THE MEASUREMENT

OF SCIENTIFIC
AND TECHNOLOGICAL ACTIVITIES

PROPOSED GUIDELINES
FOR COLLECTING
AND INTERPRETING
INNOVATION DATA

OSLO MANUAL
Final draft of the third edition,
JULY 5, 2005

1
 TABLE OF CONTENTS

Chapter 1 OBJECTIVES AND SCOPE OF THE MANUAL .......................................................................5

1. INTRODUCTION ............................................................................................................................5
2. FACTORS INFLUENCING THE SCOPE OF THE MANUAL .....................................................8
2.1 What is measurable?....................................................................................................................9
2.2 What is of value to measure?.......................................................................................................9
3. SCOPE OF THE MANUAL.............................................................................................................9
3.1 Sector coverage .........................................................................................................................10
3.2 Innovation at the level of the firm .............................................................................................10
3.3 Types of innovations .................................................................................................................10
3.4 Diffusion and the degree of novelty ..........................................................................................11
4. PROVIDING DATA ON THE KEY ISSUES................................................................................12
4.1 Innovation activities and expenditures ......................................................................................12
4.2 Factors influencing innovation ..................................................................................................12
4.3 The innovating firm and the impact of innovation ....................................................................13
4.4 Linkages in the innovation process ...........................................................................................13
5. SOME SURVEY ISSUES ..............................................................................................................13
5.1 Approach to data collection.......................................................................................................13
6. THE RELATIONSHIP BETWEEN THE OSLO MANUAL AND OTHER INTERNATIONAL
STANDARDS AND RELATED CONCEPTS .........................................................................................14
6.1 Manuals for the measurement of science and technology activities..........................................14
6.2 Other economic norms and classifications ................................................................................16
6.3 Other related concepts and surveys ...........................................................................................16
7. FINAL REMARK...........................................................................................................................18
Chapter 2 INNOVATION THEORY AND MEASUREMENT NEEDS....................................................19
1. INTRODUCTION ..........................................................................................................................19
2. ECONOMICS OF INNOVATION.................................................................................................19
3. A MEASUREMENT FRAMEWORK ...........................................................................................23
4. SECTORAL AND REGIONAL ASPECTS OF INNOVATION...................................................27
4.1 Innovation in services................................................................................................................27
4.2 Innovation in low and medium technology industries...............................................................28
4.3 Innovation in small and medium sized enterprises....................................................................28
4.4 Regional innovation...................................................................................................................28
4.5 Globalisation..............................................................................................................................29
5. AREAS FOR INVESTIGATION ...................................................................................................29
5.1 What can be measured...............................................................................................................29
5.2 Inputs to innovation...................................................................................................................30
5.3 Linkages and the role of diffusion.............................................................................................30
5.4 The impact of innovation...........................................................................................................31
5.5 Incentives and obstacles to innovation ......................................................................................31
5.6 Demand .....................................................................................................................................31
5.7 Other ..........................................................................................................................................32
Chapter 3 BASIC DEFINITIONS ...............................................................................................................33
1. INTRODUCTION ..........................................................................................................................33
2. INNOVATION ...............................................................................................................................33
3. MAIN TYPES OF INNOVATION ................................................................................................34
4. DISTINGUISHING BETWEEN TYPES OF INNOVATIONS ....................................................38
4.1 Distinguishing between product and process innovations.........................................................38

2
 4.2 Distinguishing between product innovations and marketing innovations .................................39
4.3 Distinguishing between service (product) innovations and marketing innovations ..................39
4.4 Distinguishing between process and marketing innovations.....................................................39
4.5 Distinguishing between process and organisational innovations ..............................................40
4.6 Distinguishing between marketing and organisational innovations ..........................................40
5. CHANGES THAT ARE NOT CONSIDERED INNOVATIONS .................................................41
6. NOVELTY AND DIFFUSION ......................................................................................................42
7. THE INNOVATIVE FIRM ............................................................................................................43
8. COLLECTING DATA ON INNOVATIONS ................................................................................44
Chapter 4 INSTITUTIONAL CLASSIFICATIONS ...................................................................................46
1. THE APPROACH...........................................................................................................................46
2. THE UNITS ....................................................................................................................................46
2.1 The primary statistical unit ........................................................................................................47
2.2 The secondary statistical unit ....................................................................................................49
3. CLASSIFICATION BY MAIN ECONOMIC ACTIVITY ............................................................50
4. CLASSIFICATION BY SIZE ........................................................................................................52
5. OTHER CLASSIFICATIONS........................................................................................................52
5.1 Type of institution .....................................................................................................................52
5.2 Other ..........................................................................................................................................53
Chapter 5 LINKAGES IN THE INNOVATION PROCESS ......................................................................54
1. INTRODUCTION ..........................................................................................................................54
2. INBOUND DIFFUSION ................................................................................................................55
2.1 Types of linkages.......................................................................................................................56
2.2 Collecting data on linkages in the innovation process...............................................................59
2.3 Other linkages indicators ...........................................................................................................61
3. OUTBOUND DIFFUSION ............................................................................................................62
4. KNOWLEDGE MANAGEMENT .................................................................................................63
Chapter 6 MEASURING INNOVATION ACTIVITIES ............................................................................64
1. INTRODUCTION ..........................................................................................................................64
2. THE COMPONENTS AND COVERAGE OF INNOVATION ACTIVITIES .............................65
2.1 Research and experimental development ..................................................................................65
2.2 Activities for product and process innovations .........................................................................66
2.3 Activities for marketing and organisational innovations...........................................................68
2.4 Design........................................................................................................................................69
2.5 The borderline between R&D and non-R&D innovation activities ..........................................69
2.6 The development and use of software in innovation activities..................................................70
3. COLLECTING DATA ON INNOVATION ACTIVITIES............................................................71
3.1 Qualitative data on innovation activity......................................................................................72
3.2 Quantitative data on innovation activity....................................................................................73
3.3 Other measurement issues .........................................................................................................73
3.4 Breakdown by type of expenditure............................................................................................74
3.5 Breakdown by source of funds ..................................................................................................74
3.6 The subject approach versus the object approach......................................................................75
Chapter 7 OBJECTIVES, OBSTACLES AND OUTCOMES OF INNOVATION....................................77
1. INTRODUCTION ..........................................................................................................................77
2. OBJECTIVES AND EFFECTS OF INNOVATIONS ...................................................................77
3. OTHER MEASURES OF IMPACTS ON ENTERPRISE PERFORMANCE...............................80

3
 3.1 Impact on turnover ....................................................................................................................81
3.2 The impact of process innovations on costs and employment ..................................................82
3.3 The impact of innovation on productivity .................................................................................82
4. FACTORS HAMPERING INNOVATION ACTIVITIES.............................................................84
5. QUESTIONS ON THE APPROPRIABILITY OF INNOVATIONS ............................................86
Chapter 8 SURVEY PROCEDURES ..........................................................................................................88
1. INTRODUCTION ..........................................................................................................................88
2. POPULATIONS .............................................................................................................................88
2.1 The target population.................................................................................................................88
2.2 The frame population ................................................................................................................89
3. SURVEY METHODS ....................................................................................................................89
3.1 Mandatory or voluntary survey .................................................................................................89
3.2 Census or sample survey ...........................................................................................................89
3.3 Domains.....................................................................................................................................90
3.4 Sampling techniques..................................................................................................................90
3.5 Panel data surveys .....................................................................................................................91
3.6 Survey methods and suitable respondents .................................................................................92
3.7 The questionnaire ......................................................................................................................93
3.8 Innovation and R&D surveys ....................................................................................................93
4. ESTIMATION OF RESULTS........................................................................................................95
4.1 Weighting methods....................................................................................................................95
4.2 Non-response.............................................................................................................................95
5. PRESENTATION OF RESULTS...................................................................................................96
6. FREQUENCY OF DATA COLLECTION ....................................................................................97
ANNEX 1: INNOVATION SURVEYS IN DEVELOPING COUNTRIES.................................................98
1. INTRODUCTION ..........................................................................................................................98
2. THE CHARACTERISTICS OF INNOVATION IN DEVELOPING COUNTRIES ....................99
3. INNOVATION MEASUREMENT IN DEVELOPING COUNTRIES .......................................101
3.1 Specific needs for public policies and private strategies: the potentially innovative firms.....101
3.2. Measurement priorities............................................................................................................102
4. PRINCIPAL ADAPTATIONS .....................................................................................................104
4.1 ICTs in innovation surveys......................................................................................................104
4.2 Linkages ..................................................................................................................................105
4.3 Innovation Activities ...............................................................................................................105
4.4 Additional adaptations.............................................................................................................106
5. METHODOLOGICAL ISSUES FOR DEVELOPING COUNTRY CONTEXTS......................106
5.1 Information systems specificities ...........................................................................................106
5.2 General methodological considerations...................................................................................106
6. THINKING AHEAD ....................................................................................................................108
ANNEX 2. EXAMPLES OF INNOVATIONS ........................................................................................109
1. INTRODUCTION ........................................................................................................................109
2. EXAMPLES OF INNOVATIONS ...............................................................................................109

4
 Chapter 1

OBJECTIVES AND SCOPE OF THE MANUAL

1. INTRODUCTION

1. It is widely accepted that innovation is central to the growth of output and productivity. However,
our understanding of innovation activities and their economic impacts, while greatly increased since the
first edition of this Manual, is still deficient. For example, as the world economy evolves, so does the
process of innovation. Globalisation has led to dramatic increases in access to information and new
markets for firms, along with increased international competition and new organisational forms to manage
global supply chains. With advances in technologies and in the flow of information, knowledge is
increasingly seen as a central driver of economic growth and innovation. However, we lack a full
understanding of how these factors affect innovation.

2. In order to develop appropriate policies to support innovation, several critical aspects of the
innovation process need to be better understood, including innovation activities other than R&D,
interactions among actors, and the flows of relevant knowledge. Policy development also requires further
advances in the analysis of innovation, based in part on obtaining better information.

3. The first edition of this Manual, issued in 1992, and the surveys undertaken using it, including
the Community Innovation Survey (CIS) organised by the EU and comparable surveys in Australia and
Canada, showed that it is possible to develop and collect data on the complex and differentiated process of
innovation.

4. The second edition, issued in 1997, updated the original framework of concepts, definitions and
methodology to incorporate survey experience and improved understanding of the innovation process and
also to take in a wider range of industries. It provided improved guidelines for developing internationally
comparable innovation indicators for OECD countries, along with discussion of the analytical and policy
problems to which the indicators are relevant.

5. Both the first and second editions used the technological product and process (TPP) definition of
innovation. This reflected a focus on the technological development of new products and new production
techniques by firms, and their diffusion to other firms. However, discussion of organisational innovation
and non-technological innovation was included in an annex.

6. Since 1992, the number of countries that have conducted innovation surveys has grown
dramatically, including EU countries, other OECD countries such as Canada, Australia, New Zealand and
Japan, and a large number of non OECD countries, among them several Latin American countries, Russia
and South Africa. This has provided a large amount of data and experience with innovation surveys that
has been incorporated into this third edition of the Manual.

7. The third edition expands on the innovation measurement framework in three important ways.
First, greater emphasis is given to the role of linkages with other firms and institutions in the innovation
process. Second, it recognizes that innovation is important in less R&D intensive industries, such as
services and low technology manufacturing. While the second edition covered these sectors, the framework
was still primarily oriented towards manufacturing firms. This Manual modifies a number of aspects of the
framework (such as innovation definitions and activities) to better accommodate the services sector. Third,
the definition of innovation is expanded to include two additional types of innovations, organisational

5
innovation and marketing innovation. Also new to the Manual is an annex covering innovation surveys in
non-OECD countries, reflecting the fact that a growing number of these countries now conduct innovation
surveys.

8. The evaluation of linkages is expanded because of the important role of knowledge flows among
firms and other organisations in both the development and diffusion of innovations. This highlights the role
of organisational structures and practices that promote the sharing and use of knowledge and interactions
with other firms and public research institutions. This also includes forming closer relationships with
suppliers and an ongoing development of marketing practices to better reach customers. Linkages are now
included in a separate chapter that covers a variety of interactions ranging from arms length exchanges of
information to active involvement in joint innovation projects.

9. While services were covered in the second edition of the Oslo Manual, the focus was primarily
on manufacturing industries. Innovation activity in service-oriented sectors can differ substantially from
innovation in many manufacturing-oriented sectors. Innovation in services can often be less formally
organised, more incremental in nature and less technological. In order to form a framework that better
accommodates this broad range of industries, a number of modifications have been made to the
formulation of definitions, terms and concepts used in this Manual.

10. In order to identify the full range of changes that firms undertake to improve performance, and
their success in improving economic outcomes, a broader framework is needed than technological product
and process innovation. Including marketing and organisational innovations gives a more complete
framework that is better able to capture the changes that affect firm performance and contribute to the
accumulation of knowledge.

11. The role of organisational innovation is emphasized in a paper by Lam (2004): “Economists
assume that organisational change is a response to technical change, when in fact organisational innovation
could be a necessary precondition for technical innovation.”1 Organisational innovations are not only a
supporting factor for product and process innovations; they can have an important impact on firm
performance on their own. Organisational innovations can improve the quality and efficiency of work,
enhance the exchange of information, and improve the ability of the firm to learn and utilize new
knowledge and technologies.

12. Firms can also allocate large amounts of resources to marketing research and the development of
new marketing practices, such as targeting new markets or market segments and developing new ways of
promoting their products. New marketing practices can play a central role in firms’ performance.
Marketing practices are also important for the success of new products, and market research and contacts
with customers may play a key role in product and process development through demand-led innovation.
The inclusion of organisational and marketing innovation also permits greater analysis of the interactions
between different types of innovations, in particular the importance of implementing organisational
changes in order to benefit from other types of innovations.

13. Organisational innovations were also discussed in the second edition of the Manual, and there is
now some practical experience in data collection for organisational changes. This experience includes
specialised surveys on organisational innovation2 and the inclusion in innovation surveys of organisational

1
Lam, A. (2005), “Organizational Innovation”, Chapter 5 in J. Fagerberg, D. Mowery, and R.R. Nelson
(eds.), The Oxford Handbook of Innovation, Oxford University Press.
2
See Wengel, J. et al. (2000), “Analysis of Empirical Surveys on Organisational Innovation and Lessons for
Future Community Innovation Surveys”, Scientific Follow-up of the Community Innovation Survey (CIS)
Project CIS 98/191.

6
innovation (e.g. Australian Innovation Survey 2003) or questions on organisational changes (the CIS3
survey, the Japanese National Innovation Survey 2003, and others). This type of data has also been used in
empirical analysis, for example of the relationship between organisational innovation, ICT investment and
productivity3.

14. Marketing innovations are a new addition to the Oslo Manual. While there is less experience with
marketing innovation than organisational innovation, questions on marketing changes have been included
in a number of innovation surveys4, along with substantial field testing of marketing concepts in a number
of countries.

15. In order to gain a full picture of innovation activity, it is necessary to include marketing methods
in this framework, and there are at least two reasons for including them as a separate category, as opposed
to integrating marketing methods with organisational or process innovations.

16. First, marketing innovations can be important for firm performance and an important part of the
overall innovation process. Identifying marketing innovations allows both the analysis of their impact and
of their interaction with other types of innovations.

17. Second, a defining characteristic of marketing innovations is an orientation towards customers

and markets, with a focus on improving sales and market share. These economic objectives can differ
greatly from those of process innovations, which will likely focus on productive quality and efficiency.
Including marketing innovations with organisational innovations would also be problematic both due to the
fact that some marketing practices do not fit into conceptions of organisational changes and because this
would greatly dilute data on organisational innovations, making it difficult to interpret the results.

18. The definitions and concepts in this Manual have been adapted, based on previous survey
experience in Australia and other countries, to include organisational and marketing innovations for use in
innovation surveys. As with additions to previous editions of the Oslo Manual, the definitions of these
types of innovations are still under development and are not as well-established as the definitions of
product and process innovation.

19. Innovation surveys can have different goals and scope concerning which aspects of innovation to
cover and at what level of detail. Furthermore, in expanding coverage from product and process innovation
to also include marketing and organisational innovation, surveys can follow one of a number of
approaches. Surveys could provide equal coverage of all innovation types, partly cover marketing and
organisational innovation but maintain product and process innovations as the core innovation types, or
focus exclusively on product and process innovation. This Manual provides guidelines that can be of use
for all these approaches. In addition, one or more types of innovations can be covered in greater detail
through specialised surveys.

3
See, for example, Brynjolfsson, E. and L.M. Hitt (2000),”Beyond Computation: Information Technology,
Organizational Transformation and Business Performance”, Journal of Economic Perspectives 14 (4),
pp. 23-48 or OECD (2004), The Economic Impact of ICT: Measurement, Evidence and Implications, Paris,
OECD.
4
Experience covers the inclusion in innovation surveys of questions on marketing changes (the CIS3 survey,
the Japanese National Innovation Survey 2003 and others). In addition, the Intangible Assets Survey 2004
in France included all 4 types of innovations (i.e. marketing, organisational, product and process) along
with a number of other questions on marketing practices. The CIS4 survey, which was in progress at the
time of publication of this Manual, also included all 4 types of innovations.

7
20. The addition of organisational and marketing innovation, along with the use of a broad definition
of innovation that includes activities to both develop and adopt innovations, means that an increasing
percentage of firms are likely to meet the basic requirements to be defined as ‘innovative’. This requires
methods for identifying different types of innovative firms, based on the types of innovations that they
have implemented, and on their innovative capabilities and activities. It is not enough to know whether
firms are innovative or not; there is a need to know how firms innovate and what types of innovations they
implement.

21. With these points in mind, the aim of the present chapter is to give an overview of the coverage
and contents of the Manual (see Box 1).

Box 1. Structure of the Manual

The Manual starts with a general discussion of points that are likely to have some effect on the choice of indicators
(Chapter 2):

− An adequate conceptual understanding of the structure and characteristics of the innovation process and its
implications for policymaking.
− The key unresolved problems which further data could clarify

It continues with a framework for innovation surveys:

− Basic definitions of innovation, innovation activity and the innovative firm (Chapter 3).
− Institutional classifications (Chapter 4).

After that, suggestions and recommendations are advanced for the types of issues to be covered in national and
international innovation surveys:

− Measuring linkages in the innovation process; types of knowledge and their sources (Chapter 5).
− Innovation activities and their measurement (Chapter 6).
− Objectives, barriers and impacts of innovation (Chapter 7).

The Manual also contains two Annexes:

− Innovation surveys in developing economies (Annex 1).

− A detailed list of innovation examples (Annex 2).

2. FACTORS INFLUENCING THE SCOPE OF THE MANUAL

22. The purpose of this Manual is to provide guidelines for the collection and interpretation of
innovation data. The intended uses of innovation data are many and this Manual is designed to best
accommodate these various uses. One purpose for the collection of innovation data is to better understand
innovation and its relation to economic growth. This requires both knowledge of innovation activities that
have a direct impact on firm performance (for example through improvements in demand or reductions in
costs), and on the factors that affect their ability to innovate. Another purpose is to provide indicators for
benchmarking national performance. This both informs policymaking and allows international comparison.
There is a need for the collection of new indicators and at the same time a desire to maintain existing
indicators for comparison over time. This Manual is designed to achieve a balance between these different
needs.

23. How can one decide on the appropriate scope, structure, terminology and so on for internationally
comparable data collection? The variety of subjects that both specialised and general innovation surveys
have covered is evidence that many types of data are potentially available. Obviously, a survey covering

8
all possible topics would be so cumbersome as to be quite impracticable. That means identifying priorities,
and selecting the topics, industries and survey approaches on which to concentrate. There are two main
issues: what is measurable and what is of value to measure?

2.1 What is measurable?

24. Innovation is a continuous process. Firms can continually make changes to products, processes
and collect new knowledge, etc. It is more difficult to measure a dynamic process than a static activity.
With the objective of capturing this process, the Manual presents guidelines for collecting data on the
general process of innovation (for example, innovation activities, expenditures, and linkages), the
implementation of significant changes in the firm (i.e. innovations), the factors that influence innovation
activity, and the outcomes of innovation.

2.2 What is of value to measure?

25. In constructing innovation indicators, the information needs of policymakers and analysts are a
paramount consideration. Chapter 2 reviews these needs, which are part of the broad information system
that helps to reduce uncertainty in policymaking, and which has been influenced, since the first edition of
this Manual, by developments in the economics of innovation.

26. Innovation policy has developed as an amalgam of science and technology policy and industrial
policy. It recognizes that knowledge in all its forms plays a crucial role in economic progress, and that
innovation is a complex and systemic phenomenon. Systems approaches to innovation shift the focus of
policy towards an emphasis on the interplays between institutions, looking at interactive processes both in
the creation of knowledge and in its diffusion and application. The term “National Innovation System” has
been coined for this set of institutions and flows of knowledge. This theoretical perspective influences the
choice of questions to include in an innovation survey, such as the extensive coverage of linkages and
knowledge sources.

27. Systems approaches complement theories that focus on the innovative firm, objectives for
innovating and the activities that firms undertake. The forces that drive innovation at the level of the firm
and those innovations that are successful in improving firm performance are of central importance for
policymaking. Relevant data include questions on the implementation of innovations, the interaction of
different types of innovations, and on the objectives and barriers to innovation.

3. SCOPE OF THE MANUAL

28. For reasons summarised below:

• The Manual covers innovation in the business enterprise sector only.

• It deals with innovation at the level of the firm.

• It covers four types of innovations: product, process, organisational and marketing.

• It covers diffusion up to “new to the firm”.

9
3.1 Sector coverage

29. Innovation can occur in any sector of the economy, including government services such as health
or education. The guidelines in this Manual, however, are essentially designed to deal only with
innovations in the business enterprise sector. This includes manufacturing, primary industries, and the
service sector.

30. Innovation is also important for the public sector. However, less is known about innovation
processes in non-market oriented sectors. Much work remains in examining innovation and developing a
framework for innovation data collection in the public sector5. If undertaken, this work could form the
basis for a separate manual.

3.2 Innovation at the level of the firm

31. This Manual concerns the collection of innovation data at the level of the individual firm. It does
not cover industry or economy-wide changes that were discussed by Schumpeter, such as the emergence of
a new market, the development of a new source of supply of raw materials or semi-manufactured goods, or
the re-organisation of an industry. However, in some cases it is possible to construct estimates of industry
or economy wide changes, such as the emergence of a new market or industry re-organisation, by
aggregating data for individual firms.

32. For the purposes of the first three chapters of this Manual the generic term “firm” is used. It is
given a specific statistical definition in Chapter 4, dealing with classifications. Exactly what definition is
used in a study or survey may have an impact on the results. In the case of multinational corporations their
subsidiaries may be organised in different ways. A multinational firm may introduce a given innovation
country by country, market by market, or simultaneously throughout the group.

3.3 Types of innovations

33. A firm can make many types of changes in its methods of work, its use of factors of production
and its types of output which improve its productivity and/or commercial performance. This Manual
defines four types of innovations that encompass a wide range of changes in firms’ activities: product
innovations, process innovations, organisational innovations, and marketing innovations.

34. Full definitions of the four types of innovations can be found in Chapter 3. Product innovations
involve significant changes in goods or services’ capabilities. They include both entirely new goods and
services and significant improvements to existing products. Process innovations are significant changes in
production and delivery methods.

35. Organisational innovations refer to the implementation of new organisational methods. These can
be changes in business practices, in workplace organisation or in the firm’s external relations. Marketing
innovations involve the implementation of new marketing methods. These can include changes in product
design and packaging, in product promotion and placement, and in methods for pricing goods and services.

5
An example of early work in this area is Earl, L. (2003), “Innovation and Change in the Public Sector: a
Seeming Oxymoron”, Statistics Canada SIEID Working Paper Series No. 2002-01.

10
36. An important concern in expanding the definition of innovation is maintaining continuity with
the earlier definition of technological product and process (TPP) innovation6. However, the expansion of
the Manual to also cover the service sectors requires a few minor modifications of the definitions of
product and process innovation to better reflect innovation activities in the service sectors and to reduce
orientation to manufacturing. The revised definitions of product and process innovations (see Chapter 3)
are still very comparable to the TPP definition.

37. One change is the removal of the word ‘technological’ from the definitions. The use of the word
‘technological’ raises the concern that many service sector firms might interpret ‘technological’ as ‘using
high technology plant and equipment’, and thus not applying to many of their own product and process
innovations.

3.4 Diffusion and the degree of novelty

38. This Manual deals with changes that involve a significant degree of novelty for the firm. It
excludes changes that are minor, or do not involve a sufficient degree of novelty. At the same time, an
innovation does not need to be developed by the firm itself, but can be acquired from other firms or
institutions through the process of diffusion.

39. Diffusion is the way in which innovations spread, through market or non-market channels, from
their first worldwide implementation to different consumers, countries, regions, sectors, markets, and
firms. Without diffusion, an innovation will have no economic impact. The minimum entry for a change
in a firm’s products or functions to be considered as an innovation is that it must be new (or significantly
improved) to the firm. Three other concepts for the novelty of innovations are also discussed in Chapter 3:
new to the market, new to the world, and disruptive innovations.

40. There are two main reasons for using “new to the firm” as the minimum definition of an
innovation. First, the adoption of innovations is important for the innovation system as a whole. It involves
a flow of knowledge to adopting firms. Furthermore, the learning process in adopting an innovation can
lead to subsequent improvements in the innovation and to the development of new products, processes and
other innovations in the future. Second, the main impact of innovation on economic activity stems from the
diffusion of initial innovations to other firms. This diffusion is captured by covering innovations that are
new to the firm.

41. Of note, this Manual does not cover the diffusion of a new technology to other divisions or parts
of a firm after its first adoption or commercialisation7. For example, the first implementation of new
production technology in one of five factories owned by the same firm is counted as an innovation, but the
implementation of the same technology in the remaining four factories is not.

6
“Technological product and process (TPP) innovations comprise implemented technologically new
products and processes and significant technological improvements in products and processes. A TPP
innovation has been implemented if it has been introduced on the market (product innovation) or used
within a production process (process innovation).” (OECD/Eurostat (1997) Oslo Manual, § 130).
7
An exception would be an establishment-level survey, which could collect data on the first introduction of
an innovation into each establishment.

11
4. PROVIDING DATA ON THE KEY ISSUES

4.1 Innovation activities and expenditures

42. Innovation activities include all scientific, technological, organisational, financial and
commercial steps which actually, or are intended to, lead to the implementation of innovations. Some of
these activities may be innovative in their own right, while others are not novel but are necessary for
implementation.

43. Innovation comprises a number of activities not included in R&D, such as later phases of
development for preproduction, production and distribution, development activities with a lesser degree of
novelty, support activities such as training and market preparations, and development and implementation
activities for non product and process innovations such as new marketing methods or new organisational
methods. Innovation activities may also include acquisitions of external knowledge or capital goods that
are not part of R&D. A detailed breakdown of innovation activities and their descriptions, along with
measurement issues, is in Chapter 6.

44. During a given period, the innovation activities of a firm may be of three kinds:

• Successful in having resulted in the implementation of a new innovation (though they need not
have been commercially successful).

• Ongoing, work in progress, which has not yet resulted in the implementation of an innovation.

• Abandoned before the implementation of an innovation.

45. Expenditures are measured on the basis of the sum of these three kinds of activity over a given
period of time (see Chapter 6). An alternative would be to collect information on total expenditures on
activities related to individual innovations. Firms have had difficulties in reporting a full set of data
whichever approach has been used, however this is an essential set of data for the purposes of economic
and policy analysis. It is hoped that, with successive exercises, firms will find their own interest in costing
their innovation activities.

4.2 Factors influencing innovation

46. Enterprises can engage in innovation activity for a number of reasons. These objectives can relate
to products, markets, efficiency, quality or the ability to learn and to implement changes. Identifying
enterprises’ motives for innovating and their importance aids in examining the forces that drive innovation
activity, such as competition, and opportunities for entering new markets.

47. Innovation activity can be hampered by a number of factors. There can be reasons for not starting
innovation activities at all, or factors that slow innovation activity or have a negative effect on expected
results. These include economic factors, such as high costs or lack of demand, enterprise factors such as
lack of skilled personnel or knowledge, and legal factors such as regulations or tax rules.

48. The ability of enterprises to appropriate the gains from their innovation activities is also a factor
affecting innovation. If, for example, enterprises are unable to protect their innovations from copying by
their competitors, they will have less incentive to innovate. On the other hand, if an industry functions well
without formal protection methods, promoting these can slow the flow of knowledge and technology and
lead to higher prices for goods and services.

12
4.3 The innovating firm and the impact of innovation

49. The innovative firm (defined in Chapter 3) is one that has introduced an innovation during the
period under review. The innovations need not have been a commercial success – many innovations fail.
Innovative firms can be divided into those that have mainly developed innovations on their own or in
co-operation with other firms or public research organisations, or have mainly innovated by adopting
innovations (for example new equipment) developed by other firms. Innovative firms can also be
distinguished by the types of innovations they have implemented, for example whether they have
implemented a new product or process, or whether they have implemented a new marketing method or
organisational change.

50. Impacts of innovations on firm performance range from effects on sales and market share to
changes in productivity and efficiency. Important impacts at industry and national levels are changes in
international competitiveness and in total factor productivity, knowledge spillovers of firm-level
innovations, and an increase in the amount of knowledge flowing through networks.

51. The outcomes of product innovations can be measured by the percentage of sales derived from
new or improved products (as described in Chapter 7). Similar approaches can be used to measure the
outcomes of other types of innovations. Additional indicators of the outcomes of innovation can be
obtained through qualitative questions on the effects of innovations.

4.4 Linkages in the innovation process

52. The innovative activities of a firm partly depend on the variety and structure of its links to
sources of information, knowledge, technologies, practices, and human and financial resources. Each
linkage connects the innovating firm to other actors in the innovation system: government laboratories,
universities, policy departments, regulators, competitors, suppliers, and customers. Innovation surveys can
obtain information on the prevalence and importance of different types of linkages, plus the factors that
influence the use of specific linkages.

