identified
and discussed the main problems and critique of the Kraljic matrix, referring to the
measurement of variables, the alleged disregard for the supplier’s side, the selection of
strategies based on two dimensions, the limited and deterministic character of the
strategic recommendations, and the absence of explicit movements within the matrix. The
findings of the literature study have been used as input for the case studies, to investigate
how experienced practitioners handle these issues in practice.
In a recent interview Kraljic explained that he, as a consultant, was asked to develop a
new tool for purchasing, similar to the portfolio approaches that were introduced in
marketing and strategic management, e.g. BCG matrix. Kraljic acknowledged that the
selection of dimensions was based on discussions with purchasing professionals, in
search for ‘things that really matter in purchasing’. Basically, a matter of common sense
(Gelderman and Van Haaster, 2002). The Kraljic approach does not explicitly deal with
issues of power and dependence, nor does it provide any reference to a theoretical
foundation or comprehensive perspective. However, there are indications that power and
dependence are important in the Kraljic approach, considering some of the
recommendations and the general idea of the portfolio approach: “to minimize supply
vulnerability and make the most of potential buying power” (Kraljic, 1983, 112).
Chapter 4 therefore elaborates on issues of power and dependence in buyer-supplier

relationships with a special focus of relevance on the Kraljic approach. We have analyzed
the dimensions and the categories of the Kraljic matrix, and its strategic
recommendations. We have argued that the resource dependence theory (Pfeffer and
Salancik, 1978) should be considered as the (implicitly applied) theoretical foundation for
the Kraljic portfolio approach.
In addition, we have concluded that a comprehensive view of the dyadic nature of buyer-
supplier relationships should include the assessment of (1) the difference between
buyer’s and supplier’s dependence (net dependence) which corresponds with the relative
power between parties, and (2) the sum of buyer’s and supplier’s dependence (total
interdependence) which indicates the mutual dependence and the intensity of the
relationship between parties. These concepts have been applied to the Kraljic matrix,
which
364
resulted in the following propositions (to be tested by the survey data):

- strategic: balanced power, highest interdependence
- non-critical: balanced power, lowest interdependence (independent)
- leverage: buyer dominated, moderate interdependence
- bottleneck: supplier dominated, moderate interdependence.
Chapter 5 reports on the results of the case studies, which are used to identify and
describe advanced practices with respect to purchasing portfolio models. The critique of
portfolio models, such as the Kraljic approach, does not include the experience of
practitioners. Our explorative case studies addressed this gap. The sample consisted of a
selection of Dutch companies, restricted to manufacturing companies, where purchasing
is by nature an important business area. Three case companies were selected on their
advanced and ongoing use of purchasing portfolio analysis. The case studies revealed
three distinctive methods: the consensus method (1), the one-by-one method (2), and the
weighted factor score method (3). Each method satisfies the needs and expectations of the
different users. The reason for this was found in the additional steps that have to be taken
in the portfolio analysis. Before strategic actions are determined, it is imperative to
complete a further process of interpreting and reflecting on the results. The positioning of
items in the matrix should be considered as the starting point of portfolio analysis,
definitely not the finishing point. In-depth discussions on the positions in the matrix are
considered as the most important phase of the analysis. It is felt by the users that the
Kraljic framework facilitates strategic discussions to a large extent.
based on just two basic dimensions (e.g. Dubois and Pedersen, 2002: 40)? Actually, the
answer is simple: one cannot. In addition to the various factors that constitute the two
dimensions of any matrix, it was found that experienced portfolio users always included
additional information on:
- the specific situations on supply markets, and
- capacities, intentions and competences of individual suppliers.
Comparative analysis of the case studies resulted in a conceptual model of strategic

