Managing project interdependencies: exploring new approaches

Killen, C P1, Krumbeck, B1, Kjaer, C2 and Durant-Law, G A3

1
University of Technology, Sydney, NSW, Australia
2
Optimice, NSW, Australia
3
University of Canberra, Canberra, ACT, Australia

1. Introduction

The interdependencies between projects create complexities for the management of project portfolios within
organisations. In times of uncertainty this challenge is even greater due to the difficulties in predicting the
flow-on effects from changes to projects in the portfolio. Hence, in times of disruptive change a good
understanding of project interdependencies is particularly important.

This paper outlines two related studies that aim to improve the understanding and management of
interdependencies within project portfolios. The paper first defines project portfolio management (PPM) and
highlights its growing importance for optimising organisational outcomes, especially in dynamic
environments. Project portfolio complexity and interdependencies between projects in a portfolio are then
overviewed, highlighting the challenges that these interdependencies create for effective PPM, and
introducing some of the methods used for understanding and managing these interdependencies including the
dependency matrix and the related design structure matrix. Network analysis and mapping tools are then
introduced and suggested as a novel method for improving understanding and managing project
interdependencies. Finally, an example of the use of this type of method is presented and the current research
projects are overviewed.

2. Project portfolio management (PPM)

The management of the project portfolio is central to organisational strategy and competitiveness as
organisations aim to gain maximum value from project investments, and as most organisational innovation is
done through projects. PPM capabilities aim to provide a holistic and responsive framework for the
management of the project portfolio.

Projects are temporary endeavour undertaken to meet specific goals such as the development of new
products or services or the implementation of organisational change [1, 2]. Organisations invest large
amounts in these projects, however many projects are not successful. Research on project outcomes report
success rates between 30 and 60 percent [3, 4, 5, 6]. Therefore there is significant scope for improved project
success rates, and organisations actively seek new methods that may boost the return on their project
investments. In order to maximise the return on project investments, organisations place high importance on
PPM and embark on significant learning processes to enhance their capabilities [7].

PPM processes help organisations manage their portfolios of innovation projects through a range of tools and
methods designed to generate and evaluate project information and to steer decision-making to maintain a
balanced project portfolio that is aligned with strategic goals [8, 4]. There has been a surge of interest in
PPM in recent years as innovation has become understood as the main driver of economic growth in
developed nations and as organisations have become increasingly project-based [9, 10, 11].

PPM is a rapidly developing field for innovation research and practice, and awareness and application of
PPM practices is growing. Levine [8:22] offers the following concise and generic definition of PPM:
“Project portfolio management is the management of the project portfolio so as to maximise the contribution
of projects to the overall welfare and success of the enterprise”. While project management methods focus on
‘doing projects right’, PPM is a ‘dynamic decision process’ that focuses on ‘doing the right projects’ to
ensure that organisational resources are allocated to the best combination of projects to meet financial,
strategic and other organisational goals [4].

   
The literature indicates that managing a portfolio of innovation projects presents a multi-dimensional
challenge that is often addressed through a PPM capability with a formal and structured process [4, 12, 13].
A growing body of literature on PPM outlines processes, methods and tools and identifies the ‘best practices’
associated with better outcomes [14, 15, 16, 17]. The PPM literature suggests that a variety of methods and
approaches can be applied to the problem. Some PPM processes attempt to apply numerical models similar
to those used for financial portfolio management. However, these models have not been successful due to the
complexity of the project environment, and the necessity to incorporate multiple types of information to
assist optimisation decisions along several dimensions [18]. The many types of information include financial
factors as well as many others such as those related to resources, strategy, risk, interdependence, timing, and
customers. Research indicates that ‘best practice’ organisations make PPM decisions in meetings, and that
graphical methods such as portfolio maps and other graphical and visual information displays facilitate the
ongoing group decision-making for resource allocation, reallocation and reconfiguration [17, 19, 20]. A
PPM capability can enable an organisation to effectively respond to dynamism in the environment through
this ongoing resource reconfiguration.

3. Project interdependencies

Interdependencies within project portfolios are based on multiple interrelated factors such as resource and
time constraints, financial costs, project outcomes and risk profiles. In times of uncertainty and change,
organisations face heightened challenges in managing their project portfolios – prompting initiatives to
improve upon current management techniques such as the research outlined in this paper to improve the
understanding of project interdependencies.