53. Three types of external linkages are identified. Open information sources provide openly
available information that does not require the purchase of technology or intellectual property rights, or
interaction with the source. Acquisition of knowledge and technology are purchases of external knowledge
and capital goods (machinery, equipment, software) and services embodied with new knowledge or
technology that do not involve interaction with the source. Innovation cooperation is active cooperation
with other firms or public research institutions for innovation activities (which may include purchases of
knowledge and technology).

5. SOME SURVEY ISSUES

5.1 Approach to data collection

5.1.1 Choice of the survey approach

54. There are two main approaches to collecting data on innovations:

i) The “subject” based approach starts from the innovative behaviour and activities of the firm as a
whole. The idea is to explore the factors influencing the innovative behaviour of the firm

13
 (strategies, incentives and barriers to innovation) and the scope of various innovation activities, and
above all to examine the outputs and effects of innovation. These surveys are designed to be
representative of all industries so the results can be grossed up and comparisons made between
industries.

ii) The “object” approach involves the collection of data about specific innovations (usually a
‘significant innovation’ of some kind, or the main innovation of a firm). The approach involves
collecting some descriptive, quantitative and qualitative data about the particular innovation at the
same time as data is sought about the firm.

55. From the point of view of current economic development, it is the differential success of firms
which shapes economic outcomes and which is of policy significance. This favours a subject-based
approach, although innovation surveys can combine both approaches by including general questions on the
firm and specific questions on a single innovation. It is the subject, the firm, that counts and the subject
approach has been chosen as the basis for these guidelines.

5.1.2 Survey methods

56. To reach international comparability for subject-based innovation surveys, it is important to

harmonise survey methods. Guidelines are given in Chapter 8.

57. The target population for innovation surveys concerns statistical units (innovators and
non-innovators, R&D performers and non-R&D performers) in the business enterprise sector. This
includes both goods producing and service sectors. Innovative activities take place in small and
medium-sized units as well as in large units. In order to capture innovation activity in these smaller units,
the target population should include, at a minimum, all statistical units with at least 10 employees. In the
case of sample surveys, the sample frames should correspond as closely as possible to the target
population.

58. In order to achieve a satisfactory response rate, the questionnaire should be as short as possible
and should include clearly formulated questions and instructions. This may involve expressing the formal
definitions in Chapter 3 in ways which are appropriate and meaningful to respondents in the industry
concerned.

59. In the data collection phase, particular attention should be paid to checking the reliability and
consistency of data and to the follow-up reminder procedures. International comparability of the resulting
data will be further improved by uniform methods of imputing missing values, weighting factors, methods
for presenting results, etc.

6. THE RELATIONSHIP BETWEEN THE OSLO MANUAL AND OTHER

INTERNATIONAL STANDARDS AND RELATED CONCEPTS

6.1 Manuals for the measurement of science and technology activities

60. There are two basic families of S&T indicators which are directly relevant to the measurement of
innovation: resources devoted to R&D and patent statistics. In addition, bibliometrics and several other
types of indicators provide complementary information, although they are not always available at the firm
level.

14
61. R&D data are collected through national surveys according to the guidelines laid down in the
Frascati Manual (OECD, 2002). These data have proved valuable in many studies: for example, the
effects of R&D on productivity have been estimated by econometric techniques, at the country, sector and
firm levels. These data have two main limitations. First, R&D is an input. Although it is obviously
related to technical change, it does not measure it. Second, R&D does not encompass all the efforts of
firms and governments in this area, as there are other sources of technical change, such as
learning-by-doing, which are not covered by this narrow definition.

62. A patent is a legal property right over an invention, which is granted by national patent offices.
A patent gives its owner sole rights (with limited duration) to exploit the patented invention, while at the
same time disclosing the details of the patent (which is intended to allow a broader social use of the
discovery). Patent statistics are increasingly used in various ways as indicators of the output of research
activities. The number of patents granted to a given firm or country may reflect its technological
dynamism; examination of the growth of patent classes can give some hints on the directions of
technological change. The drawbacks of patents as innovation indicators are well known. Many
innovations are not patented, other innovations are covered by multiple patents, and many patents have no
technological and economic value, while others have a very high value (see the Patent Manual, OECD,
1994).

63. These two basic families of statistics are complemented by several others, including: statistics on
scientific publications (bibliometrics), publications in trade and technical journals (“LBIO” or
literature-based indicators of innovation output), skilled human resources, the technology balance of
payments, globalisation indicators, and activity in high-technology sectors (investment, employment,
external trade). Moreover, some information on innovation and innovative activities can be drawn
indirectly from many other sources, such as business surveys or education statistics.

64. The globalisation process affects innovation in a number of ways, through increases in
international competition, in flows of goods, services and knowledge across national boundaries, and in
international interactions. Multinational enterprises (MNE) play a key role in this process. The Handbook
on Economic Globalisation Indicators (OECD, 2005) discusses a number of issues concerning
globalisation and provides a set of relevant globalisation indicators.

65. Wherever possible this Manual draws on the concepts and classifications set out in other volumes
in the set of OECD manuals for the measurement of scientific and technological activities (see Box 2),
especially the Frascati Manual on the resources devoted to R&D (OECD, 2002). This particularly applies
to a number of additional questions on R&D and other S&T activities recommended for inclusion in
innovation surveys in Chapters 6 and 7.

Box 2. Manuals and other Guidelines for the Measurement of Scientific and Technological Activities

Proposed Standard Practice for Surveys of Research and Experimental Development - Frascati Manual, sixth edition,
(OECD, 2002).

OECD Proposed Guidelines for Collecting and Interpreting Technological Innovation Data - Oslo Manual , third edition,
(OECD/EU/Eurostat, 2005).

Using Patent Data as Science and Technology Indicators -- Patent Manual, (OECD, GD 1994).

The Measurement of Human Resources Devoted to S&T - Canberra Manual (OECD/ EU/Eurostat, GD 1995).

Proposed Standard Method of Compiling and Interpreting Technology Balance of Payments Data - TBP Manual (OECD,
GD 1992).

15
Handbook on Economic Globalisation Indicators (OECD, 2005).

Measuring Productivity Manual (OECD, 2001).

A Guide for Information Society Measurements and Analysis (OECD, 2005)

A Framework for Biotechnology Statistics (OECD, 2005).

6.2 Other economic norms and classifications

66. Because of the need to place innovation in a wider context, both conceptually and in terms of
databases, United Nations guidelines and classifications are used as far as possible, notably the System of
National Accounts - SNA (CEC et al. 1994) and the International Standard Industrial Classification - ISIC
Rev. 3.1 (UN, 2002) plus, as this is a joint OECD/Eurostat Manual, the corresponding European norms,
notably the Statistical Classification of Economic Activities in the European Community
- NACE Rev. 1.1 - series 2E.

6.3 Other related concepts and surveys

67. As noted above, there are other ways of examining changes in firms which improve their
productivity and performance. Here we shall examine a few of the most relevant intangible investments:
the generation and adoption of information and communication technology (ICT), biotechnology, and
knowledge management.

68. Information and communication technology covers both hardware and software. Their
development and diffusion is believed to have had a major impact on the pattern of production and
employment in a wide range of industries. In the case of hardware it may be interesting not only to know
when a company first introduces a technologically new or improved piece of ICT equipment but also the
ICT proportion of its total stock of equipment including subsequent purchases of further machines of the
same model. Mapping the development, production, adaptation and use of software is a more complex
matter as these activities are carried out throughout the economy. Surveys have been conducted both on
uses of ICTs and on R&D activity in firms developing ICT products.

69. While less pervasive than ICT, biotechnology is also increasingly expected to have a large
economic impact, through increasing use of biotechnology to produce products and processes. A number
of biotechnology surveys have been conducted, primarily on developers of biotechnology. However, some
surveys have also examined biotechnology users8.

70. Knowledge management involves activities related to the capture, use and sharing of knowledge
by the organisation. It thus involves an important part of the innovation process. There have been a number
of surveys conducted on knowledge management practices in recent years9. These surveys cover policies
and strategies, leadership, knowledge capture, training and communications, and also reasons for use of

8
For example, The Biotechnology Use & Development Survey 1999, Statistics Canada.
9
See Foray, D. and F. Gault (eds.) (2003), Measuring Knowledge Management in the Business Sector: First
Steps, OECD/Statistics Canada. In addition, a Eurostat pilot survey on knowledge management will be
conducted in 2006.

16
knowledge management practices and the sources that prompted the development of practices. In addition,
questions on knowledge management have been included in innovation surveys10.

71. Intangible investment covers expenditures on non-routine marketing, training, software and some
other similar items, in addition to current expenditure on R&D. It covers current expenditure on
innovation but also comprises elements which are not part of current innovation expenditure (for example
it includes all of the firm’s training and marketing expenditure in general, not simply training or marketing
in connection with the implementation of innovations). It does not cover tangible investment such as
capital innovation expenditure, which includes capital expenditure on R&D, the acquisition of other
machinery and equipment related to innovations.

10
For example, the CIS3 survey in France, the J-NIS 2003 survey in Japan, and the 2003 Innovation survey
in Australia.

17
7. FINAL REMARK

72. This Manual is based on a consensus of views on the demand for innovation indicators and the
underlying policy needs and economic theory, on the definitions and coverage of innovation and on the
lessons to be learned from previous surveys. Managed jointly by OECD and Eurostat, it has been written
for and by experts from some thirty countries that collect and analyse innovation data. Finding consensus
has sometimes meant reaching compromises and agreeing to conventions. Furthermore, the complexity of
the innovation process itself makes it difficult to establish absolutely precise guidelines. Nevertheless the
goal of this Manual is to provide a robust set of guidelines which can be applied to produce meaningful
indicators of innovation.

18
 Chapter 2

INNOVATION THEORY AND MEASUREMENT NEEDS

1. INTRODUCTION

73. “The knowledge-based economy” is an expression coined to describe trends in advanced

economies towards greater dependence on knowledge, information and high skill levels, and an increasing
need for ready access to all of these by the business and public sectors. Knowledge and technology have
grown increasingly complex, increasing the importance of links between firms and other organisations in
order to acquire specialised knowledge. A parallel economic development has been the growing
importance of innovation in services in advanced economies.

74. Both innovation research and policy discussion emphasize the importance of viewing innovation
from a broad perspective. A ‘knowledge-based’ view focuses on interactive processes by which knowledge
is created and exchanged both within and outside firms and other organisations. Many knowledge intensive
industries, such as high-technology manufacturing and business services, have undergone strong growth in
many developed economies. Furthermore, a wide range of industries in both manufacturing and services
have increased their utilisation of knowledge intensive technologies for production processes and service
provision. While R&D plays a vital role in the innovation process, much innovation activity is not
R&D-based, but nevertheless draws on highly skilled workers, interactions with other firms and public
research institutions, and on an organisational structure that is conducive to learning and exploiting
knowledge.

75. The objective of this chapter is to examine theories and research on innovation and central policy
issues that are influenced by these theories, and to discuss how and to what extent innovation surveys can
provide data on these issues. Innovation theory informs the design both of policy and measurement, while
at the same time empirical analysis of data adds to our understanding of innovation, policy evaluation and
the formation of new policy initiatives.

2. ECONOMICS OF INNOVATION

76. Research on innovation spans a number of different disciplines, with economic approaches alone
using several different theoretical perspectives. While these can be presented as alternative theories, each
has significant insights and can be seen as complementary. This section evaluates these different
theoretical approaches to innovation and assesses their implications for policy and data collection. The goal
is to ensure that the design of innovation surveys obtains data of relevance to both theory and policy.

77. These theories form a basis for a number of issues for innovation policy and measurement, such
as why firms innovate, what forces drive innovation, and which factors hinder it. Related to this is the
internal functioning of firms and what types of business practices are used to promote innovation. Another
important issue is the nature of knowledge, how it is accumulated, and how it flows between actors. A final
issue is how innovation processes develop at the industry, regional or national levels.

19
78. The work of Joseph Schumpeter has greatly influenced theories of innovation. Schumpeter
argued that economic development is driven by innovation through a dynamic process in which new
technologies replace the old, a process that Schumpeter called “creative destruction”. In Schumpeter’s
view, “radical” innovations create major disruptive changes, whereas “incremental” innovations
continuously fill in the process of change. Schumpeter proposed a list of five types of innovations11:

i) Introduction of new products.

ii) Introduction of new methods of production.

iii) Opening of new markets.

iv) Development of new sources of supply for raw materials or other inputs.

v) Creation of new market structures in an industry.

79. It is crucial to know why firms innovate. The ultimate reason is to improve firm performance,
for example through increasing demand or reducing costs. A new product or process can be a source of
market advantage for the innovator. In the case of productivity-enhancing process innovation, the firm
gains a cost advantage over its competitors, allowing a higher mark-up at the prevailing market price or,
depending on the elasticity of demand, the use of a combination of lower price and higher mark-up than its
competitors to gain market share and increase profits. In the case of product innovation, the firm can gain
a competitive advantage by introducing a new product, which allows it to increase demand and mark-ups.

80. Firms can also increase demand through product differentiation, by targeting new markets, and
by influencing the demand for existing products. Changes in organisational methods can improve the
efficiency and quality of their operations, thereby increasing demand or reducing costs.

81. Innovation can also improve performance by increasing the firm’s ability to innovate. For
example, improving the capabilities of production processes can make it possible to develop a new range
of products, or new organisational practices can improve the firm’s ability to gain and create new
knowledge that can be used in the development of other innovations.

82. A Schumpeterian perspective tends to emphasise innovation as market experiments and to look
for large sweeping changes that fundamentally restructure industries and markets. Mainstream or
neo-classical economics views innovation as working through asset creation as well as through market
experiments. In this view, innovation is an aspect of business strategy, or part of the set of investment
decisions to create capacity for product development or to improve efficiency. Recent developments have
centred on the idea of “sunk costs”, irrecoverable commitments of resources to enter new markets or to
create competitive advantages through repositioning production or output in the value chain12.

83. Appropriation is an important factor for innovation, given that research results and new
technologies often have aspects of a public good, as the costs of making them available to many users are
low compared to their development costs. Once disseminated, users cannot be denied further access to an
innovation. In such a case, the firm cannot capture all the benefits generated by its innovation, which

11
Schumpeter J. (1934), The Theory of Economic Development, Harvard University Press, Cambridge,
Massachusetts.
12
John Sutton, Sunk Costs and Market Structure, MIT Press, 1992 and Technology and Market Structure,
MIT Press, 1998.

20
lessens the incentive to invest in innovative activities. Thus, the ability to protect innovations will have an
important influence on innovation activity.

84. Other work, notably within industrial organisation theory13, has emphasised the significance of
competitive positioning. Firms innovate to defend their existing competitive position as well as to seek
new competitive advantages. A firm might take a reactive approach and innovate to prevent losing market
share to an innovative competitor. Or it may take a proactive approach to gain a strategic market position
relative to its competitors, for example by developing and then trying to enforce higher technical standards
for the products it produces.

85. The decision to innovate often takes place under a great degree of uncertainty14. Future
developments in knowledge and technology, markets, product demand and potential uses for technologies
can be highly unpredictable, although the level of uncertainty will vary by sector, the life cycle of a
product and many other factors. Uncertainty is also present for the adoption of new products or processes,
or the implementation of new marketing or organisational methods. Furthermore, searching for and
collecting relevant information can be very time consuming and costly.

86. Uncertainty can lead firms to hesitate in implementing large changes, even as they face a volatile
environment that increases the pressure to introduce new products, seek new markets, and introduce new
technologies, practices and organisational methods to their production process. Uncertainty can also make
it more difficult for firms in obtaining external funding for their innovation projects.

87. Literature on organisational innovation15 focuses on the role of organisational structures, learning
processes, and adaptation to changes in technology and the environment (the latter includes the
institutional framework and markets).

88. The organisational structure of a firm can influence the efficiency of innovation activities, with
some structures better suited for particular environments. For example, achieving a greater degree of
organisational integration could improve the coordination, planning and implementation of innovation
strategies. Organisational integration can work particularly well in industries characterised by incremental
changes in knowledge and technologies. A looser, more flexible form of organisation, allowing greater
autonomy for workers to make decisions and define their responsibilities, might be more effective in
generating more radical innovations.

89. Organisational learning depends on practices and routines, patterns of interactions both within
and outside of the firm, and the ability to mobilise individual tacit knowledge and promote interaction.
Such learning can be fostered through careful design of practices, routines and relationships, or through a
more flexible, fluid organisation where individuals are encouraged to develop new ideas and ways of doing
things.

90. Marketing theories16 focus on consumer behaviour, market exchanges between buyers and
sellers, and normative approaches to implementing marketing practices. As both buyers and sellers are
heterogeneous, a major challenge for firms is in matching their products with demand. The heterogeneity

13
See for example, Tirole, J. (1995) The Theory of Industrial Organization, MIT Press.
14
Rosenberg, N. (1994) Exploring the Black Box: Technology, Economics, and History, Cambridge
University Press.
15
See for example, Lam, A. (2005), op. cit.
16
See for example, Hunt, S.D. (1983) Marketing Theory: the Philosophy of Marketing Science,
Richard D. Irwin, Inc.

21
of consumers also means that product differentiation is often just as important in capturing demand as the
development of new products. Demand can depend not only on the objective characteristics of products,
but also on their social characteristics and image. Firms can use the latter two features to influence the
demand for their products.

91. Normative marketing theories focus on the implementation of marketing practices. An example is
the Marketing Mix Model17 that focuses on the “4 Ps” of marketing: product, price, promotion and
placement.

92. Product involves changes in product design and packaging that are intended to change or
enhance the product’s appeal or to target a new market or market segment. Price involves the use of
pricing methods to market goods or services. Promotion involves promotional efforts made by firms to
improve their products’ image or to increase awareness of their products. The final P, placement, involves
both the types of sales channels that firms choose to sell their products and how these sales channels are
designed in order to best market their products.

93. The diffusion of new knowledge and technology is a central part of innovation. The diffusion
process often involves more than the mere adoption of knowledge and technology, as adopting firms learn
from and build on new knowledge and technology. Theories of diffusion18 focus on factors that affect
firms’ decisions to adopt new technologies, their access to new knowledge and their absorptive ability.

94. Sociological views on the diffusion of new technologies19 highlight firm attributes that influence
their decisions to adopt new knowledge or technologies, such as the relative advantage of the new
technology, its compatibility with existing ways of doing things, its complexity, and the ease with which
the firm is able to evaluate the new technology. Economic views on diffusion tend to focus on the costs
and benefits of adopting new technologies. These potential benefits can often be strategic, so as to keep up
with or gain an advantage over competitors.

95. Access to knowledge and technology can depend to a large extent on the connections between
firms and organisations. This is particularly the case for tacit knowledge that is held in the minds of
people, or for information that is held in the ‘routines’ of organisations. Direct interactions with the people
who hold tacit knowledge or who have access to routines is required in order to gain access to these types
of knowledge.

96. Much knowledge is codified, and thus can be accessed and used without direct interaction with
the source. Transfers of codified knowledge comprise a large share of technology diffusion and efforts
made to promote the codification of knowledge can have important impacts on productivity, growth and
innovation. An example is work on the establishment of technology standards.

97. Even for information that is openly available, finding it can pose a major challenge, in particular
since searching for new information can be very costly. Thus, ease of communication, effective channels
of information, and skills transmission, within organisations and between them, are highly important for
diffusion.

17
See for example, Perreault, W.D and E.J. McCarthy (2005) Basic Marketing: A Global Managerial
Approach, McGraw-Hill.
18
See for example, Hall, B. (2005) “Innovation and Diffusion”, Chapter 17 in J. Fagerberg, D. Mowery, and
R.R. Nelson (eds.) The Oxford Handbook of Innovation, Oxford University Press.
19
See for example, Rogers, E.M. (1995) Diffusion of Innovations, fourth edition. New York: The Free Press.

22
98. Evolutionary approaches20 view innovation as a path-dependent process whereby knowledge and
technology are developed through the interactions between various actors and other factors. The structure
of these interactions affects the future path of economic change. As an example, market demand and the
opportunities for commercialisation influence which products are developed and which technologies are
successful.

99. Closely linked to the evolutionary approach is the view of innovation as a system. The systems of
innovation approach21 studies the influence of external institutions, broadly defined, on the innovative
activities of firms and other actors. It emphasises the importance of the transfer and diffusion of ideas,
skills, knowledge, information and signals of many kinds. The channels and networks through which this
information circulates are embedded in a social, political and cultural background that guide and constrain
innovation activities and capabilities. Innovation is seen as a dynamic process in which knowledge is
accumulated through learning and interaction. These concepts were initially introduced in terms of national
systems of innovation, though they are also applicable to regional and international systems.

100. Systems approaches to innovation shift the focus of policy towards an emphasis on the interplay
between institutions, looking at interactive processes in the creation, diffusion and application of
knowledge. This emphasizes the importance of the conditions, regulations and policies within which
markets operate - and hence the role of governments in monitoring and seeking to fine-tune this overall
framework.

3. A MEASUREMENT FRAMEWORK

101. These different theories form the basis for the innovation measurement framework used in this
Manual, highlighting among other things: the driving forces behind innovation, the importance not only of
products and processes but also of marketing and organisational practices, the role of linkages and
diffusion, and the view of innovation as a system.

102. The link between innovation and economic change is of central interest. Through innovation,
new knowledge is created and diffused, which expands the economy’s potential to develop new products
and more productive operating methods. This production frontier not only depends on technological
knowledge, but also on other forms of knowledge that are used to develop product, process, marketing and
organisational innovations. Each specific type of innovation can differ greatly in its impacts on firm
performance and on economic change. For this reason, it is important to be able to identify the
implementation and impacts of different types of innovation.

103. Figure 1 depicts this framework from the perspective of the firm, which is the target for
innovation surveys. Other models of the innovation process, such as Kline and Rosenberg’s chain-link
model or the innovation dynamo22, provide a useful theoretical overview for innovation processes but are
less adapted to guiding survey design. The framework used in this Manual thus represents an integration of
insights from various firm-based theories of innovation with those of approaches that view innovation as a
system. The main characteristics of the framework described here are:

20
See Nelson R. and Winter S. (1982) An Evolutionary Theory of Economic Change, Belknap Press of
Harvard University Press, Cambridge, Massachusetts.
21
See Lundvall B-A (ed.) (1992) National Systems of Innovation: Towards a Theory of Innovation and
Interactive Learning, Pinter Publishers, London, and Nelson R. (1993) National Innovation Systems,
Oxford UP, Oxford.
22
See OECD/Eurostat (1997) The Oslo Manual.

23
 • Innovation in the firm.
• Linkages with other firms and public research institutions.
• The institutional framework that firms operate within.
• The role of demand.

Figure 1. The Innovation Measurement Framework

Infrastructure and Institutional Framework

The Firm

Product Process
Innovations Innovations

Marketing Organisational
Innovations Innovations

Innovation
Policies
Education and
Other Firms Public Research
System

Demand

104. Innovation in firms refers to planned changes in a firm's activities aiming at improving the firm's
performance. Drawing on the discussion of innovation theory above, the concept of innovation used in this
Manual refers to changes that are characterised by the following features:

a) Innovation is associated with uncertainty over the outcome of innovation activities. It is not
known beforehand what the result of the innovation activities will be: e.g. whether R&D will
result in the successful development of a marketable product or how much time and resources will

24
 be needed to implement a new production process, marketing or organisational method and how
successful they will be.

b) Innovation involves investment. Relevant investment can include acquisitions of fixed and
intangible assets as well as other activities (such as salaries, or purchase of material or services)
that might only yield potential returns in the future.

c) Innovation is subject to spillovers: the benefits of creative innovation are rarely fully appropriated
by the inventive firm. Other firms that innovate through adopting the innovation can benefit from
knowledge spillovers or from the use of the original innovation. For some innovation activities,
imitation costs are substantially less than development costs, which can require an effective
appropriation mechanism to provide an incentive to innovate.

d) Innovation involves the utilisation of new knowledge or a new use or combination of existing
knowledge. New knowledge may either be generated by the innovating firm in the course of its
innovation activities (i.e. through intramural R&D) or acquired externally through various
channels (e.g. purchasing of new technology). The use of new knowledge or the combination of
existing knowledge requires innovative efforts by the firm that can be distinguished from
standardised routines.

e) Innovation aims at improving a firm's performance through gaining competitive advantages (or
simply maintaining competitiveness) by either shifting the demand curve of the firm’s products
(e.g. through increasing product quality, offering new products or opening up new markets or
groups of customers) or a firm's cost curve (e.g. through reducing unit costs of production,
purchasing, distribution or transaction), or by improving the firm’s ability to innovate
(e.g. increasing the capability to develop new products or processes, or to gain and create new
knowledge).

105. There are two main archetypes of innovative activity, or options open to a firm that wants to
change its products, capabilities, or production, marketing and organisational systems. A firm can invest in
creative activities to develop innovations in-house, either alone or in conjunction with external partners, or
it can adopt innovations developed by other firms or institutions as part of a process of diffusion. Between
these two archetypes are a countless number of combinations, such as when a firm adopts an organisational
innovation developed by another firm and adjusts it to function with its own work routines, when it adapts
new manufacturing technology to its production line, or when a firm introduces a new component obtained
from a supplier into a consumer product. Both the creation and adoption of innovations can involve
intensive learning and interaction with other actors or minimal external linkages.

106. The specific innovation activities that firms can use to develop or acquire innovation include
R&D and/or many other activities that are summarized below.

− R&D: R&D is defined in the Frascati Manual and includes basic and applied research plus
experimental development:

i) The firm can undertake basic research to extend its knowledge of fundamental
processes related to what it produces.

ii) It can engage in research to broaden the range of applied projects that are open to
it, and applied research to produce specific inventions or modifications of existing
techniques.

25
 iii) It can develop product or process concepts or other changes to judge whether they
are feasible and viable, a stage which may involve (a) development and testing, and
(b) further research to modify designs or technical functions.

− Other innovative activity: The firm can engage in many non-R&D activities that are used in
innovation. These activities can strengthen capabilities that enable the development of innovations
or the ability to successfully adopt innovations developed by other firms or institutions:

iv) It can identify new concepts for products, processes, marketing methods or
organisational changes a) via its marketing side and relations with users, b) via the
identification of opportunities for commercialisation resulting from its own or
others’ basic or strategic research, c) via its design and development capabilities,
d) by monitoring competitors and e) by using consultants.

v) It can buy technical information, paying fees or royalties for patented inventions
(which usually require research and development work to adapt and modify the
invention to its own needs), or buy know-how and skills through engineering,
design or other consultancy services.

vi) Human skills can be developed (through internal training) or purchased (by hiring);
tacit and informal learning - “learning by doing” - may also be involved.

vii) It can invest in equipment, software or intermediate inputs which embody the
innovative work of others.

viii) It can reorganise management systems and its overall business activities.

ix) It can develop new methods of marketing and selling its goods and services.

107. All these innovation activities have the end objective of improving firm performance. They can
be intended to develop and implement new products and processes, new methods of promoting and selling
the firm’s products and/or changes in the firm’s organisational practices and structure.

108. The general institutional environment determines the broad parameters within which firms
operate. The component elements include:

• The basic educational system for the general population, which determines minimum
educational standards in the workforce and the domestic consumer market.

• The university system.

• The specialised technical training system.

• The science and research base.

• Common pools of codified knowledge, such as publications, technical, environmental and

management standards.

• Innovation policies and other government policies that influence firm innovation.

26
 • Legislative and macro-economic settings such as patent law, taxation, corporate governance rules
and policies relating to interest and exchange rates, tariffs, and competition.

• The communications infrastructure, including roads, telephones and electronic communication.

• Financial institutions determining, for example, the ease of access to venture capital.

• Market accessibility, including possibilities for the establishment of close relations with
customers as well as matters such as size and ease of access.

• Industry structure and the competitive environment, including the existence of supplier firms in
complementary sectors.

4. SECTORAL AND REGIONAL ASPECTS OF INNOVATION

109. Innovation processes differ greatly from sector to sector in terms of development activities, the
rate of technological change, linkages and access to knowledge, and in organisational structures and
institutional factors23. Some sectors are characterised by rapid change and radical innovations, others by
smaller, incremental changes.

110. In high technology sectors, R&D plays a central role in innovation activities, while other sectors
rely to a greater degree on the adoption of knowledge and technology. Differences in innovation activity
across sectors (e.g. whether mainly incremental innovations or radical innovations) also place different
demands on the organisational structure of firms, and institutional factors such as regulations and
intellectual property rights can vary greatly in their role and importance. These differences are important to
take into account when designing policy. They are also important for measurement, both in collecting data
that allows analysis across sectors and regions and in ensuring that the measurement framework is
applicable for a broad range of industries.

4.1 Innovation in services

111. There has been an increasing recognition of the importance of innovation in the service sector
and the service sector’s contribution to economic growth. This has also led to a number of studies
focussing on innovation in services24.

112. The service sector is diverse, covering a number of activities. Howells and Tether25 classify
services into four groups: services dealing mainly with goods (such as transport and logistics), those
dealing with information (such as call centres), knowledge-based services, and services dealing with

23
See for example, Malerba, F. (2005), “Sectoral Systems: How and Why Innovation Differs Across
Sectors”, Chapter 14 in J. Fagerberg, D. Mowery, and R.R. Nelson (eds.) The Oxford Handbook of
Innovation, Oxford University Press.
24
For example, de Jong, J.P.J., A. Bruins, W. Dolfsma, and J. Meijaard (2003) Innovation in Services Firms
Explored: What, How and Why?, EIM Report; Hauknes, J. (1998) Services in Innovation, Innovation in
Services, SI4S Final Report; and Howells, J.R.L. and B.S Tether (2004) Innovation in Services: Issues at
Stake and Trends – A Report for the European Commission. See also Miles, I. (2005), “Innovation in
Services”, Chapter 16 in J. Fagerberg, D. Mowery, and R.R. Nelson (eds.) The Oxford Handbook of
Innovation, Oxford University Press.
25
Howells, J.R.L. and B.S Tether (2004), op.cit.

27
people (such as health care). While this diversity should be kept in mind, there are several general
characteristics that apply to most service sectors.

113. A key element of services is that the distinction between products and processes is often blurred,
with production and consumption occurring simultaneously. Development processes for services can be
more informal than for goods with an initial phase consisting of search, idea gathering and commercial
evaluation, followed by implementation.