directions, providing an overview of the main strategic choices for the categories in the
matrix. A dichotomy has been identified between strategies to hold a position (1) and
strategies to move to another position (2) in the matrix. At the right side of the matrix (in
the bottleneck and the strategic areas) movements are pursued in order to reduce a high
level of supply risk. In terms of the matrix, this means moving to the left. Non-critical
items are preferably moved upwards and exceptional leverage positions are exchanged
for strategic positions. These are the most common movements within the matrix.
- partners of convenience, located in the leverage quadrant, which is dominated by the
buyer;
dependence;
- locked-in ‘partnerships’, located at the right side of the strategic quadrant, which is
dominated by suppliers, who are indispensable for the buyer.
Chapter 6 explains the design of the survey method and it reports on the main findings.
The survey method was aimed at measuring the variables and relationships in the
conceptual models:
- the variables to discriminate and explain the differences between users and non-users of
a purchasing portfolio approach,
- the conditions for the selection of portfolio based strategies,
- the power and dependence structure in the Kraljic matrix, and
- the determinants of buyer’s and supplier’s dependence in the Kraljic matrix.
The last three items required the development of scenarios which describe a number of
situations in terms of the Kraljic dimensions (profit impact and supply risk) and in terms
of the selection of a corresponding purchasing strategy. These scenarios are based on the
findings of the case studies. The design of the study can be characterized as a repeated
measures design, because the same respondent participates in all conditions of the
experiment. The levels of the conditions (scenarios) describe the within-subjects (W-S)
variable. Respondents were asked to evaluate the scenarios on multiple dependent
variables, which are associated with determinants of buyer’s and supplier’s dependence,
the nature of specifications, trust and commitment. Each scenario can be characterized by
a specific combination of these factors, constituting a unique relationship-dependence
profile.
The survey procedure included a pilot study aimed at enhancing the reliability and the
validity of the questionnaire. The final questionnaire was administered (in three rounds)
to 1,153 individuals, mostly purchasing managers of manufacturing companies who are
member of the Dutch Association of Purchasing Management (NEVI). Answering more
than 175 questions was likely to produce non-response problems. However, the special
366
measures turned out to be rather effective, resulting in a satisfactory response rate of

20.6% (n=238). The survey resulted in facts and figures about the actual utilization of the
purchasing portfolio analysis. Taking differences in company size into account, we found
a weighted average use frequency of almost 60% for manufacturing companies in The
Netherlands.
With respect to the explanation of the use, the most important findings are:
- Portfolio analysis delivers what it is supposed to: it provides additional understanding
of problems and possibilities of purchasing and it provides assistance in the process of
developing differentiated purchasing strategies.
- Users contrast in a positive way with non-users of the portfolio, especially on their
purchasing’s professionalism (skills) and their contributions to the competitive position
of the company. In addition it was found that the portfolio was relatively more used by
larger companies with higher purchasing shares.
With respect to the conditions-research question, statistical tests revealed the significant
differences between the relationship-dependence profiles of the scenarios in our search
for conditions under which the various purchasing strategies are selected. Estimated
Marginal Means (EMMEANS) was used for the multiple pair-wise comparisons of
profiles, within the SPSS-procedure GLM-repeated measures. In addition we have
explained the buyer’s dependence and the supplier’s dependence from a limited number
of explanatory variables (‘determinants of dependence’). We have established different
conditional relationships in the categories of the Kraljic matrix for most of the
explanatory variables. For instance, we found a positive relationship between the
logistical indispensability of a product and the buyer’s dependence in cases of high
supply risk (strategic and bottleneck items on the right side of the matrix), but no
relationship in cases of low supply risk (leverage and non-critical items on the left side of
the matrix).
Finally, we have analyzed the issue of power-and-dependence in the portfolio matrix. We

have calculated interdependencies and the power positions of the buyer in the Kraljic
categories and we have compared them with prior, theoretical expectations. The observed
levels of total interdependence were in accordance with prior expectations: ‘highest’ in
the strategic quadrant, ‘moderate’ in the bottleneck and leverage quadrant, and ‘lowest’
in the non-critical quadrant. The observed buyer dominance in the non-critical quadrant
seemed to be caused by slightly lower switching costs for the buyer and by slightly more
alternative trading partners, both in comparison with the supplier. We found an expected
supplier dominance in the bottleneck quadrant, and an expected buyer dominance in the
leverage
quadrant. Most remarkably however, was the observed overall supplier dominance in the
strategic quadrant, which may shed a different light on the buyer’s view on issues of
power and dependence: even satisfactory partnerships are considered to be dominated
by the supplier.
Chapter 7 summarizes the main conclusions and recommendations of this study,