Project interdependencies are acknowledged as an important factor in PPM decision making [21, 22].
Projects are said to be interdependent when the success of a project depends upon other project(s). Eilat,
Golany and Shtub [23] identify three types of interdependencies: resource interactions (the need to share
resources or wait for scarce resources until they are released by another project), benefit interactions
(complementary or competitive effects) and technical dependence. Other types of dependencies that may
exist between projects in a portfolio include outcome dependencies (the need to use the end result of another
project – these can be technical or other outcomes) and learning dependencies (the need to incorporate the
capabilities and knowledge gained through another project). Learning-based project interdependencies are
especially challenging to manage due to the difficulties in codifying project knowledge and transferring that
knowledge between projects in the portfolio.

Organisations manage dependencies between projects using a variety of methods. Management strategies to
improve between-project communication are recommended in the literature [24, 25]. Resource dependencies
are often addressed by scheduling optimisation systems [12]; however these types of systems require large
amounts of numerical input and are not considered useful in most PPM environments. Dependency matrices
are a more common method that are used to provide a view of interdependence between projects [26, 27].
The dependency matrix is similar to the design structure matrix (DSM) that is applied in development
environments to manage interdependent tasks [28, 29, 30]. In a DSM, the rows capture the information
inflows, whilst the columns capture the information outflows between components or sub-projects in a given
system [31], whereas in a dependency matrix, inward and outward between-project dependencies are
captured. The goal of DSM is to analyse the elements within a system and re-sequence these in order to
minimise iterations and their scope [28]. There are two types of DSMs, the static DSM that uses clustering
algorithms and the time-based DSMs that use partitioning algorithms [31]. DSM analysis can be used to
optimise and restructure operations; however this analysis does not reveal accumulated or multi-level
interdependencies.

Methods for managing project interdependencies currently lack the ability to highlight the most important
projects based on cumulative dependencies in the portfolio. Dependency Matrices use a two dimensional grid
to display and analyse bi-directional dependencies between each pair of projects in a portfolio. The rows and
columns are each labelled with the set of projects in the portfolio as illustrated in Figure 1. There are a
number of types of dependencies that may exist between projects such as resource dependencies, time-based
dependencies and outcome based dependencies in addition to information or learning-based dependencies;
the dependency matrix can be used to capture a range of dependencies.

   
 Figure 1: Dependency Matrix illustrating inward and outward dependencies
between projects A-G (based on [31:2])

However, the dependency matrix in Figure 1 does not clearly reveal the relationship between projects F and
A in Figure 1. The matrix shows that project F is dependent upon project C, and on a different line shows
that project C is dependent upon project A. The resulting dependency between project F and project A is not
clearly indicated on the dependency matrix in Figure 1. Understanding these multi level relationships is
important to fully understand project interdependence in a portfolio. Therefore, to improve the management
of project portfolio investments and enhance outcomes, there is a need for a better method to aid the
understanding of multi-level project interdependencies. Network mapping and analysis tools have the
potential to enhance this understanding. Dependency matrix and DSM concepts and methods are related to
network precedence diagrams and network analysis diagrams [28]. The application of network mapping tools
or the combination of tools may provide a powerful method for analysing the interdependency relationships
within project portfolios.

4. Network mapping and analysis

Network analysis and mapping tools build on mathematical modelling and graphical methods to help
understand and analyse networks. Network analysis has its origins in graph theory [32, 33], which is a
mathematical discipline that examines the relationships between a set of entities referred to as vertices or
nodes [34]. Nodes are connected by edges, more commonly known as links or ties. A graph consists of a
set of nodes and the ties between these pairs of nodes. Outside of strictly mathematical applications the
term network is a little more relaxed. Brandes & Erlebach [32:7], for example, say the term network refers
to “the informal concept describing an object composed of elements and interactions or connections between
these elements”. Caldarelli & Vespignan [35:5] link the notions of a system and graph to describe a network
as: “a graph whose nodes (vertices) identify the elements of the system. The set of connecting links (edges)
represents the presence of a relationship or interaction among these elements”.