114. Innovation activity in services also tends to be a continuous process, consisting of a series of
small incremental changes in products and processes. This can occasionally complicate the identification
of innovations in services in terms of single events, i.e. as the implementation of a significant change in
products, processes or other methods.

4.2 Innovation in low and medium technology industries

115. Innovation in low and medium technology industries (LMT) often receives less attention than
innovation in high technology industries. However, innovation in LMTs can have a substantial impact on
economic growth, due to the large economic size of these sectors.

116. LMT industries are generally characterised by incremental innovation and adoption activities. As
such, innovation activities are often focussed on production efficiency, product differentiation and
marketing26. An important aspect of innovation that has emerged in these industries is that innovation is
more complex than the simple adoption of new technologies. In many cases, innovation in LMTs involves
the incorporation of high technology products and technologies in their innovation activities. Prominent
examples are the use of ICTs and biotechnology (e.g. in food processing) in both the development of new
products and production processes. The use and application of advanced technologies by LMTs can place
different demands on the skills of their workforce, their organisational structure, and their interactions with
other firms and public research institutions.

4.3 Innovation in small and medium sized enterprises

117. Small and medium sized enterprises (SME) are necessarily more specialised in their activities.
This increases the importance of efficient interactions with other firms and public research institutions for
R&D, exchange of knowledge, and potentially also for commercialisation and marketing activities.

118. Finance can be a determining factor for innovation in SMEs, which often lack internal funds to
conduct innovation projects and have much greater difficulty in obtaining external funding than larger
firms. Surveys can provide data on the importance of financial constraints to the ability of SMEs to
innovate.

4.4 Regional innovation

119. Regional analysis of innovation has received increasing interest, based on the view that regional
factors can influence the innovative capacity of firms. Regional differences in the level of innovation
activity can be substantial and important elements in understanding innovation processes and forming
26
See for example, Von Tunzelmann, N. and V. Acha (2005), “Innovation in ‘Low Tech’ Industries”,
Chapter 15 in J. Fagerberg, D. Mowery, and R.R. Nelson (eds.) The Oxford Handbook of Innovation,
Oxford University Press.

28
policy lie in identifying the main characteristics and factors that promote general innovation activity and
the development of specific sectors.

120. Parallel to national systems of innovation, many regions can also be considered innovation
systems. The presence of, for example, local public research institutions, large dynamic firms, industry
clusters, venture capital and a strong entrepreneurial environment can influence the innovative
performance of regions. This includes the potential for contacts with suppliers, customers, competitors and
public research institutions. Infrastructure also plays an important role in innovation activity.

4.5 Globalisation

121. National systems of innovation build on the idea that many factors that influence innovation
activities are national, such as institutional factors, culture and values. At the same time, it is also clear that
innovation processes are, in many ways, international processes. Technology and knowledge flows across
borders. Firms interact with foreign firms and universities. Many markets, in terms of firms and their
competitors, are global. The internet has greatly enhanced opportunities for communicating and doing
business with firms in other countries.

122. Multinational enterprises are a key element in globalisation. Their activities transcend national
boundaries, involving international transfers of capital, knowledge and technology.

123. The globalisation process also presents a powerful driving force for innovation. International
competition forces firms to increase efficiency and to develop new products. Globalisation can also change
the industrial structure of economies, pushing them to develop new industries and adapt their institutional
frameworks.

5. AREAS FOR INVESTIGATION

124. It is essential to feed the debate on policy issues with relevant information and analysis of many
aspects of innovation. Ideally, a comprehensive information system should be constructed that covers all
types of factors relevant to innovation policy and research. This would place governments in a strong
position to deal appropriately with any particular policy issues that might arise. In practice, only parts of
such a system can be covered by indicators, while other parts call for qualitative information. Moreover,
as policy and indicator analysts are well aware, indicators will only occasionally relate neatly to a single
factor or issue, and more often than not will relate to a range of matters and only partially to each. Any
broad information or monitoring system will also need to be supplemented with case studies or specialized
surveys where specific and in-depth analysis is required.

5.1 What can be measured

125. Innovation surveys can provide a wide breadth of information on the innovation process at the
firm level. This includes identifying motives and obstacles to innovation, changes in the way firms operate,
the kinds of innovation activity that they engage in, and the types of innovations that they implement. In
terms of the innovation process as a system, innovation surveys can provide information on firm linkages
with other actors in the economy and on their use of different methods to protect their innovations. These
areas are discussed in greater detail below.

29
126. There are also limitations to the types of data that can be obtained by innovation surveys. First,
analysis will often require additional economic data for the firm, so that data from innovation surveys will
often need to be combined with other information sources.

127. Second, innovation is a continuous process. This is difficult to measure, particularly for firms
whose innovation activity is mainly characterised by small, incremental changes as opposed to single,
well-defined projects for implementing significant changes. Innovations are defined in this Manual as
significant changes, with the intention of distinguishing significant changes from routine, minor changes.
However, it is important to recognise that an innovation can also consist of a series of minor incremental
changes.

128. Third, quantitative information on innovation expenditures are normally not specified in firms’
financial accounts, making it difficult for firms to calculate them. While this does not preclude the
measurement of innovation expenditures, these difficulties should be taken into account both for survey
design and subsequent analysis.

129. Fourth, it is difficult for surveys to capture the timing of innovation activities, their
implementation, and their impacts. Expenditures on innovation activities are investments intended to yield
potential returns in the future. The outputs of these innovation activities, from the development and
implementation of innovations to improvements in innovative capacity to impacts on performance, will
often not be observable during the review period.

130. Finally, innovation surveys are less able to provide information on the general institutional
environment, such as the education system, the labour market and financial systems, with the exception of
how these institutional factors are experienced by the respondent firms.

5.2 Inputs to innovation

131. A broad understanding of the distribution of innovation activities across industries is of obvious
importance for innovation policy. An important goal is to understand the role of R&D and non-R&D inputs
in the innovation process and how R&D may be interrelated with other innovation inputs. Better
information on the prevalence of non-R&D innovation inputs is particularly important for many service
sectors where R&D is less frequent.

132. Innovation activities can lead to both the development and implementation of innovations over
the short term and to improvements in the innovation capacity of the firm, as firms learn from the
development and implementation of innovations, gain valuable inputs from interactions and marketing
activity, and improve their innovation capacity through organisational changes.

133. Innovation surveys can provide both qualitative and quantitative data on firm innovation inputs.
Collecting quantitative data poses practical difficulties, especially when firms have many divisions, but
data on innovation inputs are one of the most useful outputs of innovation surveys.

5.3 Linkages and the role of diffusion

134. Innovation surveys can provide data for use in analyses of technological change and productivity
growth based on tracking flows of new knowledge and technologies from one industry to another. An
example is the use of ICTs in a wide variety of products. How do firms incorporate new knowledge and
innovations that have been developed elsewhere? Also, what is the relative weight of diffusion compared
to creative innovation?

30
135. A separate but related issue concerns interactions both among organisations and between
organisations and their surroundings. Systems approaches often highlight linkages as the most vital area for
promoting innovation activity. It is important to determine both the types of interactions and the main
sources of external knowledge.

136. These interactions can consist of informal contacts and information flows, or more formal
collaboration on innovation projects. They include adjustments to the value-chain, such as closer
relationships with suppliers or users, or research on market demand or on the potential uses for
technologies. Firms can have close relationships with other firms within an industry cluster or be part of
looser networks. They can draw on published work from public research institutions or work directly with
them on collaborative projects.

137. Innovation surveys can highlight the specific kinds of information of use to innovation and the
types of institutions and transmission mechanisms that are available to firms. This includes data on
information sources, knowledge and technology flows, and collaborative partners. It also includes data on
barriers to innovation from a lack of information, for example on technologies or markets.

5.4 The impact of innovation

138. Of particular interest to innovation policy is the effect of innovation on output, productivity and
employment, both on a national level and for specific sectors and regions. Better information on the
conditions for success could help improve policies to support the economic and social benefits of
innovation.

139. Innovation surveys can provide both qualitative and quantitative data on the outcomes of
innovations. In addition, survey data can very useful as input to subsequent empirical analyses of the
impacts of innovation.

5.5 Incentives and obstacles to innovation

140. Many innovation support policies would benefit from the identification of the main forces that
drive firm innovation activity. These forces can be market related, concerned with increasing quality or
efficiency, or involve adapting the firm’s organisation to better fit its needs. Information on the objectives
for innovation can be readily obtained through innovation surveys.

141. Obstacles to innovation are significant for policy as well, since a good proportion of government
measures are in one way or another aimed at overcoming them. Many obstacles - skill shortages, problems
of competence, finance, and appropriation - are relatively straightforward to assess with survey methods.

5.6 Demand

142. Demand factors shape innovation activity in a number of ways. Demand affects the development
of new products, as firms modify and differentiate products to increase sales and market share. Demand
factors can force firms to improve their production and supply processes in order to reduce costs, allowing
lower prices. Market factors determine the commercial success of specific technologies or products and
thereby affect the path of technological change. Demand factors can also be the main driver of innovation
in many cases. Firms often spend substantial resources in researching demand and can implement
marketing measures to influence or create demand for their products. Finally, market factors can also

31
determine whether firms innovate or not. If firms do not believe that there is sufficient demand for new
products in their market, they may decide not to innovate or to delay innovation activities.

143. Both the measurement and analysis of the role of demand in innovation are problematic. It is very
difficult to isolate demand effects from supply, and little is known on how to measure demand effects in
surveys. Nevertheless, some aspects of demand could be covered in a survey, such as the extent to which
innovations in products and services, and in the way they are provided, are driven through articulated
customer requirements, or supply driven. The nature of customers and users is also a demand factor that
the innovating firm will take into account. Some information can be obtained through innovation surveys
in terms of the importance of market related factors (both positive and negative) on innovation activity.

5.7 Other

5.7.1 Human resources

144. Much innovation knowledge is embodied in people and their skills, or an appropriate skill set is
needed to make intelligent use of external sources or codified knowledge. The role of human capital in
innovation is important both at the firm level and at the aggregate level. Some questions of interest here are
the quality of the education system and how well it matches the needs of innovative firms and other
organisations; what efforts firms take to invest in the human capital of their employees; whether innovation
activity is hampered by shortages of qualified personnel; whether there are sufficient opportunities for
worker training; and how adaptive the workforce is in terms of the structure of the labour market and
mobility across regions and sectors. However, the role of human capital in innovation is less developed,
with limited information available from innovation surveys.

5.7.2 Laws and regulations

145. Laws and regulations are part of the framework that firms must work within. Well-designed
regulations and standards can provide a strong signal to support and guide innovative activities. They
affect access to information, property rights, tax and administrative burdens (in particular for small firms)
and environmental standards. All these are important for innovation policy, but policy needs can vary
greatly from sector to sector.

146. For example, policies that reduce administrative barriers for SMEs can have a significant effect
on innovation activity in smaller firms. Clear intellectual property rights are also vital for improving
incentives to innovate in some industries. Innovation surveys can provide data on these issues through
questions on obstacles to innovation and on the appropriation methods used by innovative firms.

32
 Chapter 3

BASIC DEFINITIONS

1. INTRODUCTION

147. Based on the concepts discussed in Chapter 2, this chapter aims to provide concise definitions for
types of innovations, innovation activities and innovative firms.27 Given the complexity of the innovation
process and the variations in the way innovation occurs in firms, conventions have to be adopted in order
to provide operational definitions that can be used in standardised firm surveys.

148. The Manual distinguishes innovation in four areas: product, process, marketing and
organisational. Product and process innovations are familiar concepts in the business sector, and have been
the sole focus of the previous editions of the Manual with organisational changes covered in an annex and
no coverage of marketing innovations. Both organisational and marketing innovations are extensively
discussed in this Manual. Marketing and organisational innovations are familiar concepts to firms in some
countries and there are examples of their use in innovation surveys, although their definitions are in general
not as well established as those for products and processes. The definitions of these new types of
innovations for use in surveys are still under development, in much the same way as product and process
innovations were in the first edition of the Oslo Manual.

2. INNOVATION

149. An innovation is the implementation of a new or significantly improved product (good or

service), or process, a new marketing method, or a new organisational method in business practices,
workplace organisation or external relations.

150. This broad definition of an innovation encompasses a wide range of possible innovations. An
innovation can be more narrowly categorised as the implementation of one or more types of innovations,
for instance product and process innovations. This narrower definition of product and process innovations
can be related to the definition of technological product and process innovation used in the second edition
of the Oslo Manual.

151. The minimum requirement for an innovation is that the product, process, marketing method, or
organisational method must be new (or significantly improved) to the firm. This includes products,
processes and methods that firms are the first to develop and those that have been adopted from other firms
or organisations.

27
In this chapter, the generic term, ‘the firm’, is used. The primary statistical unit for collecting data, which
in general is the enterprise, is discussed in the Chapter 4. In particular, the question of what is ‘new to the
firm’ becomes more complicated when considering large enterprise groups or multinational enterprises.
These issues will be considered in Chapter 4.

33
152. Innovation activities are all scientific, technological, organisational, financial and commercial
steps which actually, or are intended to, lead to the implementation of innovations. Some innovation
activities are themselves innovative, others are not novel activities but are necessary for the
implementation of innovations. Innovation activities also include R&D that is not directly related to the
development of a specific innovation.

153. A common feature of an innovation is that it must have been implemented. A new or improved
product is implemented when it is introduced on the market. New processes, marketing methods or
organisational methods are implemented when they are brought into actual use in the firm’s operations.

154. Innovation activities vary greatly in their nature from firm to firm. Some firms can engage in
well-defined innovation projects, such as the development and introduction of a new product, whereas
other firms can primarily make continuous improvements to their products, processes and operations. Both
types of firms can be innovative: an innovation can consist of the implementation of a single significant
change, or of a series of smaller incremental changes that together constitute a significant change.

155. An innovative firm is one that has implemented an innovation during the period under review.

156. The broad definition of an innovative firm may not be appropriate for all policy and research
needs. More narrow definitions can be useful in many cases (see sections 7 and 8 of this Chapter below),
particularly for comparisons of innovation across sectors, firm size categories, or countries. An example of
a more narrow definition is a product or process innovator:

157. A product-process innovative firm is one that has implemented a new or significantly improved
product or process during the period under review. This definition, which includes all firms that have
implemented a product or process innovation, is similar to the definition of the TPP innovative firm in the
previous edition of the Oslo Manual.

3. MAIN TYPES OF INNOVATION

158. Four types of innovations are distinguished: product innovations, process innovations, marketing
innovations, and organisational innovations. This classification maintains the largest possible degree of
continuity with the previous definition of technological product and process innovation used in the second
edition of the Oslo Manual. Product innovations and process innovations are closely related to the concept
of technological product innovation and technological process innovation. Marketing innovations and
organisational innovations broaden the range of innovations covered by the Manual compared to the
previous definition.

159. A product innovation is the introduction of a good or service that is new or significantly
improved with respect to its characteristics or intended uses. This includes significant improvements in
technical specifications, components and materials, incorporated software, user friendliness or other
functional characteristics.

160. Product innovations can utilise new knowledge or technologies, or can be based on new uses or
combinations of existing knowledge or technologies. The term “product” is used to cover both goods and
services. Product innovations include both the introduction of new goods and services and significant
improvements in the functional or user characteristics of existing goods and services.

161. New products are goods and services that differ significantly in their characteristics or intended
uses from products previously produced by the firm. The first microprocessors and digital cameras were
examples of new products using new technologies. The first portable MP3 player, which combined

34
existing software standards with miniaturized hard-drive technology, was a new product combining
existing technologies.

162. The development of a new use for a product with only minor changes to its technical
specifications is a product innovation. An example is the introduction of a new detergent using an already
existing chemical composition that was previously used as an intermediary for coating production only.

163. Significant improvements to existing products can occur through changes in materials,
components and other characteristics that enhance performance. The introduction of ABS braking, GPS
(Global Positioning Systems) navigational systems, or other subsystem improvements in cars is an
example of a product innovation consisting of partial changes or additions to one of a number of integrated
technical subsystems. The use of breathable fabrics in clothing is an example of a product innovation
involving the use of new materials that improves the performance of the product.

164. Product innovations in services can include significant improvements in how they are provided
(for example, in terms of their efficiency or speed), the addition of new functions or characteristics to
existing services, or the introduction of entirely new services. Examples are significant improvements in
internet banking services such as greatly improved speed and ease of use, or the addition of home pick-up
and drop-off services that improve customer access for rental cars. Providing on-site rather than remote
management contact points for outsourced services is an example of an improvement in service quality.

165. Design is an integral part of the development and implementation of product innovations.
However, design changes that do not involve a significant change in a product’s functional characteristics
or intended uses are not product innovations, though they can be marketing innovations, as discussed
below. Routine upgrades28 or regular seasonal changes are also not product innovations.

166. A process innovation is the implementation of a new or significantly improved production or

delivery method. This includes significant changes in techniques, equipment and/or software.

167. Process innovations can be intended to decrease unit costs of production or delivery, to increase
quality, or to produce or deliver new or significantly improved products.

168. Production methods involve the techniques, equipment and software used to produce goods or
services. Examples of new production methods are the implementation of new automation equipment on a
production line or the implementation of computer-assisted design for product development.

169. Delivery methods concern the logistics of the firm, encompassing equipment, software and
techniques to source inputs, allocate supplies within the firm, or deliver final products. Examples of new
delivery methods are the introduction of a bar coded or active RFID (Radio Frequency Identification)
goods tracking system.

170. Process innovations include new or significantly improved methods for the creation and
provision of services. They can involve significant changes in equipment and software used in
services-oriented firms or in the procedures or techniques that are employed to deliver services. Examples
are the introduction of GPS tracking devices for transport services, the implementation of a new

28
A routine upgrade involves minor changes to a good or service that are expected and planned for in
advance. The development of the upgrade is also based on routine, well-established activities. For example,
anti-virus software is purchased with the expectation of frequent upgrades to cover the appearance of new
viruses. A hotel chain will construct new hotels with the expectation that furnishings, lamps, and bathroom
fixtures will be upgraded on a regular basis, even though this may be on a ten or twenty year cycle.

35
reservation system in a travel agency, and the development of new techniques for managing projects in a
consultancy firm.

171. Process innovations also cover new or significantly improved techniques, equipment and
software in ancillary support activities, such as purchasing, accounting, computing and maintenance. The
implementation of new or significantly improved Information and Communication Technologies is a
process innovation if it is intended to improve the efficiency and/or quality of an ancillary support activity.

172. A marketing innovation is the implementation of a new marketing method involving significant
changes in product design or packaging, product placement, product promotion or pricing.

173. Marketing innovations are aimed at better addressing customer needs, opening up new markets,
or newly positioning a firm’s product on the market, with the objective of increasing the firm’s sales.

174. The distinguishing feature of a marketing innovation compared to other changes in a firm's
marketing instruments is the implementation of a marketing method not used before in the firm. They must
be part of a new marketing concept or strategy that represents a significant departure from the firm’s
existing marketing methods. The new marketing method can either be developed by the innovating firm or
adopted from other firms or organisations. New marketing methods can be implemented for both new and
existing products.

175. Marketing innovations include significant changes in product design that are part of a new
marketing concept. Product design changes here refer to changes in product form and appearance that do
not alter the product’s functional or user characteristics. This also includes changes in the packaging of
products such as foods, beverages and detergents, where packaging is the main determinant of the
product’s appearance. An example of a marketing innovation in product design is the implementation of a
significant change in the design of a furniture line to give it a new look and widen its appeal. Innovations
in product design can also include the introduction of significant changes in the form, appearance or taste
of food or beverage products, such as the introduction of new flavours for a food product in order to target
a new customer segment. An example of a marketing innovation in packaging is the use of a fundamentally
new design of bottles for a body lotion, intended to give the product a distinctively exclusive look to appeal
to a new market segment.

176. New marketing methods in product placement primarily involve the introduction of new sales
channels. Sales channels refer to the methods used to sell goods and services to customers, and not to
logistics methods (transport, storing and handling of products) dealing mainly with efficiency. Examples of
marketing innovations in product placement are the first-time introduction of a franchising system, the
first-time introduction of direct selling or exclusive retailing, and the first-time introduction of product
licensing. Innovations in product placement can also involve the use of new concepts for product
presentation. An example is the introduction of sales rooms for furniture that are redesigned according to
themes, allowing customers to view products in fully decorated rooms.

177. New marketing methods in product promotion involve the use of new concepts for promoting the
firm’s goods and services. For example, the first use of a significantly different media or technique - such
as product placement in movies or television programmes, or the use of celebrity endorsements - is a
marketing innovation. Another example of an innovation in promotion is branding, such as the
development and introduction of a fundamentally new brand symbol (as distinguished from a regular
update of the brand appearance) that is intended to position the firm's product on a new market or give the
product a new image. The introduction of a personalised information system, e.g. obtained from loyalty
cards, to tailor the presentation of products to the specific needs of individual customers can also be
considered a marketing innovation.

36
178. Innovations in pricing involve the use of new pricing strategies to market the firm’s goods or
services. Examples are the first time use of a new method for varying the price of a good or service
according to its demand (e.g. when demand is low, the price is low) or the introduction of a new method
where customers are able to choose desired product specifications on the firm’s website and then can see
the price for the specified product. New pricing methods whose sole purpose is to differentiate prices by
customer segments are not considered innovations.

179. Seasonal, regular and other routine changes in marketing instruments are generally not marketing
innovations. In order for these changes to be marketing innovations, they must involve marketing methods
not used before in the firm. For example, a significant change in a product’s design or packaging that is
based on a marketing concept that has already been used by the firm for other products is not a marketing
innovation. Nor is the use of existing marketing methods to target a new geographical market or a new
market segment (e.g. socio-demographic group of clients) a marketing innovation.

180. An organisational innovation is the implementation of a new organisational method in the firm’s
business practices, workplace organisation or external relations.

181. Organisational innovations can be intended to increase a firm's performance through reducing
administrative costs or transaction costs, improving workplace satisfaction (and thus labour productivity),
gaining access to non-tradable assets (such as non-codified external knowledge) or reducing costs of
supplies.

182. The distinguishing features of an organisational innovation compared to other organisational

changes in a firm is the implementation of an organisation method (in business practices, workplace
organisation or external relations) that has not been used before in the firm and which is the result of
strategic decisions taken by management.

183. Organisational innovations in business practices involve the implementation of new methods for
organising routines and procedures of the conduct of work. These include, for example, the implementation
of new practices to improve learning and knowledge sharing within the firm. An example is the first-time
implementation of practices for codifying knowledge, e.g. establishing databases of best practices, lessons,
and other knowledge, so that they are more easily accessible for others. Another example is the first-time
implementation of practices for employee development and improving worker retention, such as education
and training systems. The first-time introduction of management systems for general production or supply
operations, such as supply chain management systems, business re-engineering, lean production, and
quality-management systems are other examples of organisational innovations in business practices.

184. Innovations in workplace organisation involve the implementation of new methods for
distributing responsibilities and decision-making autonomy among employees, for the division of work
within and between different firm activities (and organisational units) as well as new concepts for the
structuring of activities, such as the integration of different business activities. An example of an
organisational innovation in workplace organisation is the first-time implementation of organisational
models where the firm’s employees are given greater autonomy in decision-making and employees are
encouraged to contribute with their own ideas. This may be achieved through the decentralisation of group
activity and management control or the establishment of formal or informal work teams with more flexible
job responsibilities for individual workers. On the other hand, organisational innovations can also involve
the centralisation of activity and greater accountability for decision making. An example of organisational
innovation in the structuring of business activities is the first-time introduction of build-to-order production
systems (integrating sales and production) or the integration of engineering and development with
production.

37
185. New organisational methods in a firm's external relations involve the implementation of new
ways of organising the relations to other firms or public institutions, such as the establishment of new types
of collaborations with research organisations or customers, new methods of integration with suppliers, and
first-time outsourcing or subcontracting of business activities in production, procuring, distribution,
recruiting and ancillary services.

186. Changes in business practices, workplace organisation, or external relations that are based on
organisational methods already in use in the firm are not organisational innovations. Nor is the formulation
of managerial strategies an innovation, in itself. However, the organisational changes that are implemented
in response to a new managerial strategy are innovations if they represent a first-time implementation of a
new organisational method in business practices, workplace organisation, or external relations. For
example, the introduction of a written strategy document to improve the efficient use of the firm’s
knowledge is not, by itself, an innovation. Innovation occurs when the strategy is implemented through the
actual use of new software and practices for documenting information in order to encourage knowledge
sharing among different divisions.

187. Mergers with, and the acquisition of other firms are not considered organisational innovations,
even if a firm merges with or acquires other firms for the first time. Mergers and acquisitions may involve
organisational innovations, however, in the case where the firm develops or adopts new organisation
methods in the course of the merger or acquisition.

4. DISTINGUISHING BETWEEN TYPES OF INNOVATIONS

188. It is important for survey purposes to be able to distinguish between innovation types in
borderline cases. However, many innovations can have multiple characteristics that span more than one
type of innovation. It can be both difficult and misleading, in terms of types of innovation activities
undertaken by firms, to categorise these innovations as a single type. This section provides guidelines for
distinguishing between the different types of innovations.

189. Collecting data on the different characteristics of an innovation that spans several innovation
types will rarely create problems for interpretation and, in fact, will usually improve the quality of the
results. For example, a firm that introduces a new product that also requires the development of a new
process is clearly both a product and a process innovator. The same is true for a firm that introduces a new
marketing method in order to market a new product, or a firm that adopts for the first time a new
organisational method in the course of the introduction of a new process technology.

4.1 Distinguishing between product and process innovations

190. The distinction between products and processes is clear for goods. The distinction may be less
clear for services however, as the production, delivery and consumption of many services can occur at the
same time. Some distinguishing guidelines are:

• If the innovation involves new or significantly improved characteristics of the service offered to
customers, it is a product innovation.

• If the innovation involves new or significantly improved methods, equipment and/or skills used
to perform the service, it is a process innovation.

38
 • If the innovation involves significant improvements in both the characteristics of the service
offered and in the methods, equipment and/or skills used to perform the service, it is both a
product and a process innovation.

There can be many cases where a service innovation is only of one type. For example, firms can offer a
new service or new characteristics of a service without significantly changing the methods to provide the
service. Likewise, significant process improvements, for instance to reduce delivery costs, could make no
difference to the characteristics of the service that is sold to customers.

4.2 Distinguishing between product innovations and marketing innovations

191. The main distinguishing factor for product and marketing innovations is whether there is a
significant change in the product’s functions or uses. Goods or services which have significantly improved
functional or user characteristics compared to existing products are product innovations. On the other hand,
the adoption of a new marketing concept that involves a significant change in the product design of an
existing product is a marketing innovation but not a product innovation, as long as the functional or user
characteristics of the product are not significantly changed. As an example, new clothes produced using
new fabrics with improved performance (breathable, waterproof, etc.) are product innovations, but the
first-time introduction of a new shape for clothes intended to approach a new group of customers or to give
the product a higher degree of exclusivity (and thus allow for a higher mark-up compared to the previous
version of the product), is a marketing innovation.

192. Innovations can in some cases be considered both product and marketing innovations, if firms
implement changes to existing products that both involve significant changes in the functions or uses of the
product and significant changes in the product’s form and appearance or packaging that are part of a new
marketing concept.

4.3 Distinguishing between service (product) innovations and marketing innovations

193. The main distinguishing factor for service innovations and marketing innovations is whether the
innovation involves a marketing method or a service (i.e. a product). Firms will generally be able to
distinguish between their sales/marketing methods and their products.

194. This distinction can depend on the nature of the firm’s business. An example is an innovation
involving internet sales. For a firm that produces and sells goods, the first-time introduction of e-commerce
is a marketing innovation in product placement. Firms that are in the business of e-commerce
(e.g. ‘auction’ firms, website providers where other firms can advertise or sell their products, firms
arranging the sale of travel tickets, etc.) are offering ‘sales services’. For these firms, a significant change
in the characteristics or capabilities of their website is a product (service) innovation.

195. Some innovations are both product and marketing innovations. This can, for example, be the case
if a firm implements a new sales and customer service operation, introducing both a new way of marketing
its products (direct selling) and also offering additional services (e.g. repair) and product information to its
customers.

4.4 Distinguishing between process and marketing innovations

196. Both process and marketing innovations can involve new methods of moving information or
goods, but their purposes differ. Process innovations involve production and delivery methods and other

39
ancillary support activities aimed at decreasing unit costs or increasing product quality, while marketing
innovations aim at increasing sales volumes or product prices, the latter through changes in product
positioning or reputation.

197. Borderline cases can arise for marketing innovations that involve the introduction of new sales
channels. For example, innovations that involve the introduction of a new sales channel (i.e. a new way of
selling goods and services to customers) can also include the implementation of new logistics methods
(i.e. the transport, storage and handling of products). If these innovations are aimed both at increasing
sales and reducing unit distribution costs, then they should be considered both a process and a marketing
innovation.

4.5 Distinguishing between process and organisational innovations

198. Distinguishing between process and organisational innovations is perhaps the most frequent
borderline case in innovation surveys since both types of innovation attempt - among others - to decrease
costs through new and more efficient concepts of production, delivery and internal organisation. Many
innovations will thus contain aspects of both innovation types. For example, the introduction of new
processes may also involve the first-time use of new organisational methods such as group working.
Organisational innovations such as the first-time introduction of a total quality management system may
involve significant improvements in production methods such as new production logistic systems to avoid
certain types of flaws or new and more efficient information systems based on new software and new ICT
equipment.

199. A starting point for distinguishing process and/or organisational innovations is the type of
activity: process innovations deal mainly with the implementation of new equipment, software and specific
techniques or procedures, while organisational innovations deal primarily with people and the organisation
of work. Guidelines for distinguishing the two in borderline cases are as follows:

• If the innovation involves new or significantly improved production or supply methods that are
intended to decrease unit costs or increase product quality, then it is a process innovation.

• If the innovation involves the first-time use of new organisational methods in the firm’s business
practices, workplace organisation or external relations, it is an organisational innovation.

• If the innovation involves both new or significantly improved production or supply methods and
the first-time use of organisation methods, it is both a process and an organisational innovation.