including a 5-step approach to the application of purchasing portfolio analysis,
accompanied by a number of practical guidelines for assessing and improving positions
in a portfolio matrix, conform to the principle of ‘conditional dynamics’: switching to
another position in the matrix depends on the conditions that make it desirable and
feasible. A close connection appears with the problems of how to manage power and
dependence in supplier relations. The chapter is completed by a critical reflection on the
research project and recommendations for further research. Future research could
address the immediate and long-term impact of the application of the portfolio approach,
overcoming the difficulties of attributing results to portfolio use. Another challenge
would be the development of a portfolio approach from a network perspective, in
addition to the well-known dyadic perspective. We have broken a lance for quantitative
and qualitative research to the importance, impact and determinants of power relations
in chains of interdependent companies (supply chain). Finally, we have recommended
research to the relationship between e-procurement and the purchasing portfolio
approach.
368
Samenvatting (summary in Dutch)
In de afgelopen 30 jaar is de rol van inkoop drastisch veranderd in veel bedrijven, van
een administratieve naar een strategische bedrijfsfunctie die bijdraagt aan de
concurrentiepositie van bedrijven. Deze ‘inkooprevolutie’ heeft geleid tot enorme
veranderingen in het belang, de invloed en de verantwoordelijkheden van
inkoopmanagement (Van Weele en Rozemeijer, 1996). In aansluiting op de toegenomen
aandacht voor leveranciersrelaties, constateren we steeds meer acceptatie en gebruik van
portfoliobenaderingen binnen het inkoopmanagement, gericht op het ontwikkelen van
passende en gedifferentieerde inkoop- en leveranciersstrategieën. Echter, tegelijkertijd
moeten we constateren dat er nauwelijks onderzoek is gedaan naar dit onderwerp,
hetgeen ons ertoe heeft gebracht om dit onderzoek te beginnen.
Hoofdstuk 1 verklaart de achtergronden van dit onderzoeksproject, beginnend met een

overzicht van recente ontwikkelingen in inkoopmanagement. De belangrijkste
ontwikkelingen benadrukken het belang van het managen van leveranciersrelaties.
Bedrijven hebben behoefte aan verschillende soorten relaties, zonder dat er sprake is van
‘een beste’ type relatie (zie bijvoorbeeld Young en Wilkinson, 1997; Gadde en Snehota,
2000). Heel lang was de ABC-analyse het enige instrument om onderscheid te maken
tussen belangrijke en onbelangrijke aankopen. Echter, de ABC-analyse beperkt zich tot de
financiële waarde van producten én het verschaft geen gedifferentieerde
inkoopstrategieën. In een baanbrekend artikel introduceerde Kraljic (1983) de eerste,
omvattende portfoliobenadering voor inkoop- en leveranciersmanagement. Deze
benadering omvat de constructie van een matrix waarin producten worden ingedeeld in
vier categorieën (knelpunt-, routine-, hefboom- en strategische producten) op basis van
twee dimensies: invloed op de winst en inkooprisico (‘laag’ en ‘hoog’). Elke categorie
vraagt een eigen benadering van leveranciers. Routineproducten vereisen vooral
efficiënte afhandeling en bestelroutines, met hefboomproducten kan de onderneming zijn
inkoopmacht gebruiken, bijvoorbeeld brede concurrentiestelling. Knelpuntproducten
veroorzaken ernstige risico’s en problemen, waarvoor het inkoopbeleid gericht moet zijn
op het zekerstellen van levering, extra controle van de leverancier, bewaking van
voorraden en dergelijke. Voor de strategische producten worden drie generieke
benaderingen aanbevolen (nastreven van diversificatie, een evenwichtige relatie of
exploitatie), al naar gelang de relatieve machtspositie van de onderneming in de
bijbehorende leveranciersmarkten.
Kraljic's categorieën en strategische aanbevelingen
In de loop der tijd hebben portfoliomodellen steeds meer erkenning en toepassing