Network mapping tools have the ability to map relationships between nodes in a network at multiple levels
and may reveal accumulated effects that are not evident in single-level relationship analysis [36]. The
mapping is done through the use of software-based tools that help to record, analyse and visually display the
relationships between items or ‘nodes’ in a network. Computerised tools are used for network analysis due to
their ability to handle large amounts of data and to analyse and display the data in a graphical form.
Enhanced analysis of the data is provided by using such maps to illustrate proposed or actual changes in the
network. The graphical displays provide an intuitive and easy-to-interpret format that can help reveal
patterns that are not revealed by verbal explanations or matrix displays of data [37].

Social network analysis (SNA) and the related organisational network analysis methods are a common
application of network mapping where relationships between people or organisations are analysed and
presented in a visual form [38]. The network mapping exercise involves collecting data from people
representing each ‘node’ of the network on their interaction and relationships with other ‘nodes’. For SNA
the ‘nodes’ are individual people who answer questions about their interactions with other people. SNA is

   
shown to be an aid to understanding and improving relationships between networks of people or
organisations [39, 36, 40]. For example, SNA has been shown to be effective in promoting collaboration,
supporting critical nodes in the network, and in managing and maintaining networks during organisational
restructuring [40].

There are many other existing applications for network mapping. These include mathematical, biological and
economic modelling [37]. Network mapping has also been used in conjunction with design structure matrix
(DSM) tools to manage interdependencies between tasks in product development environments [41, 31]. In
these environments network mapping and analysis is shown to help determine the key relationships in a
complex development environment and to show the impact of making changes to tasks [42]. Product
development team interactions and information flows were measured in another study that demonstrated the
use of network mapping and analysis in identifying the most important nodes in the network [41]. These
studies show the benefits of network mapping for networks of teams or tasks involved in complex product
development.

These findings indicate that network mapping and analysis may have benefits for the understanding of
project interdependencies. There are currently no published studies on the use of network mapping and
analysis to study networks of projects in a project portfolio. However, initial tests have been performed as
part of a larger PhD study on Visualising Collective Knowledge to Manage Complexity in a Portfolio of
Projects [43]. These tests have indicated that network mapping and analsysis can be useful for project,
program and portfolio management by considering each project as a ‘node’ in the network and capturing and
displaying information on the relationships or interdependencies between nodes. One benefit of network
mapping and analysis illustrated by Durant-Law’s work is shown in Figure 2. This figure presents two
network mapping visualisations of the same portfolio of projects. The projects in the view on the left are
indicated by circles sized to indicate the level of investment in the project. The level of investment is often
assumed to be a proxy measure of the relative importance of the project within the portfolio. The view on the
right displays the project size to indicate the number of projects that are dependent upon a project. This view
provides a very different picture of project importance based on the analysis of interdependencies within the
portfolio. Identical projects are circled in each view to highlight the difference in the types of information in
each choice of display.

Projects sized by investment Projects sized by dependency

Figure 2: Network maps of a project portfolio illustrating interdependencies [43]

Network maps are also useful for helping organisations visualise multi-level dependencies. For example a
network map reveals whether a project is indirectly dependent upon another project. For example, Project
A’s dependency list may include Project B but not Project C. If Project B’s dependency list includes Project
C, this indicates that Project A also depends upon Project C, however matrix based methods of analysing

   
project dependencies are difficult to analyse and ‘see’ these dependencies. Figure 3 presents a portion of a
network map showing this type of relationship and illustrating the improved visualisation of multi-level
dependencies afforded by network mapping and analysis of projects in a portfolio. This map illustrates all of
the projects that Project X depends upon, making the relationships clearer and helping to ensure that multi-
level dependencies are acknowledged. As shown in Figure 2, although Project X’s dependencies may only
include two other projects (A and Y), the multi-level dependency chain shows that Project X’s success
depends upon four additional projects (B, C, D and Z). Using this approach it is possible to see and analyse
second-order and beyond relationships.

Project Z

Project C
Project A
Project Y

Project B Project X

Project D

Figure 3: Portion of a network map illustrating multi-level dependencies for project X

5. Research method

The two studies outlined in this paper address the identified need for improved methods for understanding
interdependencies between projects in a portfolio. In the main study, network analysis and mapping tools are
being used to create visual network representations of the project portfolios at three case organisations. A
second study will extend the analysis at one of the organisations by comparing network analysis and
mapping with other methods for understanding project portfolio interdependencies. These studies are
currently being completed and final results will be available early 2010.