4.6 Distinguishing between marketing and organisational innovations

200. Borderline cases can arise here for innovations that involve the first-time introduction of both
marketing methods and organisational methods. As noted above, if an innovation has characteristics of
both types of innovations, then it is both a marketing innovation and an organisational innovation.
However, organisational innovations that involve sales activities (e.g. the integration of sales with other
departments), but do not involve the introduction of new marketing methods, are not marketing
innovations.

40
5. CHANGES THAT ARE NOT CONSIDERED INNOVATIONS

a) Ceasing to use a process, a marketing method, an organisation method, or to market a product

201. Stopping doing something is not an innovation, even if it improves a firm’s performance. For
example, it is not an innovation when a television manufacturer ceases to produce and sell a combined
television and DVD player, or a property development agency or construction company stops building
retirement villages. Stopping the use of a certain marketing or organisation method is also not an
innovation.

b) Simple capital replacement or extension

202. The purchase of identical models of installed equipment, or minor extensions and updates to
existing equipment or software, are not process innovations. New equipment or extensions must both be
new to the firm and involve a significant improvement in its specifications.

c) Changes resulting purely from changes in factor prices

203. A change in the price of a product or in the productivity of a process resulting exclusively from
changes in the price of factors of production is not an innovation. For example, an innovation does not
occur when the same model of PC is constructed and sold at a lower price simply because the price of
computer chips falls.

d) Customisation

204. Firms engaged in custom production make single and often complex items according to
customers’ orders. Unless the one-off item displays significantly different attributes to products that the
firm has previously made, it is not a product innovation.

205. Note that the above concerns product changes from customisation and not the implementation of
customised production itself. For example, the integration of production, sales and delivery operations is an
organisational innovation.

e) Regular seasonal and other cyclical changes

206. In certain industries such as clothing and footwear there are seasonal changes in the type of goods
or services provided which can be accompanied by changes in the appearance of the products concerned.
These types of routine changes in design are generally neither product nor marketing innovations. For
example, the introduction of the new season’s anoraks by a clothing manufacturer is not a product
innovation unless, for example, they have a lining with significantly improved characteristics. However, if
the occasion of seasonal changes is used for a fundamental change in product design that is part of a new
marketing approach used for the first time by the firm, then they should be considered a marketing
innovation.

41
f) Trading of new or significantly improved products

207. The situation for new products is complicated in the goods-handling services and distributive
trades (wholesale and retail distribution, transport and storage). Trading of new or improved products is
generally not a product innovation for the wholesaler, retail outlet, or transport and storage firm. However,
if such a firm begins to deal with a new line of goods (i.e. types of goods that the firm has not previously
sold) then this activity is to be considered a product innovation, as the firm is offering a new service.

6. NOVELTY AND DIFFUSION

208. By definition, all innovations must contain a degree of novelty. Three concepts for the novelty of
innovations are discussed below: new to the firm, new to the market, and new to the world.

209. Also related to novelty and diffusion are who developed the innovations, which is discussed in
Chapter 5. This establishes whether innovations are mainly developed by enterprises themselves or in
co-operation with other enterprises or public research institutions, or if innovations are mainly developed
outside the enterprise.

210. As noted above, the minimum entry level for an innovation is that it must be new to the firm. A
product, process, marketing method, or organisational method can already have been implemented by other
firms, but if it is new to the firm (or in case of products and processes: significantly improved), then it is an
innovation for that firm29.

211. The concepts new to the market and new to the world concern whether or not a certain innovation
has already been implemented by other firms, or whether the firm is the first in the market or industry or
worldwide to have implemented this innovation. Firms that are the first to develop innovations can be
considered drivers of the process of innovation. It is from these firms that many new ideas and knowledge
originate, but the economic impact of their innovations will depend on the adoption of these innovations by
other firms. Information on the degree of novelty can be used to identify the developers and adopters of
innovations, to examine patterns of diffusion, and to identify market leaders and followers.

212. Innovations are new to the market when the firm is the first to introduce the innovation onto its
market. The market is simply defined as the firm and its competitors and can include a geographic region
or product line. The geographical scope of new to the market is thus subject to the firm’s own view of its
operating market and thus can include both domestic and international firms.

213. An innovation is new to the world when the firm is the first to introduce the innovation for all
markets and industries, domestic and international. New to the world thus implies a qualitatively greater
degree of novelty than new to the market. While many surveys might find that questions on new to the
market are sufficient to examine the degree of novelty for innovations, new to the world provides an option
for those surveys that wish to examine novelty in greater detail.

214. A related concept is a radical or disruptive innovation which can be defined as an innovation that
has a significant impact on a market and on the economic activity of firms in that market. This concept
thus focuses on the impact of innovations as opposed to their novelty. These impacts can, for example,

29
By ‘firm’ is meant, the statistical unit for which data is compiled, which in general is the enterprise. New to
the firm thus implies new to the statistical unit.

42
change the structure of the market, create new markets, or render existing products obsolete30. However, it
might not be apparent whether an innovation is disruptive until long after the innovation has been
introduced. This makes it difficult to collect data on disruptive innovations within the period reviewed in
an innovation survey.

7. THE INNOVATIVE FIRM

215. The innovative status of a firm can be defined in several ways. The basic definition of an
innovative firm (see section 2 of this Chapter) is a firm that has implemented at least one innovation, while
a product-process innovator is defined as a firm that has implemented either a product or a process
innovation.

216. A number of further ways of classifying an innovative firm are possible, depending on policy or
research needs. These classifications can be used to specify the percentage of firms (by size class, sector,
country or other factor) that introduce each of the four types of innovations, or the share of firms that have
implemented combinations of innovations, such as product and marketing innovations or process and
organisational innovations. The classification by innovative status can also include other information, such
as on who developed the innovation. For instance, this information could be used to identify firms that only
adopt product and process innovations developed by other firms.

217. Firms can have innovation activities in the period under review without having actually
implemented an innovation. All activities involved in the development or implementation of innovations,
including innovations planned for implementation in the future, are innovation activities (see Section 2
above). During a given period, innovation activities can be of three kinds:

• Successful in having resulted in the implementation of an innovation (although the innovation

need not have been commercially successful).

• Ongoing, work in progress, which has not yet resulted in the implementation of an innovation.

• Abandoned before the implementation of an innovation.

Innovation activities are specified in detail in Chapter 6.

218. An innovation active firm is one that has had innovation activities during the period under
review, including those with ongoing and abandoned activities. In other words, firms that have had
innovation activities during the period under review, regardless of whether the activity resulted in the
implementation of an innovation, are innovation active.

219. A number of firms can have come into existence during the period under review. This includes
both newly established firms and firms that are the result of mergers, demergers and other kinds of
reorganisation. The innovation status of these firms, such as whether they are innovative or innovation
active, is defined in the same way as for all other firms.

30
See e.g. Christensen, C.M. (1997) The Innovator’s Dilemma: When New Technologies Cause Great Firms
to Fail. Boston: Harvard School Press.

43
8. COLLECTING DATA ON INNOVATIONS

220. Depending on survey objectives and scope, data collection on innovations can follow various
approaches. A comprehensive approach would be to cover all four types of innovations equally.
Alternative approaches would be to partly cover marketing and organisational innovations but maintain
product and process innovations as the core innovation types, or to focus exclusively on product and
process innovations. In addition, one or more types of innovation can be covered in greater detail through
specialised surveys.

221. The guidelines presented here and in subsequent chapters outline a number of options for data
collection. Clearly, coverage of all topics and subtopics in one innovation survey is not feasible, and
surveys will need to choose those questions that are deemed most relevant. A limited set of topics that are
considered particularly important for innovation surveys are recommended in bold type.

222. Data on each type of innovation can be collected through a single question or through a series of
sub-questions on separate subgroups of each type of innovation. The latter option would obtain more
detailed information on the innovations of each firm. This greater level of detail on what types of
innovations firms have implemented would be very useful for data analysis and interpretation.

223. Information on supplementary factors, such as on linkages, innovation objectives, and barriers to
innovation, can be collected separately for each innovation type, or for closely related subsets of the four
types, such as for product and process innovation combined. For some supplementary questions, collecting
data for all four types of innovation combined can lessen the ability to interpret the data. For example,
combining all four innovation types for questions on linkages would make it very difficult to determine if a
specific linkage was used to develop a product, process, marketing or organisational innovation.

224. Asking separate questions on each supplementary factor for each innovation type in innovation
surveys can be problematic, due to length constraints. Taking this into account, Chapters 5 to 7 outline
options for supplementary questions that follow a number of approaches, such as referring to all four types
of innovations combined, focussing on product and process innovations, or directing questions at
individual types of innovations. For Chapter 5 on linkages, guidelines are provided for directing questions
on linkages towards all four types (either combined or separately) or to a subset of types, such as product
and process innovations. Concerning Chapter 6, product and process innovation activities are separated
from activities for marketing and organisational innovations. Chapter 7 lists objectives, hindering factors
and other indicators for each individual type of innovation. These lists can thus be equally useful for
questions that focus solely on a subset of innovations, those that cover all innovation types, or those that
cover individual types of innovations.

225. Innovations that span more than one type of innovation and that frequently include a process and
organisational component could play an increasingly important role in firm competitiveness and in
productivity gains. For example, a restructuring of production operations could involve process,
organisational, and marketing innovations, or a marketing and organisational innovation could be
implemented in order to better profit from a product innovation.

226. One option is to include questions on the connections between different types of innovations. Of
particular interest is the link between organisational and process innovations, though it can also be of
interest whether other innovation types are connected, such as product and marketing innovations or
product and process innovations.

227. To ensure comparability between respondents, surveys must specify an observation period for
questions on innovation. It is recommended that the length of the observation period for innovation

44
surveys should not exceed three years nor be less than one year. Factors influencing the choice of the
length of the observation period are discussed in Chapter 8 on survey methods. The same observation
period should be used for all questions in innovation surveys, with the exception of a few quantitative
indicators that are difficult to collect and which should therefore refer to the most recent year of the
observation period, the reference year.

228. The concept of novelty is in principle applicable to all four types of innovations, though possibly
to differing degrees. Questions on novelty are likely to be easiest to answer for product innovations. Firms
could find questions on the novelty of process innovations more difficult to answer than for product
innovations, as they can lack full knowledge of other firms’ production and delivery methods. The concept
of novelty is applicable for most marketing innovations (such as new methods in contracting, product
placement and product promotion), though less relevant for new concepts of product design. For
organisational innovation, some new organisation methods may be specific to an individual firm,
complicating comparison with other firms, and firms can lack information on whether certain
organisational methods have been applied already by other firms.

229. It is recommended that innovation surveys ask whether any product innovations
introduced during the observation period were new to the market. Innovation surveys can also collect
data on new to the market process innovations. An additional option is to ask whether these types of
innovations were new to the world.

45
 Chapter 4

INSTITUTIONAL CLASSIFICATIONS

1. THE APPROACH

230. The institutional approach focuses on the characteristic properties of the innovative firm. All
characteristics of innovation activities, and their inputs and outputs, are classified to one class or subclass
according to the unit’s principal activity.

2. THE UNITS

231. A clear distinction should be made between the reporting unit, the observation unit and the
statistical unit. The reporting unit is the entity from which the recommended items of data are collected.
They may vary from sector to sector and from country to country, depending on institutional structures, the
legal status of data collection, tradition, national priorities and survey resources. It is therefore difficult to
make international recommendations about the reporting unit for innovation surveys. However, whenever
countries provide statistics for international comparisons, the reporting units should be specified.

232. The observation unit is the entity for which the received data refer to. The observation unit is
equivalent to the reporting unit if data received refers to the same unit as the reporting unit31. The
statistical unit may be an observation unit on which information is received and statistics are compiled, or
an analytical unit which statisticians create by splitting or combining observation units with the help of
estimations or imputations in order to supply more detailed and/or homogeneous data than would
otherwise be possible.

233. As far as possible, the statistical unit should be uniform for all countries. This goal may,
however, be difficult to achieve in practice. One reason is that legal structures are different from country to
country. Definitions of units, such as the enterprise, may also vary across countries. Another reason is the
interactions of the statistical unit with the observation or reporting unit. If the reporting or the observation
unit is larger than the statistical unit, there may be problems in distributing the data into the appropriate
statistical units.

234. Ideally, innovation data should be compiled (and collected) at the organisational level for
which decisions on innovation activity are made. Taking into account how innovation activities are
usually organised, the enterprise is in general the most appropriate statistical unit. However, there is
no single definition of an enterprise that is used in all countries. The two main definitions are those of the
ISIC and the EU. According to ISIC Rev. 3.1 §§ 54-55, an enterprise has “autonomy in respect of financial
and investment decision-making, as well as authority and responsibility for allocating resources for the
production of goods and services. It may be engaged in one or many productive activities. The enterprise is
31
They are not the same if, for example, a questionnaire is sent to an enterprise (reporting unit), but the data
are reported individually from each division of the enterprise (observation units).

46
the level at which financial and balance sheet accounts are maintained and from which international
transactions, and international investment position (when applicable) and the consolidated financial
position can be derived.” The EU definition of an enterprise is somewhat more narrowly defined: “The
enterprise is the smallest combination of legal units that is an organisational unit producing goods or
services, which benefits from a certain degree of autonomy in decision-making, especially for the
allocation of its current resources. An enterprise carries out one or more activities at one or more
locations.”32

235. While these two definitions are very similar, a central difference between the two is that a group
of enterprises is not an enterprise according to the EU definition (since it does not comprise the smallest
combination of legal units that is an organisational unit), while an enterprise group still falls within the
ISIC definition of an enterprise. Common characteristics are that enterprises exercise a certain degree of
autonomy in decision-making and have full financial accounts.

2.1 The primary statistical unit

236. The enterprise unit is the appropriate primary statistical unit in innovation surveys in most
cases. The enterprise unit should, however, not be confused with the entity “legal unit”. While legal units
are independent in a legal sense, they may not necessarily constitute independent economic entities with
decision-making autonomy for their productive activities. This point also follows from the EU definition of
an enterprise and ISIC Rev 3.1 §4933. This is important for samples that are drawn from business registers
based on legal units, as data from legal units without decision-making autonomy may not be comparable
with data from enterprises. If data are collected for legal units, then it is desirable to compile this data for
use in statistics at the enterprise level.

237. The enterprise is generally the most appropriate statistical unit, including:

• Enterprises that consist of a single legal unit that engages primarily in one kind of economic
activity.

• Enterprises that are a group of legal units, where the individual legal units cannot be considered
separate economic entities, including:

− Legal units that are vertically or horizontally integrated in the enterprise.

− Individual legal units that perform ancillary functions, including R&D.

238. There are some cases that require greater clarification. These include: 1) enterprise groups,
2) large enterprises that may have several areas of activity, 3) multinational enterprises and groups.

239. An enterprise group is an association of enterprises bound together by legal and/or financial
links. For such groups, the question of whether innovation data should be compiled at the group level or
for each individual enterprise depends on what level decisions on innovation activity are made. Here, if

32
Council Regulation (EEC) No 696/93 of 15 March 1993 on the statistical units for the observation and
analysis of the production system in the Community, OJ No L 76, p.1, section III/A of the annex.
33
I.e. “…the smallest combination of legal units that is an organisational unit…”. ISIC Rev. 3.1 §49: “In
such cases, for statistical purposes it is inappropriate and unnecessary to regard each legal entity as a
separate institutional unit.”

47
each individual enterprise unit has decision-making autonomy concerning innovation, it is preferable to
collect and compile data at the enterprise level as opposed to the group level.

240. Large enterprises may have a number of different productive activities. For the largest of these
enterprises, decision-making on innovation activity will likely not be made at the highest level of the
organisation, but instead will be undertaken for each productive activity or division. In this case, it may be
preferable, where possible, to collect and compile data at the kind of activity unit level (KAU), defined as
“an enterprise or part of an enterprise which engages in one kind of economic activity without being
restricted to the geographic area in which that activity is carried out”34. This means that the KAU may
consist of one or more legal units, or a part of a legal unit.

241. Multinational enterprises (MNE) present a number of challenges, given that many activities may
take place across national borders. For example, innovation activities in MNEs may be undertaken jointly
by units in more than one country, and many activities may be segmented, with development activities in
one country and production and sales in another. Given that innovation surveys are national surveys, data
will be limited to that of domestic unit(s) of the MNE. However, it can be very useful to obtain as much
information as possible on connections of the domestic unit’s innovation activities with that of units
abroad. The following guidelines are suggested for treating MNEs:

• Only the domestic part of the multinational enterprise forms the statistical unit to be included,
regardless of where the enterprise’s headquarters is situated. Data can be collected on whether
observation units are part of MNEs (see below).

• The foreign units of a multinational should be regarded as a separate entities (not part of the
statistical unit, though belonging to the same enterprise group):

− Joint innovation activities between the domestic and foreign units of the MNE should be
considered as innovation co-operation between two enterprises in the same enterprise group.
Exchanges of information or purchases of knowledge and technology should also be treated
as transfers between two separate enterprises.

− R&D activity or other new knowledge and technology in MNEs that is conducted abroad
(and included in the accounts of the domestic unit) should be included in Extramural R&D
and Acquisitions of external knowledge, respectively (see Chapter 6).

− For questions on who developed an innovation, ‘other enterprises’ can be broken down into
those that are part of a MNE or enterprise group, and all other enterprises.

− Innovations developed by the foreign units of a multinational enterprise and adopted by the
domestic units are new-to-the-firm innovations.

− Innovations developed by the foreign units, but not adopted by the domestic units of a
multinational enterprise, should not be included.

34
Council Regulation (EEC) N° 696/93 of 15 March 1993, OJ N° L76 of the 3 March.

48
242. Guidelines for the primary statistical unit are summarised below:

The most appropriate primary statistical unit is the enterprise, including:

• Enterprises that are single legal units with one primary economic activity.

• Enterprises that are groups of legal units that cannot be considered separate economic entities.

• Enterprises that are part of an enterprise group, where decision-making on innovation takes place at the
enterprise level.

− In some cases for very large enterprises with more than one economic activity, the appropriate primary
statistical unit can be kind-of-activity units (KAU), with the KAU consisting of one or more legal units, or part of a
legal unit.

− For MNEs, the domestic units of multinational enterprises are the appropriate statistical units, which may either
be an enterprise or part of an enterprise (e.g. a legal unit).

243. Based on the discussions above, it is recommended to collect data on the institutional status
of the observation unit:

• Whether the observation unit is part of an enterprise or enterprise group and, if so, what its
function is, for example, headquarters, research centre, administrative centre, marketing, other.

• Whether the observation unit is part of a multinational enterprise and, if so, what its function is,
and where the headquarters is situated.

2.2 The secondary statistical unit

244. In some instances, there may be interest in collecting data at a lower organisational level than the
enterprise. This may for example be the case for compiling regional statistics or for employing a two-tiered
approach to data collection.

245. In such cases, an appropriate secondary statistical unit is the establishment unit, which can be
defined as (following ISIC Rev. 3.1): “an enterprise or part of an enterprise, that is situated in a single
location, and in which only a single (non-ancillary) productive activity is carried out or in which the
principal productive activity accounts for most of the value added”. An alternative is the local unit as
defined by the EU.35 The secondary statistical unit may be useful for larger enterprises that have operations
in more than one region.

246. For regional analysis, the establishment unit or similar units may be useful36 for collecting
innovation data. However, information on some variables should not be collected at the level of
establishment (or similar) units as they refer directly to the enterprise. An example is information on

35
The local unit as defined by the EU: "The local unit is an enterprise or part thereof (e.g. a workshop,
factory, warehouse, office, mine or depot) situated in a geographically identified place. At or from this
place economic activity is carried out for which - save for certain exceptions - one or more persons work
(even if only part-time) for one and the same enterprise." (Council Regulation (EEC) N° 696/93 of
15 March 1993, OJ N° L76 of the 3 March.)
36
For a detailed discussion of the problem of the local unit as the statistical unit in innovation surveys see
Eurostat (1996), The Regional Dimension of R&D and Innovation Statistics, particularly part B.

49
innovation objectives. They refer to strategic decisions at the enterprise level, and will rarely be taken by
establishment units.

247. In particular for large enterprises, decisions on innovation activities may be decentralised and it
may be difficult for one person to provide data on all innovation activities in the enterprise. An option in
this case is to use a two-tiered approach to data collection. There is more than one method for utilizing this
approach. One is to collect data at the establishment level and then compile this data at the enterprise level.
Regional or establishment level analysis can rely on the collected establishment level data. A drawback
here is that, as mentioned above, establishments may not be in a position to respond to all questions on
innovation. Another is to collect some innovation data at the enterprise level and other data at the
establishment level. How this method is conducted in practice may depend on the preferences of the
management of each enterprise.

248. Care should be taken in aggregating establishment results to the enterprise level. For example, the
introduction of a new technology can be an innovation for an establishment but not for the enterprise if it
already has been used elsewhere in the enterprise.

3. CLASSIFICATION BY MAIN ECONOMIC ACTIVITY

249. Statistical units of innovation surveys can be broken down by different classifications. The most
important classification is the principal economic activity of the statistical unit (“industry”). The
International Standard Industrial Classification (ISIC Rev. 3.1) and the statistical classification of
economic activities in the European Community (NACE Rev. 1.1) 37 respectively are appropriate
international classifications for this purpose. Countries that use a national industrial classification system
rather than ISIC Rev. 3.1 should use concordance tables to convert their industrially classified data to
ISIC Rev. 3.1.

250. The criteria for classification of statistical units by principal activity should be determined by
“the ISIC (NACE) class in which the principal activity, or range of activities, of the unit is included”38.
The principal activity is the ISIC class that accounts for the majority of the enterprise’s value-added from
its goods and services. If this is not possible, the principal activity can be determined either on the basis of
the gross output of the goods sold or services rendered in each ISIC class, by value of sales, or by
employment39.

251. The proposed classification list is presented in Table 1, which contains the basic arrangement of
the divisions, groups and classes of ISIC Rev. 3.1/ NACE Rev. 1.1 for the purpose of innovation statistics.
The table may be further split, or aggregated, for specific purposes.

37
Revisions of the International Standard Industrial Classification (ISIC Rev. 4) and the Statistical
Classification of Economic Activities in the European Community (NACE Rev. 2) are expected to be
completed in 2007. Industrial classifications should be modified accordingly upon the implementation of
these revisions.
38
See United Nations (2002), International Standard Industrial Classification of All Economic Activities,
Rev. 3.1, § 79.
39
See ISIC Rev. 3.1, § 80 (UN, 2002).

50
 Table 1. Industrial classification proposed for innovation surveys in the business enterprise sector based on
ISIC Rev.3.1 and NACE Rev. 1.1

Title ISIC Rev. 3.1 NACE Rev. 1.1

Division/Group/Class Division/Group/Class
MINING AND QUARRYING 10 to 14 10 to 14
MANUFACTURING 15 to 37 15 to 37
Food Products & Beverages 15 15
Tobacco Products 16 16
Textiles 17 17
Wearing Apparel & Fur 18 18
Leather Products & Footwear 19 19
Wood & Cork (not Furniture) 20 20
Pulp, Paper & Paper Products 21 21
Publishing, Printing & Reproduction of Recorded Media 22 22
Coke, Refined Petroleum Products & Nuclear Fuel 23 23
Chemicals & Chemical Products 24 24
Chemical Products less Pharmaceuticals 24 less 2423 24 less 24.4
Pharmaceuticals 2423 24.4
Rubber & Plastic Products 25 25
Non-metallic Mineral Products 26 26
Basic Metals 27 27
Basic Metals, Ferrous 271+2731 27.1 to 27.3+27.51/52
Basic Metals, Non-ferrous 272+2732 27.4+27.53/54
Fabricated Metal Products (except Machinery & Equipment) 28 28
Machinery n.e.c. 29 29
Office, Accounting & Computing Machinery 30 30
Electrical Machinery 31 31
Electronic Equipment (Radio, TV & Communications) 32 32
Electronic Components (includes Semiconductors) 321 32.1
Television, Radio & Communications Equipment 32 less 321 32 less 32.1
Medical, Precision & Optical Instruments, Watches, Clocks 33 33
(Instruments)
Motor Vehicles 34 34
Other Transport Equipment 35 35
Ships 351 35.1
Aerospace 353 35.3
Other Transport n.e.c. 352+359 35.2+35.4+35.5
Furniture, Other Manufacturing n.e.c. 36 36
Furniture 361 36.1
Other Manufacturing n.e.c. 369 36.2 to 36.6
Recycling 37 37
ELECTRICITY, GAS & WATER SUPPLY 40+41 40+41
CONSTRUCTION 45 45
MARKETED SERVICES 50 to 74 50 to 74
Sale, Retail, Maintenance & Repair of Motor Vehicles & 50 50
Motorcycles
Other Wholesale Trade 51 51
Other Retail Trade 52 52
Hotels & Restaurants 55 55
Land Transport & via Pipelines 60 60
Water Transport 61 61
Air Transport 62 62
Supporting & Auxiliary Transport Activities, Travel Agencies 63 63
Post & Telecommunications 64 64
Post 641 64.1
Telecommunications 642 64.2
Financial Intermediation 65 to 67 65 to 67
Real Estate, Renting 70+71 70+71
Computer & Related Activities 72 72
Software Consultancy & Supply 722 72.2
Other Computer Services n.e.c. 72 less 722 72 less 72.2
Research & Development40 73 73
Other Business Activities 74 74
Architectural, Engineering & other Technical Activities 742 74.2+74.3
Other Business Activities n.e.c. 74 less 742+743 74 less 74.2+74.3

40
Only enterprises in the business sector should be included, following the Frascati Manual §§163-168. For
this NACE/ISIC-group (73), data on the product field should also be collected, following the Frascati
Manual §272.

51
4. CLASSIFICATION BY SIZE

252. The other essential classification of statistical units for innovation surveys is by size. Though
different variables can be used to define the size of a statistical unit in innovation surveys, it is
recommended that size should be measured on the basis of the number of employees. This
recommendation is in line with similar proposals in other manuals in the Frascati family. Given the strata
requirements in sample surveys (see Chapter 8), and given that innovation activities other than R&D are
widely performed by small and medium-sized units, it is recommended that size classes include smaller
firms. In order to maintain international comparability while at the same time allowing flexibility in the
number of size classes, the following size classes are recommended as a minimum:

Classification of statistical units for innovation surveys by size

Number of employees:

10 – 49
50 – 249
250 and above

More detailed breakdowns by size class may also be used. This also includes a size class for firms with less
than ten employees. It is important that more detailed size classes are consistent with the above groups. A
proposal would be:

Classification of statistical units for innovation surveys by size – detailed:

0
1–9
10 – 49
50 – 99
100–- 249
250 – 499
500 – 999
1 000 – 4 999
5 000 and above.

5. OTHER CLASSIFICATIONS

5.1 Type of institution

253. A further useful classification of statistical units for innovation surveys might be by type of
institution. This breakdown seems particularly important when the statistical unit is in general the
enterprise, and in view of the increasing internationalisation of innovation activities. Taking into account
these considerations, and similar proposals in the Frascati Manual and the Handbook on Economic
Globalisation Indicators, it is recommended that when enterprises are the statistical units in innovation
surveys they should be classified as follows:

Classification of statistical units for innovation surveys by type of institution

• Private enterprise:

52
 − National (no Controlled Affiliates41 (CA) abroad).

− Multinational, of which there may be three types:

− Foreign-controlled42 affiliates (CAs) (where the affiliate does not control any other
affiliates abroad).

− Foreign-controlled affiliates with CAs (parent companies under foreign control).

− Parent companies with CAs abroad (parent company not under foreign control).

• Public enterprise43, “resident non-financial corporations and quasi-corporations that are subject to
control by government units, control over a corporation being defined as the ability to determine
general corporate policy by choosing appropriate directors, if necessary.”

5.2 Other

254. Many other types of classifications of statistical units in innovation surveys can be used for
analytical purposes. They include:

General enterprise characteristics:

• Form of activity, with the categories: capital intensive / labour intensive / knowledge intensive.

• Type of goods produced, with the categories: consumer goods / intermediate goods / investment
goods.

• Export intensity, the exports of the enterprise as a ratio of sales revenue/turnover44.

• Geographic location.

Innovation indicators:

• Innovation or R&D intensity, the ratio between innovation expenditure (or R&D expenditure)
and turnover.

• Co-operation, with other enterprises / public institutions.

41
A controlled affiliate is an enterprise that is directly or indirectly controlled by a parent company. See
OECD (2005), Handbook on Economic Globalisation Indicators.
42
Control is defined in the Handbook on Economic Globalisation Indicators as having more than 50%
ownership or control of more than 50% of voting shares.
43
OECD (2002), Proposed Standard Practice for Surveys for Research and Experimental Development,
Frascati Manual 2002, Paris, §180.
44
Turnover or sales revenue is the total amount of money that the firm has earned from the sales of all its
products during a given time period.

53
 Chapter 5

LINKAGES IN THE INNOVATION PROCESS

1. INTRODUCTION

255. The innovative activities of an enterprise45 partly depend on the variety and structure of its links
to sources of information, knowledge, technologies, practices, and human and financial resources.
Linkages act as sources of knowledge and technology for an enterprise’s innovation activity, ranging from
passive sources of information to suppliers of embodied and disembodied knowledge and technology to
cooperative partnerships. This chapter makes recommendations on how to measure these linkages, with a
focus on links to external sources outside of the enterprise. Linkages can be related to any of the four types
on innovations (i.e. product, process, marketing or organisational). This chapter is designed to provide
guidelines both for measuring linkages for individual or subsets of innovation types (such as product and
process innovations) and for all innovation types combined.

256. Each linkage connects the innovating enterprise to other actors in the innovation system:
government laboratories, universities, policy departments, regulators, competitors, suppliers, and
customers. Innovation surveys can obtain information on the prevalence and importance of different types
of linkages, plus the factors that influence the use of specific linkages. Identifying the linkages in
innovation activity provides evidence of the complexity of the system, but stops short of providing the
information necessary for a dynamic model of the system, with positive and negative feedback loops and
non-linear outcomes resulting from change. However, this information can make valuable contributions to
our understanding of innovation systems and can help determine the influence of government programmes
to encourage greater knowledge sharing or technology diffusion.