gekregen in de inkooppraktijk, en dan met name de Kraljic matrix. Ondanks het
toenemende gebruik, hebben we een gebrek aan (wetenschappelijk) onderzoek
geconstateerd naar het feitelijk gebruik, de mogelijkheden en de theoretische
fundamenten van deze modellen. Daarom is dit onderzoeksproject gericht op het
verwerven van meer begrip van en kennis over:
- de theoretische en conceptuele grondslagen van portfoliomodellen in de inkoop,
- het feitelijk gebruik van portfoliomodellen in de inkooppraktijk, en
- de manier waarop deze modellen kunnen worden gebruikt voor het ontwikkelen van
gedifferentieerde inkoopstrategieën.
Het onderzoeksproject omvat drie achtereenvolgende onderzoeksmethoden: een

literatuuronderzoek, een aantal case studies en een enquête (survey). Het
literatuuronderzoek beslaat drie hoofdgebieden: portfoliomodellen in inkoop,
portfoliomodellen in verwante vakgebieden en macht-en-afhankelijkheid in inkoop-
leveranciersrelaties. We zijn begonnen met de bespreking van portfoliomodellen in
verwante gebieden, te weten investeringstheorie, strategisch management en marketing
management (hoofdstuk 2). We zijn met andere gebieden dan inkoopmanagement
begonnen, omdat we wilden leren van vakgebieden met een langere traditie en meer
ervaring in het gebruik van portfoliomodellen. Hoofdstuk 3 bespreekt de belangrijkste
portfoliobenaderingen in inkoopmanagement. De modellen zijn besproken en
geëvalueerd op hun (1) dimensies, (2) categorieën, (3) strategische aanbevelingen en (4)
370
hun acceptatie en adoptie (gebruik). De literatuuranalyse maakt duidelijk dat er meer