The main study employs a three-phase method to evaluate the project environment and a selected project
portfolio at three case organisations. This study aims to measure the relationship between the use of network
mapping tools and an organisation’s understanding of its project portfolio interdependencies. The
organisations selected for the study have complex portfolios with multiple interdependencies. This multiple-
case study method develops a rich picture of the nature of the dependencies between projects in the portfolio,
the effects on these interdependencies on the project environment and portfolio outcomes, and the
relationship between the level of understanding of the project interdependencies and the portfolio
performance at each organisation. Cross case comparisons and analysis will be conducted to evaluate and
test the findings across project environments. Each case study will focus on a selected portfolio with between
about 30 and 70 projects – these limits have been selected to ensure that enough data is available for analysis
– but to keep the scope of the research within manageable bounds.

During the first phase of the research at each organisation, the organisational environment and the nature of
the interdependencies between the projects in the portfolio will be investigated through semi-structured
interviews and analysis of other organisational documents when possible. The bounds of the portfolio for the
study will be identified during this phase. The second phase, the data collection phase, first incorporates the
findings from the first phase to customise the data collection instrument to include the projects in the relevant
portfolio and to capture the nature of dependencies based on the types of dependencies experienced by that
organisation. Project managers responsible for the identified projects each complete a survey using the
ONAsurveys survey tool 1 . The ONAsurveys tool is designed for the analysis of organisational networks and

1
ONAsurveys is a tool for capturing network data for display in network maps - www.ONAsurveys.com.

   
has been customised for this research including the addition of a range of additional questions to better
understand the project environment. During this second phase, a group of senior portfolio stakeholders will
complete a separate survey to assess their current impressions of the relative importance and critical
dependencies between projects in the portfolio and to capture their impressions on the performance of the
overall project portfolio. At the completion of the second phase, the data is analysed using network mapping
and analysis tools [44], and these findings compared with the senior project stakeholders’ impressions. The
final phase of the study involves presenting the findings to the case organisations’ senior project stakeholders
and collecting feedback and impressions through a semi-structured feedback session.

The second study extends the analysis conducted in the main study at one of the case organisations. The
second study will compare and contrast several methods of understanding project portfolio dependencies by
using multiple tools such as dependency matrices or design structure matrices to analyse the same portfolio.
Each tool’s ease of use, information gained, and applicability to a range of project portfolio management
problems such as scheduling, prioritisation and change management will be explored. This second study will
also investigate whether a combination of methods will best enhance understanding of a project portfolio.

6. Conclusion

In order to best manage their project portfolios, organisations need to understand the interdependencies
between projects. Project and portfolio complexity and dynamic environments create challenges for the
understanding of these interdependencies, and there is a need for better methods and tools to aid in this
understanding. The research introduced in this paper investigates network mapping and other graphical
methods for capturing, displaying and updating information on dependencies between projects in a portfolio.
Graphical methods and visual information displays have been shown to facilitate group PPM decision-
making indicating that the graphical and visual methods investigated in this study may also strengthen PPM
processes.

Network analysis and mapping tools use visual network mapping displays while the dependency matrix
presents data on a grid to highlight project interdependencies. Network mapping and analysis is a novel
approach for the study of relationships between projects in a portfolio. These network analysis tools have
become popular for the study of relationships in social networks and initial results indicate benefits for the
study of project relationships.

Building upon initial findings that illustrate the benefits of network mapping as a visualisation tool in a
project portfolio environment, the research projects outlined in this paper will test the applicability of
network analysis and mapping tools in a range of environments. The research will investigate relationships
between the use of network mapping, understanding of project interdependencies, and project portfolio
management outcomes, and will also compare and combine network analysis and mapping tools with
dependency matrix and DSM tools for the capture, analysis and communication of project interdependencies.