257. Linkages may depend on the nature of the enterprise and its market46. For example, the
innovative activities of an enterprise operating in a stable mature sector will be driven by the value of its
turnover and the cost of its inputs. Under these circumstances, the enterprise might focus on incremental
innovation and its principal links could be with suppliers and market signals from its customers. In a more
volatile environment, the enterprise might need to rapidly introduce new products, seek new markets, and
introduce new technologies, production methods, and organizational practices. The enterprise may develop
multiple linkages to provide new information, knowledge, technologies, production practices, and human
and financial resources. In each of these cases, information on linkages shows how the enterprise responds
to its business environment.

258. Linkages vary by source (who or what the link is with), cost (the amount of investment required),
and level of interaction (the direction of information flows and the level of interpersonal contact). Some
external sources can consistently provide information at a nominal cost, such as patent disclosures or
publications, while other sources such as consultants are usually costly. The level of interaction of a
linkage influences the characteristics of the information or knowledge that can be obtained. Less

45
Throughout the remainder of this Manual, the use of the term enterprise should be understood as referring
to the primary statistical unit.
46
See for example, M. Dierkes, 2003, “Visions, Technology, and Organizational Knowledge: An Analysis of
the Interplay between Enabling Factors and Triggers of Knowledge Generation”, in John de la Mothe and
Dominique Foray (eds), Knowledge Management in the Innovation Process, Boston: Kluwer Academic
Publishers, pp. 9-42.

54
interactive linkages that require no interpersonal contact and which are based on one-way information
flows, such as reading publications or searching patent databases, can only provide codified information.
Conversely, highly interactive linkages involving close working relationships, such as with a supplier, can
provide both codified information and tacit knowledge and real-time assistance with solving problems.
However, enterprises may avoid some types of high complexity links due to concerns over the loss of
intellectual property.

259. A linkage can be internal to the unit or external, depending on how the unit is defined (see
Chapter 4). Even if business units are formally organised as separate enterprises, they may belong to the
same enterprise group. Units may be part of multinational enterprises so that within-enterprise links cross
national boundaries. Enterprises that belong to marketing chains (for instance clothing) or highly integrated
value chains could view linkages with other enterprises in the chain more as internal than external linkages.

260. The inflow of knowledge and technology is one side of what is often referred to as diffusion.
Diffusion also involves outflows from the innovating enterprise. Outbound diffusion is relevant both for
identifying the economic effects of innovation and for establishing the shape of an enterprise’s network. As
with highly interactive linkages, outbound diffusion is influenced by concern over knowledge leakages and
the methods enterprises use to protect their intellectual property.

261. Also related to linkages are questions on who developed innovations. These questions establish
whether innovations are mainly developed by enterprises themselves or in co-operation with other
enterprises or public research institutions, or if innovations are mainly developed outside the enterprise.

262. Benefits of linkages will depend on how well knowledge is shared throughout the enterprise and
channelled into the development of new products, processes and other innovations. Knowledge
management involves practices for gaining external knowledge and interacting with other organisations,
and for sharing and utilising knowledge within the enterprise.

263. Trust, values and norms can have an important impact on the functioning of external relations
and on the exchange of knowledge within the enterprise. Given this, building social capital may be a vital
part of enterprises’ innovation strategies. The term social capital has many meanings outside of economic
analysis which can lead to confusion. Network capital has been used as an alternative.

264. Three additional topics of relevance to linkages are covered in Chapter 7. Several hampering
factors for innovation activity concern linkages, such as access to information and opportunities for
co-operation. The section on the objectives and effects of innovation includes improving the capture and
use of knowledge. Firms may use different protection methods to control the flow of knowledge to other
firms. These will be relevant to some types of highly interactive linkages.

2. INBOUND DIFFUSION

265. Diffusion is the spread of innovations, through market or non-market channels, from first
implementation anywhere in the world, to other countries and regions and to other markets and firms. The
diffusion process will often involve more than the mere adoption of knowledge and technology, as
adopting enterprises learn from and build on the new knowledge and technology. Through this diffusion
process, innovations can change and supply feedback to the original innovator.

266. Central to understanding linkages in the innovation process are identifying how transfers of
knowledge and technology take place, what the main sources of knowledge and technology flows are for
enterprises, and which of these are of greatest importance. In addition to providing a better understanding
of diffusion processes and a mapping of linkages and knowledge flows, they are of direct relevance for

55
innovation policy. For example, should policy focus on promoting active co-operation and, if so, which
types of partners are most important? Or, are knowledge and technology flows of greater importance
through networks and other informal arrangements that do not involve active co-operation?

267. This section discusses three types of linkages or flows of knowledge and technology to
enterprises: open information sources that do not involve purchases of knowledge and technology or
interaction with the source; purchases or acquisitions of knowledge and technology; and innovation
co-operation.

2.1 Types of linkages

268. Open information sources provide access to knowledge without payment for the knowledge
itself, although there may be marginal fees for access to information, such as for membership in trade
associations, attendance for conferences, and subscriptions to journals. Open information sources do not
provide access to knowledge embodied in machinery or equipment or the right to use knowledge protected
by patents and other forms of intellectual property, although the knowledge behind the patent can be
accessed through patent databases. Some open sources, such as attendance at fairs or exhibitions, can
provide access to some tacit knowledge based on personal interactions with other participants.

269. Codified knowledge can take many forms, such as published articles, standards, metrology
(methods of measuring items such as liquid or gas flow, time, chemical pollutants, etc.), or knowledge
gained from networks, arms length contact with suppliers, or trade fairs.

270. Some information, such as feedback from clients or suppliers, may be very easy to use. The
utilisation of other information, for example from universities, may be more difficult and depend greatly on
employee capabilities. An additional challenge in using information sources is locating them. Enterprises
may have limited knowledge on potential information sources. Data on information sources can aid in
designing policy initiatives concerning training, improving ICT capabilities, and supporting the
establishment of networks and support services.

271. Knowledge networks facilitate the exchange of technology and commercial information.
Informal networks will tend to be based on personal contacts or ‘communities of practice’ or simply arise
in the normal course of business. Formal or managed networks can be organised by business organisations
such as chambers of commerce, research associations, technology services companies, consultants,
universities or public research organisations or sponsored by local, regional or central governments.

272. Acquisitions of technology and knowledge involve the purchase of external knowledge and
technology without active co-operation with the source. This external knowledge can be embodied in the
form of machinery or equipment that incorporates this knowledge. It can also include the hiring of new
employees that possess the new knowledge, or the use of contract research and consulting services.
Disembodied technology or knowledge also include other know how, patents, licenses, trademarks and
software.

273. Data on sources of acquisitions provide information on flows of knowledge and technology and
on where (in terms of regions, industries) these acquisitions are most prevalent. Identifying the importance
of purchases of knowledge and technology are also useful in motivating further analysis of how well
‘trade’ in knowledge and technology functions.

274. Innovation co-operation involves active participation in joint innovation projects with other
organisations. Other organisations may be either other enterprises or non-commercial institutions. Both
partners do not need to derive immediate commercial benefit from the venture. Pure contracting out of

56
work, where there is no active collaboration, is not regarded as co-operation. Co-operation is distinct from
open information sources and acquisitions of knowledge and technology in that all parties take active part
in the work.

275. Innovation co-operation allows enterprises to access knowledge and technology that they would
not be able to utilise on their own. There is also great potential for synergies in co-operation as partners
learn from each other.

276. Innovation co-operation can take place along supply chains where customers and suppliers are
involved in the joint development of new products, processes or other innovations. The level of interaction
(i.e. whether linkages involve co-operation, or arms length exchanges of information or purchases of
technology) along supply chains can depend on the type of knowledge and technology. For example,
concerning product development, if the technology is non-modular then innovation along the supply chain
must be closely coordinated because changes in the technological configuration of one part of a product
must take account of changes in any of the others. If technologies involved are completely modular then
the assemblers of the final product can deal with suppliers of components, materials, etc. on an arms length
basis, where interactions mainly consist of purchases of equipment or services embodied with the new
knowledge. Exchange of technological and business information naturally accompanies trade in goods and
services. Information on customer needs and their experience of a supplier’s products plays a key role in
innovation.

277. Innovation co-operation can also involve horizontal collaborations, with enterprises working
jointly with other enterprises or public research institutions. Examples are the joint development of new
technologies, products or processes by enterprises that while they sell the same type of product, possess
different complementary assets, e.g. sell into different geographical markets or different market niches.
Horizontal innovation co-operation can also include strategic marketing alliances to develop and
implement new marketing concepts. It can also take place between enterprises that produce different but
highly complementary products, e.g. a new computer controlled machine tool and the software package
needed to monitor and control it.

278. While the focus of this chapter is on external linkages, the internal sourcing of information is
also important. The identification of which parts of the enterprise (e.g. R&D, marketing, production, or
distribution) are important sources of information for innovation activities provides information on the
flow of knowledge within the enterprise.

279. Potential sources are similar for the three types of linkages, though there are some sources that
are only relevant as openly available sources of knowledge and technology. The list below shows sources
for all three types of linkages and indicates for which types each source is relevant.

280. The list includes internal sources, other enterprises, public and non-profit research institutions,
and a number of general information sources. The definition of several sources must be adapted to
country-specific terminology to clearly differentiate between commercial laboratories, government
research institutes and private non-profit research institutes.

281. Multinational enterprises (MNE) comprise a special case. Interactions within the MNE can take
place across countries. Given that statistical units for innovation surveys only include the domestic part of
MNEs (see Chapter 4), interactions with foreign units of the MNE should be considered as external
linkages, with ‘other enterprises within an enterprise group’.

57
 Sources for transfers of knowledge and technology

Open Sources for Co-operation

information purchases of partners
sources knowledge and
technology
Internal sources within the enterprise: *
R&D *
Production *
Marketing *
Distribution *
Other enterprises within the enterprise group * * *
External market and commercial sources:
Competitors * * *
Other enterprises in the industry * * *
Clients or customers * *
Consultants/consultancy firms * *
Suppliers of equipment, materials, * * *
components, software or services
Commercial laboratories * * *
Public sector sources:
Universities and other higher education * * *
institutions
Government/public research institutes * * *
Private non profit research institutes * * *
Specialised public/semi-public innovation * * *
support services
General information sources:
Patent disclosures *
Professional conferences, meetings, branch *
literature and journals
Fairs and exhibitions *
Professional associations, trade unions *
Other local associations *
Informal contacts or networks *
Standards or standardisation agencies *
Public regulations (i.e. environment, security) *

58
2.2 Collecting data on linkages in the innovation process

282. It is recommended that data be collected on all three types of linkages, drawing on the list
of sources above. For use in innovation surveys, these types of linkages can be defined as:

• Open information sources: openly available information that does not require the purchase of
technology or intellectual property rights, or interaction with the source.

• Acquisition of knowledge and technology: purchases of external knowledge and/or knowledge

and technology embodied in capital goods (machinery, equipment, software) and services, which
do not involve interaction with the source.

• Innovation co-operation: active co-operation with other enterprises or public research

institutions for innovation activities (which may include purchases of knowledge and technology).

283. Designing questions on types of linkages in innovation surveys presents a number of challenges.
While all three types of linkages are of interest, separate questions on each type may result in large
response burdens and, given the questions’ similarity, may be very tiresome for enterprises to answer. The
paragraphs below discuss a number of relevant aspects and some options for covering these topics in
innovation surveys.

284. Linkages can generate knowledge and technology for any of the four types of innovations
(i.e. product, process, organisational and marketing innovations). A large share of the interactions will
involve the development of new products or processes; however linkages can, in many cases, also involve
product design, the development of new marketing techniques, or work on organisational innovations such
as the integration of enterprises with customers, suppliers or retailers.

285. Questions on linkages can refer to all innovation types combined, to individual types, or to
closely related subsets of the four types, such as product and process innovations. The ability to
identify which type of innovation a linkage primarily relates to (e.g. work on individual types of
innovations, or a subset of types such as product and process innovations) can aid greatly in interpreting
the data. For example, linkages involving the development of a new good or service can be greatly
different from linkages involving the development of new marketing methods.

286. Questions on linkages can either utilise a binary scale (i.e. yes/no) or an ordinal scale in
which enterprises are asked whether they have utilised the source and, if so, its importance. An
ordinal scale is useful for identifying which sources are most important (see also the discussion of binary
and ordinal scales in Chapter 8). However, use of an ordinal scale may also limit the options available for
designing questions on linkages.

287. Some information on acquisitions of knowledge and technology can be obtained from questions
on innovation activities (see Chapter 6), though these questions do not ask for the source of the purchase.

288. In order to detect and better understand the process of clustering or networking in the field
of innovation, additional information can be obtained by asking for the geographical location of
co-operation partners (local, national, foreign by region or country). Information on the geographic
location of sources (domestic or foreign) may also be useful concerning open information sources and
acquisitions of knowledge and technology.

59
289. In order to better interpret results on linkages, questions can be asked about the enterprise’s status
as part of an enterprise group, and its position in a value-chain.

2.2.1 Options for designing linkages questions for innovation surveys

290. Four main factors have been identified that innovation surveys can take into consideration in
designing questions on linkages: types of linkages, reference to innovation types, use of binary or ordinal
scales, and the geographical location of linkages. In order to provide some additional guidance on this,
three options are outlined below.

291. One option for designing linkages questions for innovation surveys is to include a combined
question that asks whether sources are relevant as information sources, sources of purchases of knowledge
and technology, or co-operation partners. This allows the inclusion of all three types of linkages and
eliminates repetition. For this option, it would only be feasible to utilise a binary scale (yes/no). This
question could either refer to product and process innovation or to all innovation types. However,
restricting the question to product and process innovation (as opposed to all innovation types combined)
would aid in interpreting the data. Supplemental questions could also be asked whether enterprises have
had linkages (e.g. co-operative partners, information sources, though without specifying specific types of
sources) concerning each individual innovation type. This can also be supplemented by a question on the
geographic location of enterprises’ linkages.

292. A second option that has been used in a number of innovation surveys is to include two separate
questions on linkages, one on information sources and their relative importance, and one on co-operation
partners, their relative importance, and their location. In using this option, it is important to distinguish
between information sources and co-operation partners (e.g. if no guidelines are given, then any
co-operation partner will also be considered an open information source). Advantages with this approach
are that surveys can ask for the relative importance of each source, and the geographic location of
co-operation partners. Drawbacks are that acquisitions of knowledge and technology are not covered (aside
from information from questions on innovation activity) and even with their exclusion, there is a great deal
of repetition in the two questions. As with the option above, these questions could either refer to product
and process innovation or to all innovation types, and the same supplemental questions could be asked here
as well.

293. A third option would be to include two separate questions on linkages, one on information
sources and one on co-operation partners, as above. However, instead of asking on the relative importance
of linkages, innovation surveys could ask (using a binary scale) which types of innovation(s) (i.e. product,
process, organisational, marketing) each linkage involves. The main advantage of this option is that it
allows more detailed information on what type of innovation each linkage concerns.

2.2.2 Who developed the innovation

294. Additional information can be obtained on diffusion by collecting data on who developed the
enterprise’s innovations. Questions on who developed innovations have been included in a number of
surveys. This can give an indication of how active the enterprise has been in developing its innovations,
whether it interacted with other enterprises in their development, or whether the development of the
innovation was primarily conducted outside the enterprise.

295. This provides different information than questions on the degree of novelty (see Chapter 3), as
enterprises may be developing innovations that already have been implemented by other enterprises. It thus

60
provides information on how innovative enterprises are, though not necessarily on how novel their
innovations are.

296. It is recommended that questions are asked on who developed enterprises’ innovations.
Enterprises can, for example, be asked for each type of innovation whether:

• The innovations were mainly developed by the enterprise itself.

• The innovations were developed by the enterprise in co-operation with other enterprises or
institutions.

• The innovations were mainly developed by other enterprises or institutions.

The last two categories can also be broken down into subcategories, for example by distinguishing between
innovations developed in co-operation with other firms and in co-operation with public research
institutions, or whether other enterprises are part of the same multinational enterprise or enterprise group.
As with questions on degree of novelty, questions on who developed innovations can be asked for all four
types of innovations or for a subset of them.

2.3 Other linkages indicators

297. This subsection discusses options for collecting additional data on linkages, which are primarily
relevant for specialised surveys.

2.3.1 Types of knowledge and methods of transfer

298. In addition to identifying types of linkages and their sources, it may be useful to collect more
detailed information on important characteristics of linkages, such as the kinds of knowledge that are
transferred and the method of knowledge transfer. Questions of this type will likely require a specialised
survey that may be possible to link to the results of general innovation surveys. Alternatively,
supplementary questions can be included in general innovation surveys, for example on the most important
external linkage47.

299. Questions on the kinds of knowledge that are obtained from a linkage can ask if the knowledge
was embodied or disembodied, tacit or codified, public or private, whether it is R&D-based, specific or
generic, and on its degree of novelty. Questions on how the actual transfer takes place can ask about the
use of written reports, blueprints, purchases of machinery, components and software; informal contacts,
working together, training, and presentations.

2.3.2 Social or network capital

300. Social or network capital refers to the stocks of social trust, values and norms that enterprises
possess. These have important impacts on the circulation of information within an enterprise and on
sharing knowledge in collaborative activities with other organisations. Enterprises may implement new
organisational structures or new practices to introduce a new business culture, norms and values, with the

47
See for example, the DISKO-surveys on innovation in Denmark and Norway, and OECD (2001),
Innovative Networks: Co-operation in National Innovation Systems.

61
objective of improving the enterprise’s capacity to innovate. The establishment of trust is also a key factor
in maintaining and improving relationships, both within and outside the enterprise. Long-term relationships
that can build up mutual trust will likely be more beneficial for all participants.

301. Information can be gained on enterprises’ activities to improve social capital through questions
on organisational innovation, e.g. whether enterprises have implemented new practices and routines to
introduce a new business culture and values. Additional information can be gained from specialised
surveys. For example, questions on the duration of an external linkage can provide an indicator of the level
of trust in the relationship. An option is to include questions on the date of establishment of partnerships or
strategic alliances, for example within the last five years, more than five years ago, etc. Questions on the
degree of formalisation of a relationship, such as on the use of formal contracts, add information on the
social and cultural values involved and on the degree of trust. Further details on the role of trust can be
gained by asking how the partner was identified, such as through prior knowledge, recommendations, or at
arms length through means such as advertising.

2.3.3 Additional information on innovation co-operation

302. Additional information on innovation co-operation can be collected about the formalities
regulating co-operation, specific knowledge delivered in the field of the co-operation agreement, and more
general economic information (like sector, size and age) about each partner.

303. An additional option is to ask for the number of different partners for each category. This would
make it possible to distinguish between large and small networkers. The number and duration of
relationships are also of interest. Ideally this could contribute to distinguishing the importance of the
different relationships that make up the networks surrounding the firms.

3. OUTBOUND DIFFUSION

304. The value of innovations goes far beyond the impact on the developing firm itself. It is thus of
interest to examine the effects and benefits of innovations for other firms, consumers and the general
public. Outbound diffusion can take place with the sale of a new good or service to consumers, or the sale
of a new product or process to another firm. However, outbound diffusion is much broader than this, and
can also include the sharing of information and the diffusion of organisational and marketing innovations.
Parts of this process are already discussed connected to collaboration, which by definition involves active
participation – and knowledge or technology transfer – for all partners.

305. While it may be difficult for enterprises to assess the impacts of their innovations outside of their
own enterprise or track the use of any kinds of new knowledge, it may be possible to gain information on
the users of a firm’s innovations. This may be addressed through the identification of the main users of
their innovations for the following classifications:

• Consumer markets:

− Domestic.
− Foreign.

• Inputs to other firms:

− Domestic (inside/outside group).

62
 − Foreign (inside/outside group).

This information may also be useful in identifying the structure of demand for the firm’s innovations. The
innovative activities, strategies, and problems of firms that sell to intermediate users such as other firms
could differ from that of firms that sell to final consumers.

4. KNOWLEDGE MANAGEMENT

306. Knowledge about the core processes, products and markets can be considered to be what
constitutes a firm. Decisions on how to use and exchange existing knowledge and obtain new knowledge
are essential to the operation of enterprises. Proper systems for managing knowledge can therefore improve
competitiveness and innovative ability.

307. Knowledge management involves activities related to the capture, use and sharing of knowledge
by the organisation. It thus involves both the management of external linkages and knowledge flows within
the enterprise. This includes methods and procedures for seeking external knowledge, and for establishing
closer relationships with other enterprises (suppliers, competitors), customers or research institutions. In
addition to practices for gaining new knowledge, knowledge management involves methods for sharing
knowledge and utilizing it. This includes establishing values systems for sharing knowledge and practices
for codifying routines and practices.

308. Some examples of knowledge management practices aimed at improving the internal flow and
use of information are:

• Databases of worker ‘best practices’.

• Regular education or training programs.

• Informal and formal work teams that promote worker communication and interaction.

• Integration of activities, which promotes the interaction of employees from different activities,
for example engineers and production workers.

309. There have been a number of surveys conducted on knowledge management practices in recent
years, most notably the survey conducted in Canada48 in conjunction with OECD and international experts.
The knowledge management survey in Canada covers a number of aspects of knowledge management
practices, such as policies and strategies, leadership, knowledge capture, training and communications, and
also reasons for use of knowledge management practices and the sources that prompted the development of
practices. In addition, questions on knowledge management have been included in innovation surveys49.
Both approaches have had some success in gaining information on knowledge management practices.

310. As discussed in Chapter 3, organisational innovations may involve the implementation of

significant changes in practices for knowledge management. Thus, some information on knowledge
management can be gained through questions on organisational innovation. However, detailed

48
See e.g. Foray, D. and F. Gault (2003) op.cit. and Earl, L. (2003), “Knowledge management in practice in
Canada, 2001”, Statistics Canada.
49
For example, the CIS3 survey in France, the J-NIS 2003 survey in Japan, and the Australian Innovation
Survey 2003.

63
examinations of knowledge management activities would require a specialised survey. This information
can then be linked with information from general innovation surveys.

Chapter 6

MEASURING INNOVATION ACTIVITIES

1. INTRODUCTION

311. Information on innovation activity is useful for a number of reasons. It can provide information
on the types of innovation activity enterprises engage in, for example: whether innovative enterprises
engage in R&D; whether they purchase knowledge and technology in the form of extramural R&D,
machinery and equipment, or other external knowledge; and whether the development and implementation
of innovations also involves the training of employees; and whether enterprises are engaged in activities to
change part of their organisation.

312. Innovation activities, including capital purchases, R&D and other current expenditures that are
related to innovations, can be characterised as investments in that they can yield potential returns in the
future. These returns will often go beyond the specific innovation that the activity is directed towards. For
example, investments in R&D and innovation related training will often be of an open ended nature,
allowing their application to other tasks.

313. Quantitative measures of expenditures on each innovation activity provide an important measure
of the level of innovation activity at enterprise, industry and national levels. These measures may also be
used, along with output measures, to calculate returns to innovation activities.

314. As stated in the Frascati Manual, R&D is only one step in the innovation process. Innovation
involves a number of activities not included in R&D, such as later phases of development for
preproduction, production and distribution, development activities with a lesser degree of novelty, support
activities such as training and market preparations for product innovations, and development and
implementation activities for new marketing methods or new organisational methods. In addition, many
firms may have innovation activities that do not involve any R&D.

315. In addition to innovation activities, there are a number of other factors that may affect the ability
to absorb new knowledge and technology and to innovate. Among them are firms’ knowledge bases,
workers’ abilities and academic backgrounds, the implementation of ICTs, and proximity to public
research institutions and regions with a high density of innovative firms. Identifying the main factors
necessary for firms to innovate and those factors that enhance the ability to innovate is of great importance
for policy.

316. Innovation is a complex process, as outlined in Chapter 2, and the scale of activity required for an
innovation in a firm may vary considerably. For example, the in-house development of a radically
different and sophisticated electronic product for the mass market will involve many more steps than the

64
introduction of an improved process resulting from technology incorporated in a pre-programmed machine
purchased for the purpose.

317. Innovation activities may be carried out within the firm or may involve the acquisition of goods,
services or knowledge from outside sources, including consulting services. Thus a firm may acquire
external knowledge and technology in disembodied or embodied form.

2. THE COMPONENTS AND COVERAGE OF INNOVATION ACTIVITIES

318. As defined in Chapter 3, innovation activities are all those scientific, technological,
organisational, financial and commercial steps, including investment in new knowledge, which actually, or
are intended to, lead to the implementation of innovations. These activities may either be innovative in
themselves, or required for the implementation of innovations. Also included are basic research activities
that (by definition) are not directly related to the development of a specific innovation.

319. In order to enable comparability with the results of innovation surveys based on the second
edition of the Oslo Manual, all innovation activities other than R&D can be separated between product and
process innovation activities on the one hand, and marketing and organisational innovations on the other
hand.

320. In order to both maintain continuity with earlier measures of product and process innovation
activities and expand coverage to include activities related to all types of innovations, the following
conventions are adopted:

• While most R&D is related to product and process innovations, some R&D may be related to
marketing or organisational innovations, and basic research is by definition not related to any
specific innovation. All R&D is included as innovation activity. Furthermore, R&D is defined as
a separate category that includes relevant activities for product, process, marketing and
organisational innovations, along with basic research.

• All innovation activities other than R&D that are specifically related to marketing and
organisational innovations and not related to a product or process innovation are included under
the categories preparations for marketing innovations and preparations for organisational
innovations, respectively. This includes acquisitions of other external knowledge or capital goods
and training that are specifically related to marketing or organisational innovations.

2.1 Research and experimental development

321. Research and experimental development (R&D) comprises creative work undertaken on a
systematic basis in order to increase the stock of knowledge, including knowledge of man, culture and
society, and the use of this stock of knowledge to devise new applications (as defined in the Frascati
Manual).

322. All R&D activities financed or performed by enterprises are included as innovation activities.
This includes total intramural and extramural R&D as defined in the Frascati Manual. It is also worth
emphasizing the importance of using the definition and guidelines for R&D given in the Frascati Manual
when collecting data on R&D in innovation surveys. To ensure this, surveys should state that the definition
of R&D is the same as that used in R&D surveys. This will aid comparison with R&D surveys and
facilitate the use of the R&D data separately.

65
323. Software development is classified as R&D as long as it involves making a scientific or
technological advance and/or resolving scientific/technological uncertainty on a systematic basis. Services
development is classified as R&D if it results in new knowledge or involves the use of new knowledge to
devise new applications.

324. Construction and testing of a prototype is classified as R&D as long as the primary objective is to
make further improvements. This is often the most important phase of the experimental development of an
innovation. A prototype is an original model (or test situation) that includes all the technical characteristics
and performances of the new product or process. The acceptance of a prototype often means that the
experimental development phase ends and the other phases of the innovation process begin (further
guidance on this may be found below and in the Frascati Manual).

325. Intramural R&D comprises all R&D performed within the enterprise as defined in the Frascati
Manual and as reported in R&D surveys. This includes both R&D intended to contribute to the
development and implementation of product, process, marketing or organisational innovations and basic
research that is not directly related to the development of a specific innovation. Note that, following the
Frascati definition, intramural R&D also includes the acquisition of capital goods that are directly related
to R&D.

326. Extramural R&D comprises the acquisition of R&D services. This also includes the acquisition
of R&D services from units of multinational enterprises (MNE) that are located abroad50.

2.2 Activities for product and process innovations

Acquisition of other external knowledge

327. In addition to R&D, enterprises may acquire technology and know-how in a number of forms and
from a variety of sources in connection with the development and implementation of innovations. This also
includes acquisitions from foreign units of MNEs.

328. Acquisitions of external knowledge and technology may be in the form of patents, non-patented
inventions, licenses, disclosures of know-how, trademarks, designs and patterns.

329. Acquisitions of external knowledge may also include computer services and other scientific and
technical services for product and process innovation activities.

Acquisition of machinery, equipment and other capital goods

330. Innovation activities also involve the acquisition of capital goods. This includes both the
acquisition of capital goods with improved technological performance and the acquisition of capital goods
with no improvement in technological performance that are required for the implementation of new or
improved products or processes. The category here only comprises the acquisition of capital goods for
innovation that are not included in R&D activities. Note that this category also includes acquisitions of
capital goods from foreign units of MNEs (that are not included in R&D).

50
Foreign units of MNEs are considered as separate statistical units for innovation surveys. See Chapter 4.

66
331. Capital goods for innovations are composed of acquisitions of land and buildings, of machinery,
instruments and equipment and, in line with the revised System of National Accounts - SNA, of computer
software, which is a component of intangible investment and considered as capital formation51.

332. Land and buildings includes the acquisition of land and buildings for product and process
innovation activities including major improvements, modifications and repairs.

333. Machinery, instruments and equipment includes major instruments and equipment acquired for
use in product and process innovation activities of the firm.

334. Computer software, in line with the revised SNA, includes computer software, program
descriptions and supporting materials for both systems and applications software for use in product and
process innovation activities of the firm. The acquisition, development or extension of computer databases
expected to be used for more than one year in product and process innovation activities of the firm are also
included.

Other preparations for product and process innovations

335. Enterprises’ development of innovations may include a number of in-house activities that are not
included in R&D as defined by the Frascati Manual. This includes both the later phases of development
activities and, importantly, the introduction of product and process innovations that are new to the firm, but
not new to the market (or, in terms of the definition of R&D, do not increase the stock of knowledge or do
not contain an appreciable element of novelty). Development and implementation activities for the
adoption of new goods, services and processes may comprise an important share of innovation activity.

336. Other preparations for product and process innovations include internal activities involved in the
development and implementation of product and process innovations that are not included in R&D. This
includes development activities that are either partially excluded as R&D (such as industrial design,
engineering and set-up, and trial production) or are fully excluded (such as patent and license work,
production start-up and testing), and development activities for product or process innovations that do not
meet the novelty requirement for R&D (i.e. they are new to the firm but not new to the market). Some of
the elements of this category are described in greater detail below.

337. Other development activities for services include internal activities involved in the planning and
development of new or significantly improved services that are not included in R&D (see also Frascati
Manual §§ 145-151).

338. Design can include a wide range of activities aimed at planning and designing procedures,
technical specifications and other user and functional characteristics for new products and processes.
Among these activities are initial preparations for the planning of new products or processes, and work on
the implementation and design of these, including adjustments and further changes. This also includes
industrial design as defined in the Frascati Manual, which involves the planning of technical
specifications for new products and processes. Some elements of industrial design should be included as
R&D (see Frascati Manual §§ 124-125) if they are required for R&D.