overeenkomsten zijn dan verschillen tussen de modellen. Ze gebruiken vrijwel dezelfde
dimensies als de Kraljic matrix, ze hebben dezelfde categorieën en geven vergelijkbare
aanbevelingen. Op basis van het literatuuronderzoek concluderen we dat de Kraljic
matrix de dominante benadering is binnen inkoop. Ook hebben we in hoofdstuk 3 de
belangrijkste problemen van en kritiek op de Kraljic matrix geïnventariseerd en
besproken: het meten van variabelen, de vermeende gebrekkige aandacht voor de
leverancierszijde, de selectie van strategieën gebaseerd op twee dimensies, het beperkte
en deterministische karakter van de strategische aanbevelingen en de afwezigheid van
expliciete bewegingen binnen de matrix. De bevindingen van de literatuurstudie zijn
gebruikt voor de opzet van de case studies, waarin is onderzocht hoe ervaren
inkoopprofessionals met deze problemen omgaan in de praktijk.
In een recent interview verklaarde Kraljic dat hij, als adviseur, was gevraagd om een
nieuw instrument te ontwikkelen voor inkoopmanagement, vergelijkbaar met de destijds
net ontwikkelde marketing matrices, zoals de BCG matrix. Kraljic bevestigde dat de
selectie van dimensies was gebaseerd op discussies met inkoopprofessionals, op zoek
naar ‘zaken die echt belangrijk zijn voor inkoop’. Kortom, een kwestie van gezond
verstand (Gelderman en Van Haaster, 2002). De Kraljic benadering is niet expliciet over
zaken als macht en afhankelijkheid, evenmin wordt verwezen naar theoretische
grondslagen of een overkoepelend perspectief. Desondanks zijn er aanwijzingen dat
‘macht’ en ‘afhankelijkheid’ belangrijke begrippen zijn in de Kraljic benadering. Dit blijkt
onder meer uit enkele aanbevelingen en het algemene idee van de portfoliobenadering
waarin verwezen wordt naar “het minimaliseren van kwetsbaarheid in de bevoorrading
en zo goed mogelijk gebruik maken van potentiële inkoopmacht” (Kraljic, 1983, 112).
Derhalve zijn we in hoofdstuk 4 nader ingegaan op kwesties van macht en
afhankelijkheid in de relaties tussen inkopers en leveranciers in het algemeen en de
betekenis voor de Kraljic benadering in het bijzonder. We hebben de dimensies en de
categorieën van de Kraljic matrix nader bestudeerd, alsmede de strategische
aanbevelingen. We hebben beredeneerd dat de resource dependence theory (Pfeffer en
Salancik, 1978) beschouwd moet worden als de (impliciet gekozen) theoretische
grondslagen van de Kraljic portfoliobenadering.
Vervolgens hebben we geconcludeerd dat een omvattend perspectief op het tweezijdige
karakter van inkoop-leveranciersrelaties inhoudt, dat we moeten vaststellen (1) het
verschil tussen de afhankelijkheid van de inkoper en de afhankelijkheid van de
leverancier hetgeen overeenkomt met de relatieve machtsverhouding tussen partijen en
(2) de som van de afhankelijkheid van de inkoper en de afhankelijkheid van de
leverancier (interdependentie) hetgeen aangeeft de mate van wederzijdse afhankelijkheid
en de intensiteit van de relatie tussen partijen. Deze concepten hebben we toegepast op
de Kraljic matrix, hetgeen geleid heeft tot de volgende veronderstellingen ten aanzien
van de categorieën in de matrix:
- strategisch: evenwichtige machtsverhouding met een hoge wederzijdse
afhankelijkheid
- routine: evenwichtige machtsverhouding met een lage wederzijdse
afhankelijkheid
- hefbooom: inkoopzijde is dominant met een gematigde wederzijdse
afhankelijkheid
- knelpunt: leverancierszijde is dominant met een gematigde wederzijdse
afhankelijkheid.
Deze veronderstellingen zijn getoetst met behulp van de survey data.
Hoofdstuk 5 doet verslag van de resultaten van de exploratieve case studies, waarin
geavanceerde toepassingen van de inkoopportfoliomodellen zijn beschreven. De kritiek
op portfoliomodellen, zoals met name op de Kraljic benadering, houdt geen rekening met
de ervaringen uit de praktijk. De case studies zijn met name gericht om dit hiaat te
vullen. De steekproef bestond uit een selectie van Nederlandse bedrijven uit de industrie,
waar inkoop van nature een belangrijke bedrijfsfunctie is. Drie bedrijven zijn
geselecteerd op hun geavanceerde en continue toepassing van de inkoopportfolio
analyse. In de cases studies werden drie verschillende meetmethoden gevonden: de
consensus methode (1), de één-op-één methode (2) en de methode van gewogen
factorscores (3). Iedere methode komt tegemoet aan de wensen en verwachtingen van de
diverse gebruikers. Dit valt te verklaren uit de additionele stappen die moeten worden
ondernomen in iedere portfolio analyse. Voordat strategische acties kunnen worden
bepaald, is het noodzakelijk om de matrix te onderwerpen aan een nader proces van
interpretatie en reflectie. De posities in de matrix moeten worden beschouwd als
vertrekpunt van de analyse, zeker niet als het eindpunt. Grondige discussies over de
gevonden matrixposities worden beschouwd als het meest belangrijke onderdeel van de
portfolio analyse. De gebruikers zijn van mening dat het Kraljic raamwerk deze
belangrijke discussies in hoge mate faciliteert.
Sommigen beweren dat eenvoudige aanbevelingen niet mogelijk zijn voor complexe
beslissingen in het bedrijfsleven. Hoe kunnen strategieën worden afgeleid uit een
portfolio analyse die gebaseerd is op slechts twee dimensies (zie bijvoorbeeld Dubois en
Pedersen, 2002: 40)? Het antwoord is simpel: dat is niet mogelijk. In aanvulling op de
diverse factoren die worden gebruikt voor de twee dimensies, blijken ervaren portfolio
gebruikers altijd additionele informatie nodig te hebben over:
- de algemene bedrijfsstrategie (met name situaties op eindmarkten),
- de specifieke situaties op leveranciersmarkten en
- capaciteiten, intenties en competenties van individuele leveranciers.
372
Een vergelijkende analyse van de case studies heeft geresulteerd in een conceptueel
model dat een overzicht geeft van de belangrijkste strategische keuzes in alle categorieën
van de matrix. Een tweedeling is vastgesteld tussen strategieën waarmee een positie
wordt behouden (1) en strategieën waarmee een andere positie wordt nagestreefd (2).
Aan de rechterkant van de matrix (in de knelpunt- en strategische kwadranten) kan men
proberen het hoge inkooprisico te beperken, bijvoorbeeld door naar alternatieve
leveranciers te zoeken. Bezien vanuit de matrix betekent dit een beweging naar links.
Routineproducten kunnen wellicht worden gebundeld en naar boven verplaatst (richting
hefboom), terwijl hefboomposities slechts bij uitzondering worden ingeruild voor
strategische posities door de relatie te verdiepen tot strategische samenwerking (naar
rechts in de matrix). Vanuit inkoopperspectief werd een nieuwe indeling van
partnerships gevonden, gerelateerd aan de portfolio matrix:
- partners of convenience (gemaksleveranciers), te vinden in het hefboomkwadrant waar
relaties worden gedomineerd door de inkoopzijde;
- strategische partnerships, te vinden rond het midden tussen het hefboom- en het
strategische kwadrant, waar relaties worden gekenmerkt door een evenwichtige
machtsverhouding en hoge wederzijdse afhankelijkheid;
- locked in ‘partnerships’ (gedwongen ‘partnerships’), te vinden in de rechterhelft van
het strategische kwadrant, waar relaties worden gedomineerd door onmisbare
leveranciers.
Hoofdstuk 6 beschrijft het ontwerp van de survey en belangrijkste bevindingen. De