Network mapping and analysis tools and dependency matrix tools have been chosen for the study because of
their potential to enhance the understanding of project interdependencies to ensure that decisions are better
informed. These informed decisions are expected to improve the outcomes from investments in an
organisation’s project portfolio. The project will contribute to the engineering and project management
community by investigating new methods of understanding project interdependencies and using the findings
to improve organisational PPM capabilities – thus enabling those organisations to more dynamically respond
to environmental change.

7. References

[1] Webb A., Managing innovative projects, Chapman & Hall, London ; New York, 1994.
[2] Wideman R. M., A management framework for project, program and portfolio management,
Trafford Publishing, Victoria B.C., 2004.
[3] Griffin A., "PDMA research on new product development practices: Updating trends and
benchmarking best practices", Journal of Product Innovation Management, Vol. 14, Iss. 6, 1997, pp.
429-458.

   
[4] Cooper R. G., Edgett, S. J. and Kleinschmidt, E. J., Portfolio management for new products, Perseus,
Cambridge, Mass., 2001.
[5] Tidd J., Bessant, J. and Pavitt, K., Managing innovation: Integrating technological, market and
organizational change, John Wiley and Sons, Chichester, 2005.
[6] Jenner S., Realising benefits from government ict investment - a fool's errand?, Academic Publishing
International, Reading, UK, 2009.
[7] Killen C. P., Hunt, R. A. and Kleinschmidt, E. J., "Learning investments and organisational
capabilities: Case studies on the development of project portfolio management capabilities",
International Journal of Managing Projects in Business, Vol. 1, Iss. 3, 2008, pp. 334-351.
[8] Levine H. A., Project portfolio management : A practical guide to selecting projects, managing
portfolios, and maximizing benefits, Jossey-Bass ; John Wiley distributor, San Francisco, Calif.
Chichester, 2005.
[9] Cleland D. I., "The strategic context of projects," Project portfolio management: Selecting and
prioritizing projects for competitive advantage, L. D. Dye and J. S. Pennypacker (Editors), Center
for Business Practices, Havertown PA, 1999.
[10] Edwards M. and Croker, M., "Major trends and issues, innovation and productivity in services",
Industry, Services & Trade, OECD publications service, Vol. 33, 2001, pp. 7-16.
[11] Galende J., "Analysis of technological innovation from business economics and management ",
Technovation, Vol. 26, Iss. 3, 2006, pp. 300-311.
[12] Archer N. P. and Ghasemzadeh, F., "An integrated framework for project portfolio selection",
International Journal of Project Management, Vol. 17, Iss. 4, 1999, pp. 207-216.
[13] Cauchick Miguel P. A., "Portfolio management and new product development implementation: A
case study in a manufacturing firm", International Journal of Quality & Reliability Management,
Vol. 25, Iss. 1, 2008, pp. 10-23.
[14] Loch C., "Tailoring product development to strategy: Case of a european technology manufacturer",
European Management Journal, Vol. 18, Iss. 3, 2000, pp. 246-258.
[15] Cormican K. and O'Sullivan, D., "Auditing best practice for effective product innovation
management", Technovation, Vol. 24, Iss. 10, 2004, pp. 819-829.
[16] Jeffery M. and Leliveld, I., "Best practices in it portfolio management", MIT Sloan Management
Review, Vol. 45, Iss. 3, 2004, pp. 41-49.
[17] Killen C. P., Hunt, R. A. and Kleinschmidt, E. J., "Project portfolio management for product
innovation", International Journal of Quality and Reliability Management, Vol. 25, Iss. 1, 2008, pp.
24-38.
[18] Coldrick S., Longhurst, P., Ivey, P. and Hannis, J., "An R&D options selection model for investment
decisions", Technovation, Vol. 25, Iss. 3, 2005, pp. 185-193.
[19] Mikkola J. H., "Portfolio management of R&D projects: Implications for innovation management",
Technovation, Vol. 21, Iss. 7, 2001, pp. 423-435.
[20] De Maio A., Verganti, R. and Corso, M., "A multi-project management framework for new product
development", European Journal of Operational Research, Vol. 78, Iss. 2, 1994, pp. 178-191.
[21] Söderlund J., "On the broadening scope of the research on projects: A review and a model for
analysis", International Journal of Project Management, Vol. 22, Iss. 8, 2004, pp. 655-667.
[22] Stummer C. and Heidenberger, K., "Interactive R&D portfolio analysis with project
interdependencies and time profiles of multiple objectives", Engineering Management, IEEE
Transactions on Vol. 50, Iss. 2, 2003, pp. 175-183.
[23] Eilat H., Golany, B. and Shtub, A., "Constructing and evaluating balanced portfolios of R&D
projects with interactions: A dea based methodology", European Journal of Operational Research,
Vol. 172, Iss. 2006, 2006, pp. 1018-1039.
[24] Platje A., Seidel, H. and Wadman, S., "Project and portfolio planning cycle : Project-based
management for the multiproject challenge", International Journal of Project Management, Vol. 12,
Iss. 2, 1994, pp. 100-106.