339. Testing and evaluation comprise the testing of new or significantly improved products or
processes, while testing of prototypes is part of R&D, and so excluded here. For manufacturing, it includes

51
It is foreseen that the current revision of the 1993 SNA will change the treatment of R&D expenditure from
consumption to capital formation.

67
trial production and pilot plants as far as they are not already included in R&D. Trial production is
included in R&D if production implies full-scale testing and subsequent further design and engineering,
while pilot plants are included in R&D as long as the primary purpose is R&D. Also included here are
trials and testing activities for the provision of services, such as tests of how the provision of services
functions with the use of new technologies or trials to examine the performance of significant
improvements in existing services.

340. Setup and engineering comprise changes in production and quality control procedures, methods,
standards and associated software required to produce the new or improved product or to use the new or
improved process. Also included are changes in procedures and software required for the provision of new
services or the use of new delivery methods.

Market preparations for product innovations

341. Market preparations for product innovations can include preliminary market research, market
tests and launch advertising for new or significantly improved goods and services.

Training

342. Training is a product or process innovation activity when it is required for the implementation of
a product or process innovation, for example in order for production workers to be able to identify the
desired consistency of a new type of yoghurt in a food factory, for a marketing manager to understand the
characteristics of the improved braking system on a new model of car in order to prepare the market
launch, or for staff to be able to use different Linux programs after the introduction of a Linux-based PC
network in the firm. This category should exclude training that is already included in R&D.

343. Training is not a product or process innovation activity when it is not oriented towards a specific
product or process innovation in the firm. For example, the following are not innovation activities:
training in existing production methods for new employees, general upgrading training for individuals
(supervisors, managers, etc.), ongoing computer training, and language classes. Training for the first-time
introduction of new marketing methods or new organisational methods is part of activities for marketing
and organisational innovations (see section 2.2.).

2.3 Activities for marketing and organisational innovations

Preparations for marketing innovations

344. Preparations for marketing innovations comprise activities related to the development and
implementation of new marketing methods that have not been used by the firm before. Preparations
include development and planning of new marketing methods and work involved in the implementation of
these new methods. Activities related to marketing innovations comprise only those for developing and
implementing new marketing methods, but not expenditures for running these methods in daily business
(such as expenditures on running advertisements, event marketing or sponsoring in the course of a newly
introduced marketing method). Note that this category also includes acquisitions of other external
knowledge, acquisitions of machinery, equipment and other capital goods, and training activities that are
specifically related to marketing innovations.

68
345. Preparations for marketing innovations may be related to the development and implementation of
four types of marketing instruments that are typically distinguished in the business world: preparations for
the introduction of new marketing methods in product design or packaging, in pricing methods, in product
placement, and in product promotion.

Preparations for organisational innovations

346. Preparations for organisational innovations include the development and planning of new
organisational methods and the work involved in the implementation of these new methods. Note that this
category also includes acquisitions of other external knowledge, acquisition of machinery, equipment and
other capital goods, and training activities that are specifically related to organisational innovations.

347. Preparations for organisational innovations may be distinguished by new methods in business
practices, new methods in workplace organisation and new methods in organising external relations.

2.4 Design

348. The term product design, as used in the definition of marketing innovations, refers to products’
form and appearance, and not to products’ technical specifications or other user or functional
characteristics. However, design activities may be understood by enterprises in more general terms, to also
be an integral part of the development and implementation of product or process innovations, as described
in paragraph 338. The categorisation of design activities will thus depend on which type of innovation they
are related to.

349. All design activities for the development and implementation of product innovations (including
work on products’ form and appearance) and process innovations should be included either in R&D or
Other preparations for product and process innovations.

350. Work related to changes in product design that are marketing innovations (and not product
innovations, i.e. where the functional characteristics or intended uses of the product in question are not
significantly improved) should be included in Preparations for marketing innovations.

2.5 The borderline between R&D and non-R&D innovation activities

351. Enterprises may have difficulty in differentiating between R&D expenditure and other innovation
activities, especially at the borderline between R&D and non-R&D activity. Care must be taken to exclude
from R&D activities that are part of the innovation process but rarely involve any R&D (e.g. patent work,
licensing, market research, manufacturing start-up, process re-engineering, tooling up). At the same time,
some activities are at least partly counted as R&D (e.g. pilot plants, prototypes, industrial design, process
development).

352. Distinguishing R&D and other innovation activities is particularly difficult for services (see
Frascati Manual §§ 145-151), due in part to the fact that innovation activities in services tend to be less
formally organised, and that R&D in services is less well defined than for manufacturing.

353. The basic criterion for distinguishing R&D activities from non-R&D innovation activities “is the
presence in R&D of an appreciable element of novelty and the resolution of scientific and/or technological
uncertainty” or that they “result in new knowledge or use of knowledge to devise new applications” (see
the Frascati Manual, §84 and §146). These criteria imply “that a particular project may be R&D if

69
undertaken for one reason, but not if carried out for another” (Frascati Manual, §85). It is recommended
that the guidelines in the Frascati Manual, §§ 110-130 and 145-151, should be applied to innovation
surveys.

2.6 The development and use of software in innovation activities

354. The development, acquisition, adaptation and use of software pervade innovation activities.
Developing new or substantially improved software, either as a commercial product or for use as an
in-house process (an innovation in its own right), involves research and experimental development and a
range of post-R&D activities. In addition, all types of innovations may involve the acquisition and
adaptation of software, where the software is not an innovation in itself but is required for the development
and implementation of innovations.

70
3. COLLECTING DATA ON INNOVATION ACTIVITIES

355. Summarising the section above, innovation activities can be broken down in the following
way:

Research and experimental development

Intramural (in-house) R&D: Creative work undertaken on a systematic basis within the enterprise in
order to increase the stock of knowledge and the use of this stock of knowledge use to devise new
applications. This comprises all R&D conducted by the enterprise, including basic research.

Acquisition of R&D (extramural R&D): Same activities as intramural R&D, but purchased from public or
private research organisations or from other enterprises (including other enterprises within the group).

Activities for product and process innovations:

Acquisition of other external knowledge: Acquisition of rights to use patents and non-patented inventions,
trademarks, know-how and other types of knowledge from other enterprises and institutions such as
universities and government research institutions, other than R&D.

Acquisition of machinery, equipment and other capital goods: Acquisitions of advanced machinery,
equipment, computer hardware or software, and land and buildings (including major improvements,
modifications and repairs), that are required to implement product or process innovations. Excluded are
acquisitions of capital goods that are included in intramural R&D activities.

Other preparations for product and process innovations: Other activities related to the development and
implementation of product and process innovations, such as design, planning and testing for new products
(goods and services), production processes, and delivery methods that are not already included in R&D.

Market preparations for product innovations: Activities aimed at the market introduction of new or
significantly improved goods or services.

Training: Training (including external training) linked to the development of product or process
innovations and their implementation.

Activities for marketing and organisational innovations:

Preparations for marketing innovations: Activities related to the development and implementation of new
marketing methods. Includes acquisitions of other external knowledge and other capital goods that are
specifically related to marketing innovations.

Preparations for organisational innovations: Activities undertaken for the planning and implementation
of new organisation methods. Includes acquisitions of other external knowledge and other capital goods
that are specifically related to organisational innovations.

356. Total expenditure for innovation activities comprises current and capital expenditure incurred for
the innovation activities defined above. Current innovation expenditures are composed of labour costs and
other current costs. Capital expenditures for innovations are composed of gross expenditures on land and
buildings, on instruments and equipment and on computer software. Capital expenditures that are part of

71
R&D are included in intramural R&D, while non-R&D capital expenditures linked to product and process
innovations are included in acquisition of machinery, equipment and other capital goods. Non-R&D
capital expenditures specifically linked to marketing or organisational innovations are included in
preparations for marketing innovations and preparations for organisational innovations, respectively. The
remaining categories of innovation activity consist solely of current expenditure.

357. Innovation surveys can collect both qualitative and quantitative data on innovation activities.
Qualitative data involves questions on whether or not enterprises have engaged in an innovation activity.
Quantitative data involve questions on expenditures for an innovation activity.

358. Innovation expenditure data are among the most important and are in demand both for research
and policy. However, innovation expenditure questions are also among the most difficult and time
consuming to answer. Innovation surveys may consider limiting the number of categories of innovation
activities for quantitative questions.

359. Difficulty in reporting innovation expenditure raises two issues that innovation surveys can
consider in designing quantitative questions of innovation activities. The first issue is response burdens.
While activities such as R&D may be confined to one department, innovation activity can take place
throughout the enterprise. In addition, expenditures on a number of activities may not be directly available
from enterprises’ accounting systems. Expenditure questions are thus likely to be the most time consuming
to answer and detailed questions on expenditure may therefore have an impact on unit and item response
rates. This is of particular concern for non-mandatory surveys.

360. A second, related issue is the quality of the data. Data quality is likely to be best for those
expenditures that can be extracted from accounting information, while other expenditures may be rough
estimates, if given at all. Questionnaire design, the number of expenditures categories and wording may be
very important for the quality of the data collected on innovation expenditure.

3.1 Qualitative data on innovation activity

361. It is recommended that qualitative data be collected on innovation activities. Questions on

whether or not firms have engaged in the above activities can either refer to a single year or the
entire observation period. Surveys may wish to collect qualitative data on all categories of innovation
activities or on a subset of these. The multi-year approach has the advantage of capturing innovation
activity for enterprises that may not perform innovation activity on a regular basis.

362. Additional information can also be collected on individual types of innovation activities.
Examples are to ask whether R&D activity is performed on a continuous or an occasional basis, to ask
which types of external knowledge the enterprise has acquired, or to include a separate question on
software expenditures.

3.1.1 Other qualitative indicators of innovation activity

363. Information on employee characteristics, such as the level of education and number of technical
staff, can also be collected in innovation surveys. For example, the share of employees with a higher
education certificate or degree (ISCED 5-6) and the share of personnel involved in innovation or R&D
activity can be used as supplementary measures of the capability for innovation that lies in the firm’s own
knowledge stock and people. Furthermore, most enterprises are likely to already possess information on
their employees’ education. An additional qualitative indicator is whether enterprises participate in national

72
or supranational programmes that provide financial support for employee education/training or for the
employment of research personnel.

3.2 Quantitative data on innovation activity

364. For the collection of quantitative data on innovation expenditures, it is recommended that a
breakdown by type of activity is used (see § 353). Surveys may wish to collect quantitative data on all
categories of innovation activities or, as innovation expenditures are difficult to measure, surveys may opt
to collect data on a subset of these.

365. Innovation expenditure can also be broken down by type of expenditure (current innovation
expenditure vs. capital expenditures for innovations) and by source of funds. These are discussed in
sections 3.4 and 3.5.

366. While collecting data on innovation activities for a multi-year period may be feasible for
qualitative questions on innovation activities, restricted availability of data within firms is a serious
obstacle to the multi-year approach for quantitative data. Therefore, it is recommended that quantitative
questions on innovation expenditure refer only to the last year of the observation period, the
reference year.

367. Concerning capital purchases, expenditures for the category, acquisition of machinery,
equipment and other capital goods, should exclude purchases of capital goods that are already
included in intramural R&D. Purchases of capital goods should be included in full for the period when
they took place. All depreciation provisions for building, plant and equipment, whether real or imputed,
should be excluded from the measurement of intramural expenditure.

368. Enterprises often face severe problems in supplying reliable estimates of capital expenditure for
innovation activities. To assist them, data on total capital expenditure (including capital expenditure not
related to innovation activities) can be collected as well. This will also help in checking the reliability of
innovation expenditure data.

369. Some innovations can span more than one innovation type. An example is the implementation of
a marketing innovation that is connected to a product innovation. In order to avoid double counting,
innovation surveys should ensure that innovation expenditures are included in no more than one
category of innovation activities.

3.3 Other measurement issues

3.3.1 Intramural and extramural expenditure

370. While a large part of innovation activities are separated into intramural and extramural
expenditures, a full separation is likely not feasible for most enterprises, and thus is not recommended in
this Manual.

371. The inclusion of extramural expenditures is important at the micro level for measuring
expenditures for individual enterprises. However, special care has to be taken when aggregating
individual enterprise amounts to industry or national figures, because of double-counting.
Double-counting is most likely to be prevalent for extramural R&D and acquisitions of other external
knowledge.

73
3.4 Breakdown by type of expenditure

372. Expenditure for innovation activities can also be broken down into current and capital
expenditure. This can be useful if the data are to be compared with those on intangible investment, with
which innovation expenditure is sometimes confused (see §373). Among the categories of innovation
activities, capital expenditures for innovation are included in intramural R&D, acquisitions of machinery,
equipment and other capital goods, and potentially also in preparations for marketing innovations and
preparations for organisational innovations. All other categories consist solely of current expenditures.
Thus a breakdown of innovation expenditure by type of expenditure can potentially be obtained by asking
enterprises to specify R&D, preparations for marketing innovations and preparations for organisational
innovations according to current and capital expenditures. However, given that this would increase
response burdens and that R&D expenditure is collected in considerable detail through other surveys in
most countries, this breakdown is not recommended for general innovation surveys.

373. Current innovation expenditures are composed of labour cost and other current costs:

• Labour costs comprise annual wages and salaries and all associated costs of fringe benefits such
as bonus payments, holiday pay, contributions to pension funds and other social security
payments, payroll taxes and so on. The labour costs of persons not involved in innovation
activities (such as security personnel and maintenance staff) should be excluded here and
included under other current costs.

• Other current costs comprise non-capital purchases of materials, supplies, services and
equipment to support innovation activities performed by the firm in a given year.

374. Capital expenditures for innovations are defined above in section 1.1.3.

3.4.1 The relation between intangible investment and innovation expenditure

375. Intangible investment covers all non-capital expenditure for the firm’s development which is
expected to give a return over a longer period than the year in which it is incurred. It is generally taken to
cover expenditure on non-routine marketing, training, software and some other similar items, in addition to
current expenditure on R&D.

376. Current expenditure on innovation is clearly a part of intangible investment, but intangible
investment comprises elements which are not part of current innovation expenditure. For example, only
training in connection with the introduction of innovations is classified as innovation expenditure, whereas
intangible investment includes all of the firm’s training expenditure. Marketing in connection with the
introduction of new products or the development and implementation of new marketing methods is
classified as innovation expenditure. Intangible investment, on the other hand, includes all marketing
expenditure in general.

377. At the same time, innovation expenditure includes tangible investment such as capital
expenditure on R&D, and the acquisition of new machinery and equipment related to innovations.

3.5 Breakdown by source of funds

378. It is important to know how innovation expenditure is financed, for instance in order to evaluate
the role of public policy and internationalisation in the innovation process. The following classification by
source of funds can be used:

74
List of sources of funds:

• Own funds.
• Funds from related companies (subsidiary or associated companies).
• Funds from other (non-financial) enterprises.
• Funds from financial companies (bank loans, venture capital, etc.).
• Funds from government (loans, grants, etc.).
• Funds from supranational and international organisations (EU, etc.).
• Other sources.

In addition, external sources of funds can be broken down into domestic and international sources of funds.

379. Where survey designers wish to cover these topics, it may be enough, for a variety of policy and
research issues, to collect information on whether or not each source is used, instead of seeking an
estimate, probably imprecise, of the amount (either in monetary or percentage terms) contributed by each
source. This will considerably reduce the response burden on firms, and hence increase the total response
rate to the survey as well as cutting item non-response to this question. Due to concerns on response
burdens for general innovations surveys, this Manual makes no recommendation on the extent of coverage
of sources of funds.

380. To evaluate the role of government procurement in innovation processes, it is useful to know
whether or not a firm participates in government procurement at regional, national or international levels
related to innovative products and processes.

3.6 The subject approach versus the object approach

381. Survey questions on innovation expenditure may be formed in two ways:

• The total expenditure on innovation activities for the firm in a given year or period
(= the subject approach).

• The total expenditure for specific innovations implemented in a given year or during a given
period regardless of the year in which the expenditure occurs (= the object approach).

382. There is a fundamental difference between the two approaches, hence the results obtained are
different. Since both have been used in a number of innovation surveys, it seems worthwhile to clarify the
relation between them.

383. The subject approach covers expenditure for implemented, potential and aborted innovation
activities as defined above. In this respect, it is a straightforward extension of traditional R&D
measurement.

384. In the object approach the sum reported comprises total expenditure on defined innovations, or
on the main innovation(s), that have been implemented during a given period. It excludes expenditure on
innovation projects that have been aborted or are still in progress, and on general R&D not connected to a
specific application. This approach seems particularly suitable for innovation surveys starting from a set of

75
identified innovations, but it could also be used in surveys of the innovation activities of enterprises in
general.

385. In light of the advantages and disadvantages of both approaches, the subject approach is
recommended for reporting on innovation expenditure. The guidelines in this Manual are therefore
oriented to the subject approach.

76
 Chapter 7

OBJECTIVES, OBSTACLES AND OUTCOMES OF INNOVATION

1. INTRODUCTION

386. This chapter discusses incentives and obstacles for engaging in innovation activity and the
impacts of innovations. Identifying the factors that drive innovation and those that hinder it is of great
value both for understanding the innovation process and in forming innovation policy. Interest in
measuring innovation lies in its relation to the performance of enterprises, industries, and the economy as a
whole. Measures of the impact of innovation on enterprise performance are thus among the most important
innovation indicators, though they are also among the most difficult to obtain.

387. Impacts of innovations on enterprise performance range from effects on turnover and market
share to changes in productivity, efficiency. Important impacts at industry and national levels are changes
in international competitiveness and in total factor productivity, knowledge spillovers of firm-level
innovations, and an increase in the amount of knowledge flowing through networks.

388. Objectives and barriers to innovation can vary by type of innovation. For example, objectives of
product or marketing innovations will primarily be demand oriented (e.g. improving product quality,
increasing market share, entering new markets), while process or organisational innovations will tend to be
supply oriented (e.g. reducing costs, improving production capabilities). Some barriers can relate to all
types of innovation (e.g. cost factors) while others can relate to a subset of innovation types.

389. This chapter both describes a number of indicators for innovation surveys and discusses other
measures that, while very relevant, may be difficult to include in general innovation surveys though may
be included in specialised surveys. Surveys can choose to refer questions on objectives, barriers and other
indicators to all types of innovations, to subsets such as product and process innovations, or to individual
innovation types. This chapter is designed to provide guidelines for all these approaches.

2. OBJECTIVES AND EFFECTS OF INNOVATIONS

390. Enterprises may engage in innovation activity for a number of reasons. These objectives can
relate to products, markets, efficiency, quality or the ability to learn and to implement changes. Identifying
enterprises’ motives for innovating and their importance aids in examining the forces that drive innovation
activity, such as competition, and opportunities for entering new markets. Data on objectives can also
provide additional information on the characteristics of types of innovations.

391. Enterprises may or may not be successful in achieving their objectives in implementing
innovations, or innovations may have other or additional effects than those that initially motivated their
implementation. While objectives concern enterprises’ motives for innovating, effects concern the actual
observed outcomes of innovations. While their interpretation is different, objectives and effects can refer to
the same set of factors. The list below contains objectives or effects for all four types of innovations. A

77
number of these factors can be relevant for more than one type of innovation. In particular, product and
marketing innovations or process and organisational innovations may have a number of factors in common.

78
 Objectives and effects of innovation

Relevant for: Product Process Organi- Marketing

innovations innovations sational innovations
innovations
Competition, demand and markets
Replace products being phased out *
Increase range of goods and services *
Develop environment-friendly products *
Increase or maintain market share * *
Enter new markets * *
Increase visibility or exposure for products *
Reduced time to respond to customer needs * *
Production and delivery
Improve quality of goods and services * * *
Improve flexibility of production or service * *
provision
Increase capacity of production or service * *
provision
Reduce unit labour costs * *
Reduce consumption of materials and energy * * *
Reduce product design costs * *
Reduce production lead times * *
Achieve industry technical standards * * *
Reduce operating costs for service provision * *
Increase efficiency or speed of supplying * *
and/or delivering goods or services
Improve IT capabilities * *
Workplace organisation
Improve communication and interaction *
among different business activities
Increase sharing or transferring of knowledge *
with other organisations
Increase the ability to adapt to different client * *
demands
Develop stronger relationships with customers * *
Improve working conditions * *
Other
Reduce environmental impacts or improve * * *
health and safety
Meet regulatory requirements * * *

392. The factors are designed to shed light on a number of forces driving enterprises’ innovation
activity. Competition, demand and markets concern the main incentives for product innovations and in
some cases also for marketing innovations. The intent of questions on these factors is to determine the
importance of motives for product innovations, such as: short product life spans that necessitate the
development of new products; the need to diversify product portfolios; or efforts to increase or avoid a
decline in market share.

393. In addition, there are a number of factors with the aim of identifying the main motives for
changes in production and delivery, i.e. whether their main intent is to improve quality, flexibility, or

79
efficiency/cost reduction. In particular, factors on cost reduction are made specific to enable better
interpretation of results. Factors concerning workplace organisation identify the main forces behind
organisational change, whether changes are oriented towards customer relations, operational efficiency, or
improving the capture and sharing of knowledge.

394. It is recommended that data are collected on the objectives or effects of innovations
implemented by enterprises during the period under review. The use of an ordinal scale is
recommended in which enterprises are asked whether the factor is relevant and, if so, its
importance. Both questions on the objectives of innovation and their actual effects can provide valuable
information on enterprises’ innovation activity. However, it may not be possible to include both questions
in innovation surveys. The choice between the two questions may depend on which is considered to be
most useful for policy. However, a drawback with questions on effects is that the impacts of recent
innovations may not yet have taken effect within the time period of the survey. The disadvantage of
questions on objectives is that the actual effects could differ substantially from expectations.

395. Questions on objectives or effects can either refer to all types of innovations or to a subset of
these, such as product and process innovations. Limiting the types of innovations this question refers to can
improve interpretation of the data (given that most factors are relevant for at least two types, for example
product and marketing innovations or process and organisational innovations). An additional option is to
include questions for each innovation type (or subsets of types) separately.

3. OTHER MEASURES OF IMPACTS ON ENTERPRISE PERFORMANCE

396. This section considers additional output indicators to those discussed in section 2, in particular
quantitative indicators of innovation performance.

397. The success of an innovation may depend on a number of factors. It will clearly depend on the
quality of the innovation. The impact of innovations may also vary greatly from sector to sector or region
to region. In addition, it may depend on other changes in the enterprise that may support the innovations.
For example, the success of product innovations may depend to a large degree on marketing initiatives to
launch the product. Likewise, the impact of process innovations may depend on organisational changes to
take advantage of these new processes. A well-documented example is the importance of organisational
changes for the impact of investments in ICTs on productivity52.

398. It is generally difficult to ask for quantitative measures of the effects of innovations in surveys,
even for very rough estimates, as these calculations will often require substantial analysis on the part of the
enterprise. This section discusses some output indicators that can potentially be used in innovation surveys.

399. An important aspect for all analyses of the impact of innovations is the time lag between
innovations and their impacts. Some effects may materialise over the course of the observation period,
while others may take longer. Having innovation data on a large cross section over time (i.e. panel data) is
of great value for these types of analyses. Panel surveys can thus open up interesting possibilities for
analysing the results of innovation.

400. In order to aid in forming a picture of how innovation affects general performance, some general
data on the enterprise can be collected, for the beginning and end of the observation period. This can
include data on turnover, exports, employees, and operating margins. This data can then be used in

52
See for example, Brynjolfsson, E. and L.M. Hitt (2000) op.cit. and OECD (2004), op.cit.

80
subsequent analysis to examine what effect various indicators have on the above variables. Data can be
collected via the innovation survey, or taken from other available sources.

3.1 Impact on turnover

3.1.1 Proportion of turnover due to new or significantly improved products

401. An indicator on the proportion of turnover due to new or significantly improved products
provides important information on the impact of product innovations on the overall makeup of turnover
(i.e. the share of turnover from new products) and on the degree of innovativeness of the enterprise.

402. Questions on the proportion of turnover due to product innovations should refer to the impact of
product innovation over the observation period on turnover in the reference year (i.e. the last year of the
observation period). It is recommended that enterprises are asked to estimate the percentage share of
total turnover in the reference year due to:

• New or significantly improved goods and services introduced during the observation period that
were new to the market (as defined in Chapter 3).

• New and significantly improved goods and services introduced during the observation period that
were new to the firm, though not new to the market.

• Products that were unchanged or only marginally modified during the observation period.

403. A number of firms can have come into existence during the period under review. This includes
both newly established firms and firms that are the result of mergers, demergers and other kinds of
reorganisation. These firms should be treated in the same way as for all other firms when constructing this
indicator.

404. Respondents should supply their best estimates of the actual percentages. When presenting the
results by industry, firm size, region or country, the percentages should be calculated as the ratio of total
turnover due to new products to total turnover for the industry, etc.

Product life cycles

405. Outcome indicators are directly influenced by the length of product life cycles. They are likely to
be higher in product groups where life cycles are short and innovation can be expected to occur more
frequently.

406. In order to take the effects of product life cycles into account, an option is to ask enterprises to
give an estimate of the average length of its products’ life cycles. This information could be used to
weight the outcome indicators for turnover shares. An alternative way of formulating this question is to
ask how often the enterprise usually introduces innovations.

3.1.2 Process innovations

407. Innovation surveys can also ask enterprises to estimate the percentage of turnover that is
impacted by process innovations. This can provide an indication of how extensive process innovations are
in terms of the enterprise’s total operations.

81
3.1.3 Marketing innovations

408. The development and implementation of marketing innovations constitutes an important

innovation activity for many enterprises in terms of impact on performance. In order to gain an idea of the
scope of marketing innovations, an option is to ask enterprises to estimate the percentage of total turnover
that is affected by marketing innovations. Innovation surveys can ask two separate questions concerning
marketing innovations. The first is to provide an estimate of the percentage of turnover due to goods and
services with significant improvements in product design or packaging. The second is to provide an
estimate of the share of turnover impacted by new marketing methods in pricing, promotion or placement.
Note that questions on turnover due to changes in product design should not be combined with questions
on turnover due to product innovations (i.e. these two questions should be asked separately), since some
new or improved products might be both product and marketing innovations. Nor should questions on new
marketing methods in product design be combined with questions on the share of turnover due to other
new marketing methods. As with other questions concerning the impact on turnover, enterprises will likely
only be able to provide rough estimates at best.

3.2 The impact of process innovations on costs and employment

409. As discussed above in section 2, innovation surveys can include questions on the effects of
innovations and their relative importance. An option for obtaining additional information on process
innovations is through questions on their impacts on costs or employment.

410. There are a number of options for asking questions on the impact of process innovations on costs.
Enterprises can be asked first, whether process innovations implemented during the observation period led
to an increase, decrease or no change in costs. A ‘yes’ response to a decrease or increase can be followed
by further questions to quantify the amount of change.

411. These questions can either be asked with respect to average costs or to specific costs, for example
changes in the cost of material, energy, or labour inputs. Quantitative questions can either ask for an
interval estimate of the percent change in costs, or ask enterprises to choose from a set of pre-defined
categories (e.g. an increase or decrease of less than 5%, 5% to 25%, over 25%). Experience from earlier
surveys has indicated that enterprises find the latter method to be easier to answer, resulting in much higher
item response rates. The same techniques can also be used to ask about the effect of process innovations on
employment, i.e. whether employment increased or decreased, and by how much.

412. This approach can also be used for organisational innovations. In this case, questions should be
asked with respect to average costs as opposed to specific costs.

3.3 The impact of innovation on productivity

413. Concerning impacts on productivity, there are a number questions that are of great interest, for
example whether process innovations or organisational innovations improve efficiency. A detailed
examination of these issues would require separate analyses using both innovation data and other economic
data on firm performance. In many cases, panel data on innovation would be necessary, though some
analysis can be performed using innovation data from a single survey combined with economic data for
more than one year. An example is empirical analysis of ICT investments and organisational innovation,

82
which generally have found that the effects of ICT investments on productivity depend greatly on
organisational innovation53.

53
See for example, Brynholfsson, E. and L.M. Hitt (2000).

83
4. FACTORS HAMPERING INNOVATION ACTIVITIES

414. Innovation activity may be hampered by a number of factors. There may be reasons for not
starting innovation activities at all, or factors that slow innovation activity or have a negative effect on
expected results. These include economic factors, such as high costs or lack of demand, enterprise factors
such as a lack of skilled personnel or knowledge, and legal factors such as regulations or tax rules.

415. Questions on barriers to innovation can provide information on a number of issues relevant for
innovation policy. Small and medium sized enterprises may identify a lack of available finance as an
important barrier to investments in innovation. Enterprises may be concerned that there will be a lack of
demand for new products at the prices that enterprises would need to charge to make the innovation
worthwhile. Enterprises may not have the skilled personnel needed to engage in innovation activities, or
their innovation activities may be slowed because they are unable to find the necessary personnel on the
labour market. A lack of infrastructure may be an important barrier to innovation, in particular outside of
large cities. Other reasons may be that the enterprise lacks knowledge on technologies or markets needed
to develop an innovation, or that the enterprise is unable to find suitable partners for joint innovation
projects. A list of factors that could act as barriers to innovation is given below.

416. These barriers can be related to a specific type of innovation or to all types. For example, cost
factors can be relevant for all types of innovations, and market factors can affect both the development of
product innovations and work on product design (i.e. marketing innovations). Also shown in the list is
which types of innovations are relevant for each barrier.

417. It is recommended that data are collected on barriers to innovation activity and their
relative importance for the period under review. Questions on barriers to innovation activity should
be asked of both innovative and non-innovative enterprises. Questions on barriers can refer to all types
of innovations or to a subset of types, such as product and process innovations.