survey is gericht op het meten van variabelen en relaties in de conceptuele modellen:
- de variabelen die de verschillen aangeven tussen gebruikers en niet-gebruikers van de
inkoopportfolio-benadering,
- de condities voor het selecteren van inkoopstrategieën in de matrix;
- de machts- en afhankelijkheidsstructuur in de Kraljic matrix, en
- de determinanten van inkopersafhankelijkheid en leveranciersafhankelijkheid in de
Kraljic matrix.
De laatste drie onderdelen vereisten de ontwikkeling van scenario’s, waarin een aantal
situaties worden beschreven in termen van de Kraljic dimensies (invloed op de winst en
inkooprisico) en in termen van een gekozen inkoopstrategie. De scenario’s zijn gebaseerd
op de resultaten van de case studies. Het ontwerp van de studie is een repeated measures
design, omdat alle respondenten alle condities van het experiment ondergaan. De
niveaus van de condities (scenario’s) beschrijven de within-subjects (W-S) variabele. Aan
respondenten is gevraagd om de scenario’s te beoordelen aan de hand van een aantal
(afhankelijke) variabelen, gerelateerd aan de determinanten van de mate waarin inkopers
afhankelijk zijn van leveranciers, idem voor de mate waarin leveranciers afhankelijk zijn
van de inkopende partij, de aard van de specificaties, vertrouwen en commitment. Ieder

scenario kan worden gekenmerkt door een specifieke combinatie van deze factoren, dat
een uniek relatie-afhankelijkheidsprofiel oplevert.
Een vooronderzoek is gehouden met het oog op de betrouwbaarheid en de validiteit van

de vragenlijst. De uiteindelijke vragenlijst is (in drie ronden) aan 1.153 individuen
gestuurd, veelal inkoopmanagers van industriële bedrijven die lid zijn van de
Nederlandse Vereniging voor Inkoopmanagement (NEVI). Het beantwoorden van meer
dan 175 vragen leidt al gauw tot non-response problemen. Echter, de genomen
maatregelen bleken effectief te zijn, gezien de bevredigende response van 20,6% (n=238).
De enquête heeft onder meer geresulteerd in feiten met betrekking tot het gebruik van de
portfolio analyse. Als we rekening houden met verschillen in bedrijfsomvang, dan komen
we tot een gewogen gemiddeld gebruik van bijna 60% voor Nederlandse industriële
bedrijven.
Het literatuuronderzoek heeft onder meer geleid tot de identificatie van variabelen
waarmee de verschillen tussen gebruiken en niet-gebruikers kunnen worden beschreven
en verklaard: de bedrijfsomvang, het inkoopaandeel, de positie van inkoop in het bedrijf,
de inkoopprofessionaliteit en de oriëntatie van inkoop. De belangrijkste resultaten met
betrekking tot het verklaren van portfoliogebruik zijn:
- Portfolio analyse ‘doet wat het belooft’: het voorziet in meer begrip voor de problemen
en mogelijkheden van inkoop en het biedt ondersteuning voor het ontwikkelen van
gedifferentieerde inkoopstrategieën.
- Gebruikers onderscheiden zich op positieve wijze van niet-gebruikers, vooral gezien
hun inkoopprofessionaliteit (vaardigheden) en hun bijdragen aan de concurrentiepositie
van het bedrijf. Verder blijkt dat de portfolio relatief vaker wordt gebruikt door grotere
bedrijven en door bedrijven met relatief grotere inkoopaandelen.
Teneinde vast te stellen onder welke condities welke strategieën worden gekozen,
hebben we Estimated Marginal Means (EMMEANS) toegepast voor de paarsgewijze
vergelijking van de profielen, binnen de SPSS-procedure GLM-repeated measures.
Tevens hebben we de mate waarin inkopers afhankelijk zijn van leveranciers verklaard
uit een beperkt aantal verklarende variabelen. Hetzelfde hebben we gedaan voor de mate
waarin leveranciers afhankelijk zijn van de inkopende partij (‘determinanten van
afhankelijkheid’). Het onderzoek heeft geleid tot een aantal voorwaardelijk relaties in de
Kraljic categorieën, bijvoorbeeld tussen de logistieke onmisbaarheid van een product en
de mate waarin de inkoper afhankelijk is van de leverancier: een positieve relatie in geval
van hogere inkooprisico’s (aan de rechterzijde van de matrix) en geen relatie in geval van
lagere inkooprisico’s (aan de linkerzijde van de matrix).
374
Tot slot hebben een nadere analyse gemaakt van het vraagstuk van macht en
afhankelijkheid in de Kraljic matrix. Daarbij zijn berekeningen gemaakt voor de mate van
wederzijdse afhankelijkheid en de machtsverhoudingen in de Kraljic categorieën. De
uitkomsten hebben we vergeleken met eerdere, theoretische verwachtingen. De niveaus
van de wederzijdse afhankelijkheid waren in overeenstemming met de verwachtingen:
het hoogst in het strategische kwadrant, matig in het knelpunt- en het hefboomkwadrant
en het laagst in het routinekwadrant. De dominantie van de inkoper in het
routinekwadrant lijkt veroorzaakt te worden door lagere switching costs voor de inkoper
en bij meer alternatieve handelspartners, beide in vergelijking met de positie van de
leverancier. We vonden bevestiging voor de verwachte dominantie van de leverancier in
het knelpuntkwadrant. Meest opvallend echter was de waargenomen, algemene
dominantie van de leverancier in het strategisch kwadrant, vanuit inkoopperspectief.
Deze resultaten kunnen aanleiding geven om een ander licht te werpen op macht en
afhankelijkheid bezien vanuit inkoopperspectief: zelfs bevredigende partnerships
worden blijkbaar gedomineerd door de leverancier.
Hoofdstuk 7 geeft de belangrijkste conclusies en de aanbevelingen, inclusief een 5-

stappenplan voor de toepassing van de inkoopportfolio en een aantal praktische
aanbevelingen voor het beoordelen en verbeteren van matrixposities, volgens de
principes van ‘conditionele dynamiek’: een beweging naar een andere positie in de
matrix is afhankelijk van de condities en omstandigheden die dit mogelijk maken.
Daarbij geldt een nauwe relatie met het managen van macht en afhankelijkheid in
leveranciersrelaties. Het hoofdstuk wordt afgesloten met een kritische reflectie op het
onderzoeksproject en aanbevelingen voor nader onderzoek. Toekomstig onderzoek zou
zich kunnen richten op het onmiddellijke, maar ook op de lange termijn effecten van
portfoliogebruik, waarbij de onderzoeker zich moet bezighouden met vraagstuk van het
toerekenen van resultaten. Een andere uitdaging zou zijn het ontwikkelen van een
portfoliobenadering vanuit een netwerkperspectief, in aanvulling op het bekende
dyadische perspectief. Verder zouden we kwalitatief en kwantitatief onderzoek
verwelkomen naar het belang, de invloed en de determinanten van machtsverhoudingen
binnen ketens van bedrijven (supply chains). Tot slot hebben we aangegeven dat er
onderzoek zou moeten worden gedaan naar de relatie tussen e-procurement en de
inkoopportfolio.
About the author
Cees J. Gelderman (1956) studied Business Economics at the State University of