84
 Factors hampering innovation activities
Relevant for: Product Process Organisational Marketing
innovations innovations innovations innovations

Cost factors:
Excessive perceived risks * * * *
Cost too high * * * *
Lack of funds within the enterprise * * * *
Lack of finance from sources outside the enterprise:
Venture capital * * * *
Public sources of funding * * * *
Knowledge factors:
Innovation potential (R&D, design, etc.) insufficient * * *
Lack of qualified personnel:
Within the enterprise * * *
In the labour market * * *
Lack of information on technology * *
Lack of information on markets * *
Deficiencies in the availability of external services * * * *
Difficulty in finding co-operation partners for:
Product or process development * *
Marketing partnerships *
Organisational rigidities within the enterprise:
Attitude of personnel towards change * * * *
Attitude of managers towards change * * * *
Managerial structure of enterprise * * * *
Inability to devote staff to innovation activity due to * *
production requirements
Market factors:
Uncertain demand for innovative goods or services * *
Potential market dominated by established * *
enterprises
Institutional factors:
Lack of infrastructure * * *
Weakness of property rights * *
Legislation, regulations, standards, taxation * * *
Other reasons for not innovating:
No need to innovate due to earlier innovations * * * *
No need because of lack of demand for innovations * *

85
5. QUESTIONS ON THE APPROPRIABILITY OF INNOVATIONS

418. The ability of enterprises to appropriate the gains from their innovation activities is an important
factor affecting innovation. If, for example, enterprises are unable to protect their innovations from
copying by their competitors, they will have less incentive to innovate. On the other hand, if an industry
functions well without formal protection methods, promoting these may slow the flow of knowledge and
technology and lead to higher prices for goods and services.

419. Policy plays a central role in the design of legal methods of protecting innovations. Data on
which types of methods are used and their relative importance can help inform policies to maximise the
economic and social benefits from intellectual property rights.

420. The following list of methods of protection is suggested:

Formal methods:
• Patents.
• Registration of design.
• Trademarks.
• Copyrights.
• Confidentiality agreements and trade secrecy.

Informal methods:
• Secrecy that is not covered by legal agreements.
• Complexity of product design.
• Lead time advantage over competitors.

An additional formal method that is used in some countries is petty patents or utility patents, which are
rights for protection of inventions that are granted without any formal review.

421. Patent data, both applications and grants, function as an intermediate output indicator for
innovation activity and also provide information on the innovative capabilities of the enterprise. For
example, an enterprise that has applied for patents can be assumed to be capable of developing innovations
that are new to the world (occasionally only new to the market, depending on the patent strategies of other
firms). Data on whether or not enterprises have applied for or been granted a patent can thus provide useful
data for innovation surveys and be used in specialized surveys on intellectual property rights (IP). Note that
patent data should refer to the country where the patent was developed and not the country where the
patent application was made. Patent data is discussed in greater detail in the Patent Manual.

422. Registration of design is primarily a method of protecting the aesthetic design of products, to
prevent other enterprises from using the same design. Enterprises may also register trademarks related to
the enterprise as a whole or a product line. This method protects the enterprise’s image, and the association
of products with the enterprise. Copyrights relate to the final use of some types of products and establish
rights to claim payment for the use of copyright-protected products.

86
423. Patents are methods for protecting research and development results. Confidentiality agreements
between enterprises and other organisations are also designed to protect R&D work, while at the same time
allowing the enterprise to interact with other organisations on this work.

424. It is recommended that data are collected on whether or not enterprises have used various
methods of protection for their innovations during the observation period. Questions can either use a
binary or an ordinal scale. Protection methods can be relevant for all types of innovations, though asking
questions for all innovation types combined can lessen the ability to interpret the data (i.e. to link
protection methods to specific types of innovations).

425. Among the options for questions on methods of protection are:

• Refer to product and process innovation only. A supplementary question could be asked on
marketing and organisational innovations (e.g. whether any formal methods of protection were
used for these innovations). An ordinal scale can be used here to ask on the relative importance of
different methods.

• Refer to all types of innovations combined. An ordinal scale could also be used here.

• Refer to each individual innovation type, allowing enterprises to choose those innovations that
are relevant for each protection method. This would allow the greatest level of detail on the use
of protection methods, such as which formal methods are used for marketing methods, which
innovations patents are used for, and whether secrecy or other methods are used for
organisational or other innovations.

87
 Chapter 8

SURVEY PROCEDURES

1. INTRODUCTION

426. The correct application of statistical methodology is crucial for the collection and analysis of
innovation data. Based on theoretical knowledge and on practical experience in recent innovation surveys
at national and international levels, this chapter discusses and provides guidelines on key elements for the
collection and analysis of innovation data.

427. Following these guidelines will generally lead to comparable results over time and across
countries. Particular circumstances may require a country to use another methodology. In such cases, the
methods used should be considered carefully so as to ensure that the results are still comparable.
Divergence from the guidelines should be documented in sufficient detail to explain comparability issues
with data from other countries.

2. POPULATIONS

2.1 The target population

428. Innovation activities take place in all parts of an economy: in manufacturing, the service
industries, public administrations, the health sector and even private households. In reality, for various
theoretical and practical reasons, a survey will not cover all possible units. The concept of innovation may
be less clear in some parts of the economy, especially with respect to non-market oriented activities.

429. As discussed in Chapter 1, it is therefore recommended that innovation surveys should refer to
innovation activities in the business enterprise sector. Hence, the target population for innovation surveys
concerns statistical units (innovators and non-innovators, R&D performers and non-R&D performers) in
the business enterprise sector. This includes both goods producing and service industries. A proposed list
of industrial classifications to be included in the target population for innovation surveys is shown in
Chapter 4.

430. Innovative activities take place in small and medium-sized units as well as in large units. In
order to capture innovation activity in these smaller units, it is recommended that the target population
should include, at a minimum, all statistical units with at least 10 employees. This threshold may be
higher for specific industries, such as Construction and Retail Trade. Surveys may also find it useful to
include units with less than 10 employees. Innovation activities in these smaller units are of considerable
policy interest for a number of sectors, such as high technology manufacturing and knowledge intensive
services.

88
2.2 The frame population

431. The units from which a survey sample or census is drawn form the frame population. When
preparing a survey, the target and frame populations should be as close as possible. In practice the
population from which the sample or census is drawn may not be the same as the target population. For
example, the frame population underlying the survey (such as a register) may include units that no longer
exist, or units that no longer belong to the target population. At the same time, it may not contain units that
in fact belong to the target population.

432. The frame population is based on the last year of the observation period for surveys. Special
attention should be paid to a number of changes to units that may have taken place during the observation
period. Among these are: changes in industrial classifications, new units created during the period,
mergers, splits of units, and units that ceased activities during the last year of the period.

433. An ideal frame is an up-to-date official business register established for statistical purposes.
National statistical offices usually keep such registers. Other registers can be used as well, provided their
quality is comparable to that of the official register. The units in the register will in many countries be legal
units, so the units may not all be statistical units in terms of the innovation survey, as defined in Chapter 4.
Guidelines to check the validity of the responding statistical unit should be given in the questionnaire.

434. If the register forms the basis for several surveys, such as the innovation survey, the R&D survey
and the general business statistics survey, the information collected in the innovation survey can be
restricted to issues specific to innovation. Other information, for example on R&D or on general economic
variables such as employment, turnover, exports or investments, can be taken directly from the other
surveys based on the register. Accordingly, basing different types of surveys on a single business register
compiled for statistical purposes is desirable.

3. SURVEY METHODS

3.1 Mandatory or voluntary survey

435. Innovation surveys can be either mandatory or voluntary. If they are voluntary, higher
non-response rates should be expected. Low response rates mean a smaller sample than expected and thus
higher variance. This could be compensated to some extent, in the case of sample surveys, by higher
sampling fractions. But, increasing the sampling fractions does not solve the basic problem of bias in the
estimates for the target population due to a high non-response rate, making further analyses less
representative.

3.2 Census or sample survey

436. Innovation data can be collected through census or sample surveys. Resource limitations and
response burdens will in most cases rule out a survey of the entire population (census). If sample surveys
are utilised, the units should be selected on the basis of a random procedure (random sample surveys with
known selection probabilities). Sample surveys should be representative of the basic characteristics of the
target population, such as industry, size and region, so a stratified sample is necessary.

437. A census may be unavoidable in some cases. It may be a legal requirement that all business
surveys have to be censuses. In addition, when the frame population is fairly small (e.g. in small
countries), proper sampling may produce sample sizes for some strata that are relatively close in size to the

89
frame population of the strata. In such cases censuses may be worth considering. Finally, it may be
decided that all units in the frame with more than a given number of employees should be included.

438. For sample surveys, the sample of enterprises should be large enough to give reliable results for
the units in the target population and by characteristics of interest of the target population, such as specific
sectors. Estimates of the acceptable coefficients of variation can be used to estimate the required number of
responses for reliable results. The total sampling fraction will vary depending on the size of the total frame
population, while the sampling fractions of each stratum will depend on the number of units, the size of the
units and the variability of the main indicators. In general, the necessary sample fraction will decrease with
the number of units in the population and increase with the size of the units and the variability of the
sample.

3.3 Domains

439. Particular subsets of the target population may be of special interest to users or users may need
detailed information at sector or regional levels. These subsets are called domains (or sub-populations). To
get representative results for the domains, the domains must be subsets of the sampling strata. The most
frequent approach is “over-allocation” in order to produce reliable results for the domains. Additionally,
establishing domains may allow for the coordination of different business enterprise surveys, as well as for
comparisons over time between enterprises with similar characteristics. Some potential sub-populations
that can be considered are: industry groupings, size classes, regions, units that engage in R&D, and
innovation-active units.

440. Some guidelines for the use of domains are the following:

• Statistical units and classifications should be the same in all parts of the sample including
domains.

• Methods used (e.g. weighting methods) to calculate results for subsets should be consistent with
those used for results from the main sample.

• Deviations in data treatment or differences in the quality of the results from the domains should
be documented.

3.4 Sampling techniques

441. Innovation surveys are in general random sample surveys. The relevant literature offers various
sampling techniques, such as the simple random sample technique, stratification techniques, cluster sample
techniques and pps-sample54 techniques. The techniques may even be combined. In the past, stratified
sample surveys have proved to lead to reliable results.

442. If stratification techniques are used, some general rules with regard to the selection of the
stratification variables should be observed. In principle, stratification of the population should lead to
strata that are as homogeneous as possible in terms of their innovation or non-innovation activities. Given
that the innovation activities of units in different industries and in different size classes can differ

54
Pps-sampling = The units are included with Probabilities Proportional to their Size, often measured as
number of employees in the business sector.

90
significantly, it is recommended that the stratification of random sample innovation surveys should
be based on the size and principal activity of the units

443. The size of units should be measured by number of employees. Recommended size classes are
provided in Chapter 4. Some recommendations for analytical purposes, which might also be used for
stratification, are given below.

444. The stratification of units according to their principal activities should be based on the
ISIC Rev. 3.1 / NACE Rev. 1.1 classifications. The classification level largely depends on national
circumstances. Take as an example an economy specialised in the production of wood (Division 20 of
ISIC Rev. 3.1 / NACE Rev. 1.1). For this country a further subdivision at group or even class level might
be useful, in contrast to another economy where the production of wood is unimportant. However,
sampling strata should not be aggregated above the division level (second digit level of
ISIC Rev. 3.1 / NACE Rev. 1.1).

445. If regional aspects are of importance, the stratification should also include a regional dimension.
An appropriate regional classification should be used. See also the discussion of regional analyses in
Chapter 4.

446. The sampling fractions should not be the same for all strata. It is generally recommended that the
sampling fraction of a stratum should be higher for more heterogeneous strata (optimal allocation), and
also higher for smaller strata. The sampling fractions should be up to 100 per cent in strata with only a few
units, as may be the case in strata consisting of large units in certain industries (or certain regions). The
size of the units should also be taken into consideration by following the approach of pps-sampling and
thus using lower sampling fractions in strata with smaller units. Also, the units in each stratum may be
sorted by size or turnover and then sampled systematically. Another factor, which should be taken into
account when fixing the individual sampling fractions, is the expected response rate in each stratum.

3.5 Panel data surveys

447. The standard approach for innovation surveys is repeated cross sections, where a new random
sample is drawn from a given population for each innovation survey. An alternative or supplementary
approach is to impose an explicit panel data structure, where a given sample of units is surveyed more
frequently and in every subsequent survey using the same set of questions.

448. Panel data provide the opportunity to follow the development of the innovation process at a
microeconomic level over time. In particular, it allows the analysis of effects of various innovation
indicators on economic variables such as sales, productivity, exports and employees over time.

449. Panel data surveys can be conducted parallel to larger cross-sectional innovation surveys.
However, a number of guidelines should be followed:

• Units should be integrated with full-scale cross-sectional surveys in years where both are
conducted, in order to reduce burdens on units and to ensure an acceptable level of consistency
between the results from the two surveys.

• Panels should be constructed in such a way that they do not affect the main cross-sectional
survey.

91
 • If possible, information from other surveys on employment, sales, value-added, and investment
should be linked to the panel survey as well as the larger cross-sectional innovation survey for
empirical analyses.

3.6 Survey methods and suitable respondents

450. A variety of methods can be used to conduct innovation surveys, including postal surveys and
personal interviews. These methods each have different strengths and weaknesses. Postal surveys are well
established and comparatively less expensive, but may present problems as well. Several reminders
including telephone reminders are usually necessary to increase response rates to an acceptable level.
Additional action can be taken to increase response rates further: contacting respondents prior to
conducting the survey, sending a cover letter from the minister, sending basic results of previous
innovation surveys, providing the possibility of reporting to an “intelligent questionnaire” via the Internet
or a promise to send respondents the main findings from the current survey55.

451. Many of the problems with postal surveys can be avoided when data are collected by personal
interview, using for example, CATI or CAPI techniques. Interviewers can give guidelines on answering the
questionnaire. The quality of the results for CAPI (Computer Assisted Personal Interviews) techniques is in
general expected to be higher and item non-response rates are expected to be lower. However, CAPI
methods in particular are more costly than postal surveys.

452. A drawback to CATI (Computer Assisted Telephone Interviews) interviews is in collecting

quantitative data on innovation activity. This generally will take time to calculate, so that respondents may
not be able to answer the entire questionnaire in a single telephone call. Additionally, in large units,
questionnaires are answered jointly by different offices or branches, implying that several telephone calls
may be needed to complete a single questionnaire.

453. An alternative approach is the use of online or automated data exchange collection technologies.
A difference with this approach concerns the complexity of the questionnaire and the use of filtering
questions. In a paper questionnaire respondents will see all the questions and are able to change their
responses to a filtering question. An electronic questionnaire can be designed so that respondents do not
see all questions and therefore will not alter their answers in response to seeing additional information (the
same situation can apply to CATI and CAPI formats). One option could be to allow all respondents to see
the entire questionnaire, including those questions that respondents may not be required to answer. Other
issues with this approach are confidentiality and continuity (e.g. returning to the questionnaire several
times before it is finished).

454. Choosing the most suitable respondent in the units is particularly important in innovation
surveys, as the questions are highly specialised and can be answered by only a few people in the unit, often
not the ones who complete other statistical questionnaires. In small units, managing directors will often be
good respondents. In larger units, several people will often be involved, but one must be responsible for
co-ordinating the replies. A special effort to identify respondents by name, before data collection
starts, is highly recommended.

55
Further guidelines for improving response rates for postal surveys can be found in Dillman, D., The Total
Design Method, Wiley, 1978, and in Moore, D. and Baxter, R., Increasing Mail Questionnaire Completion
for Business Populations: The Effects of Personalization and a Telephone Follow-up Procedure as
Elements of the Total Design Method, 1993.

92
3.7 The questionnaire

455. Some basic rules should be followed when designing the questionnaire for an innovation survey.
Each questionnaire should be tested before it is used in the field (pre-test). Pre-testing can include both
interviewing a group of managers or experts concerning their understanding of the draft questionnaire, and
sending the questionnaire to a small sample of units. Both these steps can be valuable for improving the
quality of the questionnaire.

456. The questionnaire should be as simple and short as possible, logically structured, and have clear
definitions and instructions. Generally, the longer the questionnaire, the lower the unit and item response
rates. This effect can be minimised by devoting special attention to the design and layout and by giving
clear and sufficient explanatory notes and examples. It is particularly important to design the questionnaire
in such a way that units with no innovation activities will nonetheless reply and answer the questions that
are relevant for them.

457. Respondents’ understanding of the questionnaire may well increase as they move from question
to question. This means that their answers may depend on the order of the questions. Adding or deleting a
question may influence subsequent answers.

458. Questions on a number of qualitative indicators can either utilize a binary scale (yes or no), or an
ordinal scale in which enterprises are asked whether the factor is relevant and, if so, its importance. The
binary scale has the advantage of being very simple and reliable, but it only provides limited information
on the factors considered. However, it may introduce a high degree of subjectivity if the answer cannot be
based on facts, due to differences in the interpretation of the question. An ordinal scale allows the ranking
of factors in terms of their importance, though this also introduces a degree of subjectivity. Analytical
methods are available, however, to minimize such problems with ordinal response scales.

459. In the case of international innovation surveys, special attention should be given to the translation
and the design of the questionnaire. Even minor differences between national questionnaires can restrict
the comparability of the results. Such differences can stem, for example, from translation, from changes in
the order of questions, or from adding or deleting categories. A sound translation taking account of
particular local circumstances (such as a country’s legal system) will help avoid misunderstandings of
concepts and definitions.

3.7.1 Short form questionnaires

460. For many small units and units in sectors with little innovation activity, the response burden for a
full innovation questionnaire may be quite large relative to their innovation activity. Unit non-response
rates may also be higher for these units. In these cases, reduced form survey questionnaires, focussing on a
set of core questions, can be useful. Short form questionnaires can also be used in surveying units that have
not reported innovation activity in previous innovation surveys. Conversely, for individual units in the
above-mentioned groups (small units or less-innovative sectors) that have previously reported substantial
innovation activity, full questionnaires may be used.

3.8 Innovation and R&D surveys

461. R&D and innovation are related phenomena which can lead some countries to consider the
combination of R&D and innovation surveys. There are a number of points for and against:

93
 • With a combined survey, the overall response burden of the reporting units will be reduced (a
single questionnaire, instead of two separate surveys asking partly the same questions).

• However, if the length of the questionnaire for combined surveys is much longer than for a
separate survey, this could lead to a decline in response rates.

• A combined survey offers scope for analysing the relations between R&D and innovation
activities at the unit level. There is less scope for this with separate surveys, especially when
different institutions carry them out.

• There is a risk that units not familiar with the concepts of R&D and innovation can confuse them
in a combined survey.

• Combined surveys offer an efficient method of increasing the frequency of innovation surveys.

• Country experiences (for example, in Denmark, Finland, the Netherlands, Norway and Spain)
indicate that it is possible to obtain reliable results for R&D expenditures in combined surveys.

• The frames for the two surveys will generally be different. For example, the frame population for
innovation surveys may cover industrial classifications (and small units) that are not included in
R&D surveys. Combining them might involve sending questions about R&D to a large number
of non-R&D performers that are included in the frame population for the innovation survey, and
this would increase the cost of the joint survey.

462. In principle, business surveys other than R&D can also be merged with innovation surveys. Some
experiences in merging innovation surveys with structural business surveys have been conducted (for
example, in Bulgaria, Italy and the Netherlands). Additionally, business surveys on the diffusion of
information and communication technologies, and on the adoption of knowledge management practices
can be considered for integration with innovation surveys.

463. While the use of combined surveys is not recommended in this Manual, country experiences
indicate that they provide a feasible option for increasing the frequency of data collection. Some guidelines
for conducting combined surveys are:

• In order to reduce the risk of conceptual confusion between R&D and innovation, two distinct
sections of the questionnaire should be used. A separate section should also be used when
combining innovation with other types of surveys.

• To avoid declines in response rates, individual sections for R&D and innovation should be
smaller than for separate surveys, so that the overall length of the combined survey is comparable
to that of a separate survey.

• Comparisons of results from combined surveys with those from separate innovation surveys
should be done with care, and surveying methods should be reported.

• Samples to carry out such surveys should be extracted from a common business register in order
to avoid inconsistencies in the frame populations.

94
4. ESTIMATION OF RESULTS

4.1 Weighting methods

464. The results of sample surveys need to be weighted to obtain information that is representative for
the target population. There are various methods for weighting sampling results. The simplest is
weighting by the inverse of the sampling fractions of the sampling units, corrected by the unit
non-response. If a stratified sampling technique with different sampling fractions is used, weights should
be calculated individually for each stratum.

465. The weights can be further refined by calibration if the frame population includes some
quantitative or qualitative information on all units, for instance number of employees, turnover, legal status
or region. The calibration will ensure that the weighted sample sums to the total population or distribution
and in that way increase precision and reduce bias. Effective calibration software, in particular CLAN from
Statistics Sweden, CALMAR from INSEE (France) and CALJACK from Statistics Canada, are available
for use by other countries.

466. Weights are most commonly based on the number of enterprises in a stratum. However, it may be
beneficial for quantitative variables to weight the results by number of employees or turnover. In
international and other comparisons it is important to ensure that the same weighting method is used.

4.2 Non-response

467. In practice the responses to innovation surveys are always incomplete, irrespective of the survey
method used. Two types of missing values can be distinguished: item and unit non-responses. Unit
non-response means that a reporting unit does not reply at all. Possible reasons are, for example, that the
surveying institute cannot reach the reporting unit or that the reporting unit refuses to answer. In contrast,
item non-response refers to the response rate to a specific question and is equal to the percentage of blank
or missing answers among the reporting units. Item non response rates are frequently higher for
quantitative questions than for questions using binary or ordinal response categories.

468. Item and unit non-responses would be less a problem if the missing values were randomly
distributed over all sampling units and all questions. In reality, however, both types of missing values may
be biased with respect to certain characteristics of the population and the questionnaire.

469. Disregarding missing values and applying simple weighting procedures based only on the
responses received implicitly assumes that non-respondents are distributed in the same way as respondents.
If non-respondents do not follow the same distribution, for example if non-response units have a lower
propensity to innovate, then this practice will give biased results.

470. A number of methods can be used to minimize the problems of non-response. As different
methods may lead to different results, some general guidelines should be followed. An appropriate first
step in dealing with missing values is to contact the respondent to collect the missing information.

471. For practical as well as theoretical reasons, one suggested way to minimize the problem of item
non-response is to use imputation methods to estimate missing values on the basis of additional
information. The idea is that the use of additional information will allow more accurate estimates of
missing values than simply using the mean observed value and minimize non-response bias.

95
472. Among imputation methods, cold-deck techniques can be employed first, in which missing
information is estimated using data from other statistical surveys (including previous surveys) or from
other related sources. For any remaining missing values, surveys may consider hot-deck methods.
Hot-deck methods cover a large variety of options such as replacing the missing values for each variable
by the mean of the strata, by predicting the value using regression techniques, or by using nearest
neighbour techniques where the missing values are replaced by the values of the unit which is most similar
with respect to other relevant variables. The decision upon the most appropriate hot-deck method should
also be based on the type of variable (quantitative versus qualitative variables).

473. The choice of method to treat the problem of unit non-response will depend on the level of
non-response. If the non-response rate is fairly low,56 the weighting should be calculated on the basis of
the units that replied. This procedure is based on the assumption that the innovative behaviour of
responding and non-responding units is identical. This assumption could be tested through a non-response
analysis. Even if the assumption is wrong, the bias introduced can be disregarded as long as the fraction of
non-responding units is fairly small.

474. In contrast, if the unit non-response rate is very high, no method can be recommended to solve
the problem. In such a case the results of the innovation survey can only be used as case studies. No
conclusions should be drawn about the target population in general, as the bias may be too high.

475. In all other cases, i.e. when the unit non-response rate is beyond a lower threshold but less than
an upper threshold, some more complicated and partly more expensive techniques can be used. One
solution would be to select reporting units that have answered randomly until the response rate is 100 per
cent, i.e. to use the results of randomly selected units twice or even more often.

476. Other methods are based on the results of a non-response analysis. The objective of
non-response analysis is to obtain information on why reporting units did not answer. Non-reporting units
should be contacted by phone or by mail (using a very simple questionnaire not exceeding one page) and
should be asked for general information such as their sector of activity and size (if not already available
from other sources), the reason they did not answer, and for answers to a few key questions in the original
survey to see whether the results are biased. This information can then be used to adjust the weights. The
results of the non-response analysis should only be used if the response rate is very high.

5. PRESENTATION OF RESULTS

477. The results of innovation surveys can be used either for descriptive or for inferential purposes.
The objective of descriptive analysis is to describe the statistical units in terms of their innovative or
non-innovative activities without any conclusions for the underlying survey or target population (if it is not
a census). In this type of analysis the results are taken without further weighting, as they were observed for
the individual units. No generalisation of the results at the level of the survey or target population is
possible, because the figures only refer to the participating units. For this kind of analysis, the unit
non-response rate is of minor importance.

478. In contrast, the objective of inferential analysis is to draw conclusions about the target
population. In this case the results should give a representative estimation of the situation for the observed
and unobserved statistical units taken together. Inferential analysis requires weighted results. For this type

56
It is difficult, if not impossible, to define when a unit non-response rate is deemed to be high or low.
However, it is generally acknowledged that the higher the unit non-response rate, the lower the
comparability of the results of innovation surveys.

96
of analysis, the unit non-response rate is of great importance: if the unit non-response rate exceeds a
certain threshold, the possible bias may be so large that inferential analysis is useless.

479. As mentioned above, most innovation surveys are carried out as random sample surveys. The
results of these surveys will include two types of error: random errors due to the random process used for
the selection of the units, and systematic errors containing all non-random errors (bias). To get at least an
idea of the variance for the results, it is recommended to not only calculate (average) values for
innovation indicators, but also their coefficients of variation and/or confidence intervals. Such
intervals include the true but unknown values in the survey population with a very high probability,
assuming no bias. Standard errors give a lower threshold for the unknown total error of the indicators
under consideration.

480. The presentation of results should contain meta-data, including information on the procedure
used in collecting data, sampling methods, procedures for dealing with non-response, and quality
indicators. This will allow users to better interpret the data and judge its quality.

6. FREQUENCY OF DATA COLLECTION

481. Theoretical and practical considerations, as well as user needs at the international, national and
regional level, determine the frequency of innovation surveys. The increasing importance of innovation
for the growth of economies requires more frequent and more up-to-date data. From this viewpoint,
information on innovation activities should ideally be collected on an annual basis. This view is
strengthened by theoretical considerations indicating that innovation activities come in waves, making the
results of non-annual surveys very dependent on the time at which the survey is carried out. Only a few
countries, however, can afford or are willing to run innovation surveys every year.

482. Taking into account both practical considerations and user needs, it is recommended that
innovation surveys be conducted every two years. However, where this is not economically feasible, a
frequency of 3 or 4 years may be chosen.

483. To ensure comparability between respondents, surveys must specify an observation period for
questions on innovation. The choice of the length of the observation period is a compromise between
different requirements. A long observation period allows the collection of data on intermittent innovation
activities and the effects of innovations. For example, firms with products that have longer life cycles may
innovate less frequently. Conversely, a short observation period improves respondent recall and the
accuracy of the results. For longer observation periods, organisational memory can be poor, due to either
staff turnover or the recall of respondents. Other issues concern the relation between the collection
frequency and the observation period. An observation period that is longer than the collection frequency
(creating overlap in the coverage of innovation surveys) has some drawbacks. Due to the overlap in
coverage, it might be difficult to attribute innovation fully to the time period since the last survey. This
overlap can also affect comparison of results over time, since it can be unclear whether changes in results
are mainly due to innovation activities in the period since the last survey or in the year that is also covered
in the previous survey. As stated in §227 in Chapter 3, it is recommended that the length of the
observation period for innovation surveys should not exceed three years nor be less than one year.

97
 ANNEX 1: INNOVATION SURVEYS IN DEVELOPING COUNTRIES

1. INTRODUCTION

484. This Annex provides guidelines for the implementation of innovation surveys in developing
countries. The term “developing countries” is used here with the assumption that it does not refer to a
homogeneous set of countries, and that the content of this Annex needs to reflect the different
characteristics of economies and societies in a rapidly evolving “developing World”.

485. After the publication of the second edition of the Oslo Manual, many developing countries
conducted innovation surveys, in various regions of the World. The design of these surveys was usually
intended to comply with Oslo Manual standards. However, almost every innovation measurement exercise
in developing countries resulted in adaptations to the proposed methodologies, in order to be able to
capture the particular characteristics of innovation processes in countries with different economic and
social structures than the more developed OECD countries. Adaptations were prepared by each country
separately and with different approaches. Outside the OECD and EU, the first effort to compile these
particularities and guide the design of cross-nationally comparable innovation surveys was conducted in
Latin America by RICYT (Ibero American Network on Science and Technology Indicators – Red
Iberoamericana de Indicadores de Ciencia y Tecnología), resulting in the publication of the Bogotá
Manual. This Manual was later applied in most innovation surveys conducted in Latin American countries,
and extended to other regions. The importance and impact of this standard setting work inspired the
production of this Annex to the Oslo Manual.

486. Preparation of the Annex was co-ordinated by the UNESCO Institute for Statistics (UIS). A base
document provided by RICYT57 was submitted to a panel of researchers and practitioners with experience
in innovation surveys in developing countries58. The present Annex is based on the integrated conclusions
57
Gustavo Lugones and Fernando Peirano “Proposal for an Annex to the Oslo Manual as a Guide for
Innovation Surveys in Less Developed Countries Non-Members of the OECD”, Centro REDES/RICYT,
September 2004. This document was based on the contributions made as part of the activities of RICYT by
a group of Latin American experts (M. B. Baptista (DINACYT-Uruguay), J. E. Cassiolato
(IE/UFRJ-Brazil), M. Mainieri (SENACYT-Panama), F. Malaver Rodríguez and M. Vargas Pérez
(Comcyt/OCyT-Colombia), A. Martinez Echeverria (INE-Chile); M. Salazar Acosta (Simon Fraser
University, Canada).
58
The UIS panel was co-ordinated by Simon Ellis and Ernesto Fernández Polcuch, and included as
contributors the authors of the base document (Gustavo Lugones and Fernando Peirano, RICYT);
Pierre Tremblay, IDRC, Canada; Gao Changlin, and Jiancheng Guan, China; Javier Revilla Diez, Germany
(with experience in Thailand, Singapore and Penang State, Malaysia); Annamária Inzelt, Hungary;
Laxman Prasad, India; Antoine Zahlan, Lebanon; Fadzilah Ahmad Din (with Anita Bahari and
Dr. Cassey), MASTIC, Malaysia; Anna Ong, Penang State, Malaysia; Michael Kahn (with
William Blankley and Simon Mpele) and Tinus Pretorius (with Andre Buys), South Africa;
Bitrina Diyamett, Tanzania; and Patarapong Intarakumnerd, Thailand. Valuable comments were received
from the OECD Secretariat, as well as various NESTI members (particularly Carter Bloch and
Frank Foyn). However, contents of this annex are the sole responsibility of the UNESCO Institute for
Statistics and the publishers of the Oslo Manual.