Groningen. During his study he was a teacher at a secondary school
(‘Ommelandercollege’ in Appingedam). He graduated in 1984 (cum laude) in the field of
Marketing and Marketing Research. After graduating he worked as an assistant professor
of marketing at the State University of Groningen until 1986. In the same year he joined
the Open University of the Netherlands as an assistant professor of marketing and
purchasing management. He developed several courses, all based on the principles of
distance education in various fields such as Purchasing management, Marketing
management, Marketing Research, Marketing Planning and Business Marketing. In the
nineties he published the book “Business marketing” which he co-authored with
professor Hein van der Hart (Eindhoven University of Technology). In collaboration with
Bé Albronda he also wrote and edited the book “Professioneel inkopen” (professional
purchasing). In 1998, he started the Ph.D. research project on the possibilities of a
portfolio approach to the development of differentiated purchasing and supplier
strategies. Cees Gelderman has written many papers for journals, magazines, handbooks
and international conferences. He continues to work at the open University of the
Netherlands, where he will combine research activities with the development of courses
in the areas of marketing and purchasing management.
View publication stats

Gelderman2003PhDthesis PDF

Uploaded by

Copyright:

Available Formats

Gelderman2003PhDthesis PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Gelderman2003PhDthesis PDF

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

A portfolio approach to the development of differentiated

Article · January 2003

The user has requested enhancement of the downloaded file.

Cornelis Johannes Gelderman

Eindhoven University of Technology

CIP-DATA LIBRARY TECHNISCHE UNIVERSITEIT EINDHOVEN

A portfolio approach to the development of

Key words: Purchasing/Purchasing strategy/Portfolio

Printed by: Universiteits Drukkerij/Technische

© 2003 C.J. Gelderman, Maastricht

All rights reserved. No part of this publication may be

ter verkrijging van de graad van doctor

Cornelis Johannes Gelderman

prof.dr. A.J. van Weele

Chapter 1 Introduction to the study 15

Chapter 2 Portfolio approaches in related disciplines 35

Chapter 3 Portfolio approaches in purchasing management 63

Chapter 4 A power-dependence perspective 107

Chapter 6 The survey 197

Chapter 7 Conclusions and recommendations 257

Samenvatting (in Dutch) 368

About the author 375

Writing a dissertation is practically impossible without a supportive home base to fall

1 Introduction to the study

For long, purchasing was considered as a relatively unimportant clerical function in

1.1 Developments in purchasing management

Nowadays, there is a growing consensus that purchasing is an important contributor to a

Managing supplier relations is increasingly considered as a strategically important

1.2 Need for differentiation

differentiation is needed in managing supplier relationships. Relationships require

1.3 The portfolio concept

In this study we define a portfolio model as:

1.4 From Pareto to Kraljic

Categories Values Main tasks/strategies

non-critical items low on both dimensions efficient processing

FIGURE 1.1 Kraljic's categories and strategic recommendations

The Kraljic portfolio approach can be considered as an important breakthrough in the

1.5 Statement of the problem

To conclude, in contrast with the increased adoption of purchasing portfolio models,

- the theoretical and conceptual foundations of purchasing portfolio models,

3) How are portfolio models employed by experienced purchasing professionals?

5) What is the role of power and dependence in the Kraljic approach?

1.6 Relevance of the stated problem

FIGURE 1.2 Overview of the successive research steps

(a) Literature study

(b) Case studies

1.8 Structure of the thesis

2.1 Investment theory: the roots of all portfolio

2.1.1 The Markowitz model

The Markowitz model is based on several assumptions regarding investor behavior:

Under these assumptions, a portfolio of assets is considered to be efficient if no other

2.1.2 Relevance for purchasing: what can we learn?

‘rewards’, albeit that the context of decision making is different in comparison to

2.2 Portfolio models in strategic management

2.2.1 The BCG-approach

FIGURE 2.1a An example of a growth-share matrix

FIGURE 2.1b Cach flows in a growth-share matrix

2.2.2 Assumptions, criticism and adoption