98
of this exercise. Proposals and recommendations may be more or less applicable depending on the
characteristics of the region and countries concerned.

487. Recommendations in this Annex are mainly based on the experience of countries that have
already conducted innovation surveys, most of which pertain to the group of higher and medium income
countries of the developing world, where innovation has already become a policy issue. Nevertheless, the
knowledge gained by these countries should help other developing countries to acquire their own
experience without having to build exclusively on innovation measurement exercises carried out in
developed countries.

2. THE CHARACTERISTICS OF INNOVATION IN DEVELOPING COUNTRIES

488. It is widely accepted that dissemination mechanisms and incremental change account for most of
the innovation occurring in developing countries59. This is rooted in particular characteristics of the society
and the economy in many of these countries, which influence innovation processes in many ways.

Size and structure of markets and firms

489. The size and structure of firms and markets are key to understanding developing countries’
innovation processes. While the sector of small and medium enterprises is very significant in size
(including a large number of micro and small and, in some countries, medium size businesses, that are
often unregistered), even enterprises considered in most developing countries to be ‘big’ usually operate in
suboptimal production scales, with higher unit costs and far from optimal efficiency. Competitiveness is
mostly based on the exploitation of natural resources or cheap labour, rather than on a quest for efficiency
or differentiated products. This leads to the informal organization of innovation and fewer R&D projects.

490. Important market failures related to economies of scale and externalities present high barriers to
innovation. For instance, productive processes and more specifically innovation activities are subject to
indivisibilities and a lack of economies of scale, influencing the viability of R&D projects.

491. A number of exogenous systemic factors deeply shape the innovation landscape in developing
countries, such as: macroeconomic uncertainty; instability; physical infrastructure (sometimes lack of basic
services such as electricity or ‘old’ communications technologies); institutional fragility; lack of social
awareness about innovation; risk-averse nature of enterprises; lack of entrepreneurs; existence of barriers
for business start-up; lack of public policy instruments for business support and management training.

Instability

492. Instability among micro and small businesses may mean that some have good potential to
upgrade national innovative performance and function as cradles of innovators, while some lack resources
and support for any innovation. Macro level uncertainty limits any long term innovation activity.

59
For example, the first South African innovation survey found, for instance, that 86% of innovations in the
South African industry are of incremental nature.

99
Informality

493. Developing countries’ economies have an important degree of informal practice. Informality is
not a favourable context for innovation. The sometimes great creativity invested in problem-solving in the
informal economy is not applied systematically, and tends to result in isolated actions which neither
increase capabilities nor help establish an innovation based development path.

Particular economic and innovation environments

494. Many enterprises in developing countries operate in unusual economic and innovation
environments due to the existence and in some cases prevalence of state-owned enterprises (China) or
massive para-statal enterprises (some Arab states), where a lack of competition sometimes discourages
innovation or drains local markets of innovative potential, while in other situations big state-owned
enterprises (for example in sectors such as oil, aerospace or telecommunications) can become
technological leaders through important investments in experimental development work (as in some Latin
American countries). Moreover, in countries with less developed economic systems, major government
S&T policies and programmes may have more impact on innovation than the activities and strategies of
private enterprises.

495. Past techno-economic paradigms have a continuing economic importance in some cases, and in
other cases a paradigm switch is delayed due to high costs involved, insufficient supplies of local capital
and lack of credit for big technological investments.

496. Local markets in developing countries tend to be small, in some cases, due to a less developed
infrastructure, reducing the scope of the enterprise’s actions and the relevance of actual innovations (“new
to the market” may have a different meaning in such environments).

497. Innovations in the agricultural sector have a high economic impact, due to the sector’s significant
overall economic weight.

Reduced innovation decision-making powers

498. The dominant presence of externally-controlled or multinational corporations results in reduced

decision-making for local enterprises or subsidiaries (especially in the field of innovation), due mainly to a
international - division of functions in these organisations. In the last few years, this division has even been
extended to independent local enterprises in the framework of international manufacturing networks.
Technology transfer from multinational corporations and from abroad is therefore a fundamental source of
innovation.

Weak innovation systems

499. Fewer resources are devoted to innovation activities system-wide, thereby reducing innovation
potential of enterprises. The government is a major player in R&D execution and funding, mainly due to a
low level of resources devoted to R&D by businesses.

500. Flows of information within national systems of innovation are fragmented, and in some cases
there is an absence of linkages between science and enterprises. Weak or absent linkages challenge the

100
capacities of firms to overcome (technology-related) problems, and draw firms towards solutions implying
mostly acquisition of embodied technology.

501. Barriers to accumulation of capabilities by the enterprises are high and difficult to tackle,
particularly in the case of highly qualified human capital, local and international linkages and tacit
knowledge incorporated into organisational routines.

Characteristics of innovation

502. The acquisition of embodied technology (equipment) for both product and process innovation is a
major component of innovation.

503. Minor or incremental changes can be the most frequent type of innovation activity in some
developing countries, together with innovative applications of existing products or processes.

504. Organisational change is extremely significant in the innovation process. Besides its direct
impact on enterprise performance, it also contributes to the enterprise’s ability to absorb new technologies
incorporated in machinery and other equipments (the most frequent type of innovation). Heterogeneity
frequently prevails with regards to firms’ technological, organisational and managerial patterns, with ‘high
tech’ firms coexisting with informal businesses, and with many enterprises lacking a formal organisational
structure. This creates a large need for organisational change, often independent from product and process
innovation.

3. INNOVATION MEASUREMENT IN DEVELOPING COUNTRIES

505. Innovation measurement in developing countries has to produce results comparable to the ones
obtained in developed countries, applying the Oslo Manual, in order to enable benchmarking and construct
a coherent international system of innovation indicators. At the same time, innovation surveys need to
respect and be able to pick up the characteristics of innovation in developing countries presented in
section 2. Therefore, the definition of innovation, its subtypes (product, process, marketing and
organisational innovation), innovation activities and the innovative firm, as presented in Chapter 3, should
be applied in innovations surveys in developing countries..

506. While most of the issues presented in section 2 are addressed below, some still remain unsolved
from the measurement point of view. This is mainly due to the difficulty in applying existing definitions.
One of the main issues, also mentioned in section 5, is the problem of measuring incremental changes,
which may not result in "new or significantly improved" products or processes. Another pending
discussion is related to the scope of innovations, since concepts such as “new to the market” may have
different interpretations in environments with less developed infrastructure.

3.1 Specific needs for public policies and private strategies: the potentially innovative firms.

507. Innovation surveys in developing countries need to be based on methods and procedures that take
into account the aspects mentioned in section 1 of this Annex, in order to become useful instruments for
public and private decision-making. In developing countries the main reason for conducting innovation
surveys is to inform the public-policy-making process, and the design of business strategies, with the main
focus on the generation, diffusion, appropriation and use of new knowledge in businesses, and less priority
given to cross-country comparisons and benchmarking.

101
508. Measurement exercises should therefore focus on the innovation process rather than its outputs,
strengthening the way in which capabilities, efforts and results are dealt with. The efforts made by firms
and organizations (innovation activities) and the capabilities shown (stocks and flows) are, for this reason,
equally or even more important to know and analyze than the results obtained (innovations). Factors
hampering or facilitating innovation are seen as key indicators in this context.

509. One particular subject of interest in developing countries is the “potentially innovative firm”.
Innovation active firms are those “that have had innovation activities during the period under review,
including those with ongoing and abandoned activities”. Potentially innovative firms are a subset of these,
those that have made innovation efforts (i.e. conducted innovation activities), but have not achieved results
(innovations) during the period of analysis.

510. Inside this group, we might find businesses that have innovated in the past, or businesses that
might achieve an innovation in the near future. Nevertheless, considering that products and processes
become obsolete rapidly, the existence of a high number of potentially innovative enterprises could suggest
strong barriers to innovation, or premature attempts to innovate without assembling the necessary
resources. A key element in innovation policies in developing countries is to assist potentially innovative
firms to help them overcome the obstacles that prevent them from being innovative; that is, converting
their efforts into innovations.

3.2. Measurement priorities

511. Measurement priorities in developing countries have to take into account different
responses to common questions (why do we measure innovation, what should we measure, and how
should we measure), which result in different priorities when designing an innovation survey. The
first question refers to the aims or the main functions of these surveys. The remaining two questions help to
clarify the object we are going to measure as well as the most suitable methods and procedures for this.
This last question is closely related or even derived from the first one.

512. In developing countries, the responses sought from innovation surveys are less the number of
innovative enterprises, or even innovation counts, but information for public and private stakeholders to
analyse the various innovation strategies present in the innovation system under scrutiny, and to evaluate
and understand how these patterns contribute to strengthening the competitiveness of particular enterprises
and more generally to economic and social development. These data also allow for the construction of
different models of technological behaviour followed by enterprises.

513. This approach requires linking analysis at micro, meso and macro-economic levels, relating
innovation data to the technological content of exports; studying strengths and weaknesses of particular
industries or innovation systems in general; assessing the absorption capacity of innovation systems;
identifying networks; exploring the relationship between the formal education system and employment; as
well as obtaining indications about the effectiveness of different public instruments for supporting and
promoting innovation.

514. In order to construct indicators for potentially innovative firms, measurement instruments need to
address all the enterprises (i.e. both innovative and non-innovative), particularly when addressing the main
issues related to innovation strategies, such as innovation activities, obstacles, capabilities, linkages,
and results.

102
3.2.1 Innovation capabilities

515. Innovation capabilities are extremely helpful for describing the different stages in which firms,
and industrial sectors, can be classified in developing countries. The capabilities accumulated by the firm
are the main factor that allows the firm to take advantage of market opportunities. The most significant
innovation capability is knowledge accumulated by the firm, mainly embedded in human resources, but
also in procedures, routines and other characteristics of the firm. Innovation capabilities, as well as
technological capabilities, are the result of learning processes, which are conscious and purposeful, costly
and time-consuming, non-linear but path-dependent and cumulative. Because of its interactive,
technology-specific, and culturally-influenced nature, there is no single trajectory but a range of possible
development-paths, particularly in developing countries where innovativeness and entrepreneurship are
sometimes less frequently present and have particular characteristics.

516. Knowledge about innovation capabilities and the firm’s efforts to increase these are key to
understanding its present and future performance. Innovation capabilities condition the design of strategies
for the introduction of changes, improvements and/or innovations (innovation strategies). If innovation
strategies are at the heart of policy interest, innovation capabilities are the highest priority in the design of
an innovation survey in developing countries.

517. There are many difficulties in measuring innovation capabilities, since it implies measuring
knowledge that is not codified, but ‘stored’ in individual’s minds or organizational routines. At the same
time, it is not easy to find reliable data from firms about the exchange of knowledge with other agents or
organizations.

518. The priority given in developing countries to measuring innovation capabilities motivates
placing additional emphasis on a number of aspects of surveys:

• Human resources.
• Linkages.
• Information and communication technologies (ICTs), their incorporation and use.

519. There is also an increased need to examine more complex issues such as the types of
decision-making support systems put in place by the firm’s direction and management, as well as the
firm’s actual potential for knowledge absorption.

3.2.2 Expenditure on innovation activities

520. In order to measure firms’ innovation efforts appropriately, it is essential to understand the
intensity of innovation activities carried out. It is therefore recommended to obtain more details about
which innovation activities were undertaken by the firm in the reference period and, where deemed
feasible, to collect data on expenditure by innovation activity, as stated in Chapter 6. This indicator serves
as a powerful discriminator of firm’s behaviour and strategies. In order to explain firm development,
innovation expenditure needs to be complemented with more general information on the development of
the sector of economic activity in which the firm is active. This information could be obtained through
innovation surveys, if it is not readily available through other sources at National Statistics Offices.

103
3.2.3 Organisational innovation

521. The absorption of new technologies, mostly incorporated in machinery and other equipment, can
require significant organisational change for many enterprises in developing countries. Since innovation
in developing countries has a strong component of absorption of technologies generated in industrialized
countries, organisational change acquires substantive relevance. The enterprise’s behaviour in this field
becomes therefore a key element for explaining differences in performance and competitiveness.60

522. In order to gain additional information on the innovative capabilities of enterprises in developing
countries, questions on the implementation of organisational innovations can be supplemented by questions
on human resources and training, and the incorporation of ICTs. Both these elements would aid in
providing an indication of enterprises’ innovative capabilities.

4. PRINCIPAL ADAPTATIONS

523. There are three principal topics which might be taken into account in adapting innovation surveys
for developing countries: ICTs, linkages, and innovation activities.

4.1 ICTs in innovation surveys

524. The role of ICTs in innovation is related both to “front office” and “back office” applications (see
Annex Box 1). In developing countries, the incorporation of ICTs in enterprises has been frequently
limited to sophisticated “front office” applications (such as web-page, call-centre, e-mail, or digital
brochures and catalogues). However, it is considered that the main impact on the enterprise’s performance
can be obtained by implementing ICTs to support or automate critical activities or processes (“back
office”). Thus, innovation surveys in developing countries should focus on the use of ICTs. This is a
powerful tool for differentiating situations, since frequently even medium and large size firms do not have
a consolidated management system installed, constituting a serious obstacle for achieving better
performance in the different areas of the business.

Annex Box 1. “Front office” vs “Back office”

Front Office applications include those focused on sales and marketing, customer self service, web portals, and
call centers. The common thread is a high level of employee or customer interaction.

A back office is a part of most corporations where tasks dedicated to running the company itself take place. Back
office activities are the internal operations of an organisation that support core processes and are not accessible or
visible to the general public. Back office applications support or automate critical activities or processes.

525. This analysis of ‘deeper’ dimensions of ICT implementation in firms would help to expose the
differences between enterprise characteristics in developing countries and in industrialized ones, especially
in those medium-income countries where this difference might not be so evident when analysing only the
more ‘superficial’ dimension of front office ICTs. It is still not possible to carry out a definitive statistical
analysis on the relation between deeper and more complex integration of ICTs and business performance in

60
In the particular case of many Latin American countries, the need for firms to permanently adapt and adjust
to recurrent alterations in the economic context reinforces the idea that organisational change is an essential
dimension of firms’ competitiveness.

104
developing countries. However, partial evidence and case studies have shown the potential for further
research, particularly through innovation surveys. Further evidence will therefore help to clarify the
relationship between ICTs and innovation, complementing the more frequently available literature on “ICT
and productivity”.

526. Taking into account the fact that questionnaires should be kept as short and simple as possible, in
cases where no specific surveys on ICTs in businesses are available, innovation surveys should enquire
about available infrastructure; use of ICTs (separating front- and back-office activities) and ultimate
purpose of ICT usage; existence of internal ICT management and development capabilities; ICT
expenditure and its relationship with organisational innovation.

4.2 Linkages

527. Following the recommendations in Chapter 5, strong emphasis should be put on measuring
linkages. In order to enable the weighting of the different linkages the firm is involved in, a proxy measure
of complexity can be developed by crossing ‘type’ and ‘objective’ of the linkages. This could be done by
establishing a matrix of linkage agents (i.e. universities, technical and vocational training institutions,
technological centres, test labs, suppliers, clients, head office, enterprises belonging to the same group,
other firms, consultants, R&D firms, public S&T agencies), and types of linkage (including open
information sources, acquisition of knowledge and technology, and innovation co-operation, supplemented
also by other complementary activities, particularly access to new sources of financing, and access to
commercial information).

528. The innovation success of developed countries is related to the fact that they offer regional
environments in which firms are able to get the knowledge sources they need, combined with a wide
access to leading global knowledge-bases. Firms in developing countries most frequently do not have
access to high-end knowledge, and the local environment is therefore even more important for them.
Consequently, it is recommended to include questions on the geographical location of linkages. One
possible break-down includes local, regional, national and international locations.

4.3 Innovation Activities

529. The need to focus on the activities and capabilities of the firms will lead to more attention on
some specific innovation activities, as classified in Chapter 6. In order to comply with the priority criteria
presented, it is recommended to include activities such as:

• “Hardware purchase”, and “Software purchase” (separately, not just included in “Acquisition of
machinery, equipment and other capital goods”).

• “Industrial design”, and “Engineering activities” (separately, not just included in “Other product
and process development”).

• “Lease or rental of machinery, equipment and other capital goods”.

• “In-house software system development”.

• “Reverse engineering”.

105
4.4 Additional adaptations

530. It is recommended to collect data on human resources, both from the perspective of its
composition (by qualification, type of occupation – see Chapter 6 - and sex), and its management. From
the point of view of HR management, it is particularly important to collect information on actions taken by
firms with regards to training, including the resources involved. In order to gain information on the
innovative capabilities of enterprises, data can be collected not only on training activities that are linked to
innovations, but also on general training in areas such as management and administrative training, ICT,
industrial security, quality control.

5. METHODOLOGICAL ISSUES FOR DEVELOPING COUNTRY CONTEXTS

5.1 Information systems specificities

531. The design and planning of innovation surveys in developing countries needs to take into account
the relative weakness of statistical systems. Linkages between different surveys and data sets tend to be
weak or nonexistent, preventing the use of information from other surveys both in the design of the
exercise, and in the analysis of its results. The weakness in, or sometimes lack of official business registers
which are normally used as sample frames is another example of this type of problem.

532. It is particularly important that National Statistics Offices (NSOs) are involved in innovation
surveys, even if these surveys may not constitute a high priority in national statistical programmes in some
developing countries. The involvement of NSOs might sometimes include the signing of formal
agreements between the various institutions involved (frequently including government departments, and
universities). The participation of NSOs in innovation surveys brings to the table experience in the design
and application of surveys, and allows the resolution of problems in registers and other background
information. It also helps to obtain higher response rates, and raises the potential to make the survey
compulsory. Moreover, if the sample used is the same as for other economic surveys, it presents the
possibility of wider ranging analysis.

533. Statistical systems in developing countries frequently lack the necessary information about firm
performance (such as data on sales, investments, exports), only have outdated data, or cannot provide data
in a suitable form for statistical analysis. In those cases, a number of basic variables can be included in the
innovation survey in order to enable analysis of the relationship between actions taken by firms on
innovation, and market performance (competitiveness). However, a tension between the need for extra
questions and response burden exists. In order to maximize response rates and improve coverage, a balance
should be sought, taking into account that operational simplicity and fluidity may be detrimental to the
analytical potential of the exercise.

5.2 General methodological considerations

5.2.1 Application of the survey

534. Interviews made in person (instead of surveys by mail or by phone) and by adequately trained
staff (for instance, undergraduate or graduate students) are recommended, since they have proved to have a

106
positive impact on the response rate and on the quality of the results obtained.61 This is particularly
important in developing country contexts where postal services may not be reliable. Moreover, interviews
conducted by qualified staff provide the respondent with immediate and relevant assistance for completion
of the questionnaire, improving the quality of the results.

5.2.2 Questionnaire design

535. The questionnaire can be designed to allow different sections to be separated in order for
different persons in the firm to reply to different sections. This is particularly valid for questions related to
general economic data of the firm, which might be provided by the finance division, or specific questions
about the innovation process, which might be completed by the product or plant manager. This might result
in more reliable information.62. It is nevertheless important to avoid delaying the survey or losing a
partially completed questionnaire due to this strategy.

536. It is also advisable to include guidance in the main questionnaire, in order to help the respondent
understand better and avoid losing the instructions. It is important to recognize that in some developing
countries respondents might not know the concept (or even the word) of “innovation”. If this might be the
case, questions have to include definitions.

537. Special attention has to be given to the language in which the questionnaire is written. The
wording needs to be adapted to meet the knowledge and experience of an “average” respondent. In some
cases, there will be a need to present questionnaires in more than one language, if it would assist
respondents.63

5.2.3 Frequency and other recommendations

538. While Chapter 8 of the Manual recommends that innovation surveys be conducted every two
years, it is recognized that in developing country contexts this should rather happen every three or four
years. If possible, timing of the innovation surveys should coincide with the major international innovation
surveys, such as the Community Innovation Survey (CIS) rounds in Europe, in order to obtain comparable
data for similar periods of time. It is also convenient to update a minimum set of variables every year (the
main quantitative ones, for example), if resources permit. A less costly strategy for this is to attach a
significantly reduced questionnaire to an existing business survey.

539. As stated in Chapter 8, the results of the innovation surveys should be published and distributed
widely, in order to encourage further participation by businesses in future rounds and to increase awareness
and use by researchers and policy makers. Diffusion mechanisms need to be included in the budget early in
the exercise.

540. A frequent difficulty in obtaining reliable information on innovation in countries with a less
developed statistical tradition is the difficult environment created by the low appreciation of the
importance of innovation and its public policy instruments by business. Managers can frequently be

61
This was found to be of particular importance in Latin American, and also in Africa, where experience has
shown that the response rate to mailed questionnaires is extremely low.
62
The multiple informants approach has proved to increase reliability and validity of the innovation survey in
China.
63
In the case of the Thai survey, questions were presented in English and Thai, since it was considered that
sometimes people can understand technical terms in English easier than in their native languages.

107
secretive on finance, and qualitative information can sometimes be more reliable than quantitative
information. Surveys need to be very clear in their purpose and questions. Under these circumstances, an
adequate legislative base for the collection of innovation statistics can be another factor to foster the
success of such an exercise. In some cases, simplified questionnaires could be designed to cover small
firms, in order to encourage their participation in innovation surveys.

6. THINKING AHEAD

541. A number of important questions regarding innovation measurement in developing countries

remain unsolved. However, various approaches have been tried out in different countries and deserve
further research, involving issues such as:

• The role of entrepreneurs and their attitudes towards innovation.

• The intention to capture innovations driven by factors other than market forces, and in particular
innovations conducted by the public sector.64

• The adaptation of methodology to measure innovation in the primary sector (particularly in

agriculture).

• The development of indicators reflecting sub-national (regional) innovation systems.

542. The application of the suggestions in this Annex will lead to a broader experience on innovation
surveys in developing countries. It is expected that some countries with better S&T statistics experience
will soon go beyond conducting R&D surveys and include innovation surveys in their statistical
programmes. The consolidation of standards, concepts, formats and other issues better suited for
developing countries, such as the ones proposed in this Annex, should help building awareness and
capacity. Further efforts in direct capacity building for innovation surveys will be key to the success of this
endeavour.

64
As proposed in Salazar M. and Holbrook, A., “A debate on innovation surveys”, Science and Public
Policy, 31, 4, 2004.

108
 ANNEX 2. EXAMPLES OF INNOVATIONS

1. INTRODUCTION

543. This Annex provides a list of examples for each type of innovation. These lists are intended as
illustrations of examples and should not be considered exhaustive by any means. They are intended to give
survey practitioners a better understanding of each innovation type, though they are not designed to be
shown to enterprises as examples of innovations. There are two reasons for this. First, their inclusion might
bias firms into excluding other innovations that are not on the list. Second, the list is dated, with many new
innovations impossible to foresee. It is also worthwhile emphasizing that two central criteria for
innovations are that they represent significant changes and that they are new to the firm. Thus, a change
can be an innovation for one firm and not for another. Often, more detailed descriptions are needed to
determine whether a change is to be classified as an innovation – and which type.

2. EXAMPLES OF INNOVATIONS

544. A product innovation is the introduction of a good or service that is new or significantly
improved with respect to its characteristics or intended uses. This includes significant improvements in
technical specifications, components and materials, incorporated software, user friendliness or other
functional characteristics.

545. Product innovations exclude the following:

• Minor changes or improvements.

• Routine upgrades.
• Regular seasonal changes (such as for clothing lines).
• Customisation for a single client that does not include significantly different attributes compared
to products made for other clients.
• Design changes that do not alter the function, intended use or technical characteristics of a good
or service.
• The simple resale of new goods and services purchased from other enterprises, though goods and
services developed and produced by foreign affiliates are included.

546. Examples of product innovations:

Goods

• Replacing inputs with materials with improved characteristics (breathable textiles, light but
strong composites, environmentally-friendly plastics, etc).
• Global Positioning Systems (GPS) in transport equipment.
• Cameras in mobile telephones.
• Fastening systems in clothing.

109
 • Household appliances that incorporate software that improves user friendliness or convenience,
such as toasters that automatically shut off when the bread is toasted.
• Anti-fraud software that profiles and tracks individual financial transactions.
• Inbuilt wireless networking in laptops.
• Food products with new functional characteristics (Margarine that reduces blood cholesterol
levels, yoghurts produced using new types of cultures, etc.).
• Products with significantly reduced energy consumption (energy efficient refrigerators, etc.).
• Significant changes in products to meet environmental standards.
• Programmable radiators or thermostats.
• IP (Internet protocol) telephones.
• New medicine with significantly improved effects.

Services

• New services that significantly improve customers’ access to goods or services, such as home
pick-up and drop-off service for rental cars.
• DVD subscription service where for a monthly fee customers can order a pre-defined number of
DVDs via the Internet with mail delivery to the home, with return via a pre-addressed envelope.
• Video on demand via broadband internet.
• Internet services such as banking, or bill-payment systems.
• New forms of warranty, such as an extended warranty on new or used goods, or bundling
warranties with other services, such as with credit cards, bank accounts, or customer loyalty
cards.
• New types of loans, for example variable rate loans with a fixed rate ceiling.
• Creation of web sites on the Internet, where new services such as product information and
various support functions can be offered to clients free of charge.
• The introduction of smart cards and multipurpose plastic cards.
• A new, self-service bank office.
• Offering customers a new “supply control system” which allows clients to check that deliveries
from contractors meet specifications.

547. A process innovation is the implementation of a new or significantly improved production, or

delivery method. This includes significant changes in techniques, equipment and/or software.

548. Process innovations exclude the following:

• Minor changes or improvements.

• An increase in production or service capabilities through the addition of manufacturing or
logistical systems which are very similar to those already in use.

549. Examples that are process innovations:

Production

• Installation of new or improved manufacturing technology, such as automation equipment or

real-time sensors that can adjust processes.
• New equipment required for new or improved products.
• Laser cutting tools.
• Automated packaging.
• Computer-assisted product development.

110
 • Digitisation of printing processes.
• Computerized equipment for quality control of production.
• Improved testing equipment for monitoring production.

Delivery and operations

• Portable scanners/computers for registering goods and inventory.

• Introduction of bar-coding or passive radio frequency identification (RFID) chips to track
materials through the supply chain.
• GPS tracking systems for transport equipment.
• Introduction of software to identify optimal delivery routes.
• New or improved software or routines for purchasing, accounting or maintenance systems.
• Introduction of electronic clearing systems.
• Introduction of automated voice-response system.
• Introduction of electronic ticketing system.
• New software tools designed to improve supply flows.
• New or significantly improved computer networks.

550. A marketing innovation is the implementation of a new marketing method involving significant
changes in product design or packaging, product placement, product promotion or pricing.

551. Marketing innovations exclude the following:

− Changes in product design or packaging, product placement, product promotion or pricing that
are based on marketing methods that have previously been used by the enterprises.
− Seasonal, regular and other routine changes in marketing instruments.
− The use of already applied marketing methods to target a new geographical market or a new
market segment (e.g. socio-demographic group of clients).

552. Examples of marketing innovations:

− Marketing innovations can refer to any marketing method (product design/packaging,

placement, pricing, promotion) as long as it is used for the first time by the firm.

Design and packaging

− Implementation of a significant change in the design of a furniture line to give it a new look and
widen its appeal.
− Implementation of a fundamentally new design of bottles for a body lotion intended to give the
product a distinctively exclusive look.

Placement (sales channels)

− First-time introduction of product licensing.
− First-time introduction of direct selling or exclusive retailing.
− Implementation of a new concept for product presentation such as sales rooms for furniture that
are designed according to themes, allowing customers to view products in fully decorated
rooms.
− Implementation of a personalised information system, e.g. obtained from loyalty cards, to tailor
the presentation of products to the specific needs of individual customers.

Pricing

111
 − Introduction of a new method where customers are able to choose desired product specifications
on the firm’s website and then can see the price for the specified product.
− First time use of a method for varying the price of a good or service according to its demand.
− First-time use of in-store special offers that are only accessible to holders of the store’s credit
card or reward card.

Promotion
− First-time use of trademarks.
− First-time use of product placement in movies or television programmes.
− Introduction of a fundamentally new brand symbol intended to position the firm’s product on a
new market.
− First-time use of product seeding through opinion leaders, celebrities, or particular groups that
are fashion or product trend setters.

553. An organisational innovation is the implementation of a new organisational method in the

firm’s business practices, workplace organisation or external relations.

554. Organisational innovations exclude the following:

− Changes in business practices, workplace organisation or external relations that are based on
organisational methods already in use in the firm
− Changes in management strategy, unless accompanied by the introduction of a new
organisational method
− Mergers with, and the acquisition of other firms

555. Examples of organisational innovations:

− Organisational innovations can refer to any organisational method in a firm’s business practices,
workplace organisation or external relations as long as it is used for the first time by the firm.

Business practices
− Establishing of a new database of best practices, lessons and other knowledge so that they are
more easily accessible for others.
− First-time introduction of an integrated monitoring system for firm activities (production,
finance, strategy, marketing).
− First-time introduction of management systems for general production or supply operations,
such as supply chain management, business re-engineering, lean production, quality-
management system.
− First-time introduction of training programmes to create efficient and functional teams that
integrate staff from different backgrounds or areas of responsibility.

Workplace organisation
− First-time implementation of decentralised job responsibility for the firm’s workers, such as
giving substantially more control and responsibility over work processes to production,
distribution or sales staff.
− First-time establishment of formal or informal work teams to improve the access and sharing of
knowledge from different departments, such as marketing, research, production.
− First-time implementation of an anonymous incident reporting system to encourage the
reporting of errors or hazards in order to identify their causes and reduce their frequency.

External relations

112
− First-time introduction of quality control standards for suppliers and subcontractors.
− First-time use of outsourcing of research or production.
− First-time entering into research collaboration with universities or other research organisations.

113