   
[25] Nobeoka K. and Cusumano, M. A., "Multiproject strategy, design transfer, and project performance:
A survey of automobile development projects in the us and japan", IEEE Transactions on
Engineering Management Vol. 42, Iss. 4, 1995, pp. 397-409.
[26] Dickinson M. W., Thornton, A. C. and Graves, S., "Technology portfolio management: Optimizing
interdependent projects over multiple time periods", IEEE Transactions on Engineering
Management, Vol. 48, Iss. 4, 2001, pp. 518-527.
[27] Slade M., "Managing it project and programme interdependencies across the home office,"
Presentation to the PPM Standards Group London, 2009.
[28] Browning T. R., "Use of dependency structure matrices for product development cycle time
reduction," Fifth ISPE International Conference on Concurrent Engineering: Research and
Applications, Tokyo, Japan, 1998.
[29] Makumbe P. O., "Systems development technical interactions and innovation: A networks-based
investigation," Systems Engineering Division, vol. Master of Science in Engineering Systems,
Massachusetts Institute of Technology, Boston, MA, 2002.
[30] Yassine A. A., "An introduction to modeling and analyzing complex product development processes
using the design structure matrix (dsm) method," Product Development Research Laboratory,
University of Illinois at Urbana-Champaign, Urbana-Champaign, 2004.
[31] Batallas D. A. and Yassine, A. A., "Information leaders in product development organizational
networks: Social network analysis of the design structure matrix," Understanding complex systems
symposium, University of Illinois at Urbana-Champaign, 2004.
[32] Brandes U. and Erlebach, T. (Editors), Network anlysis: Methodological foundations, Springer,
Berlin, 2005.
[33] Newman M., Barabasi, A. and Watts, D. (Editors), The structure and dynamics of networks,
Princeton University Press, Princeton, 2006.
[34] Aldous J. and Wilson, R., Graphs and applications: An introductory approach, Springer, London,
2007.
[35] Caldarelli G. and Vespignani (Editors), Large scale structure and dynamics of complex networks:
From information technology to finance and natural science, World Scientific Publishing,
Singapore, 2007.
[36] Scott J., Social network analysis: A handbook, Sage publications, Thousand Oaks, CA, 2008.
[37] Hanneman R. A. and Riddle, M., Introduction to social network methods, University of California,
Riverside (published in digital form at http://faculty.ucr.edu/~hanneman/), Riverside, CA 2005.
[38] Anklam P., Cross, R. and Gulas, V., "Expanding the field of vision", The Learning Organization,
Vol. 12, Iss. 6, 2005, pp. 539-251.
[39] Wasserman S. and Faust, K., Social network analysis: Methods and applications, Cambridge
University Press, Cambridge, 1994.
[40] Cross R., Borgatti, S. P. and Parker, A., "Making invisible work visible: Using social network
analysis to support strategic collaboration", California Management Review, Vol. 44, Iss. 2, 2002,
pp. 25-46.
[41] Bradley J. A. and Yassine, A. A., "On the use of network analysis in product development teams,"
ASME 2006 International Design Engineering Technical Conference (DETC) and 18th International
Conference on Design Theory and Methodology (DTM), Philadelphia, PA, USA, 2006.
[42] Collins S., Yassine, A. A. and Borgatti, S. P., "Evaluating product development systems using
network analysis", Systems Engineering, Vol. 12, Iss. 1, 2009, pp. 55-68.
[43] Durant-Law G. A., "Visualising collective knowledge to manage complexity in a portfolio of
projects (working title)," Faculty of Business and Government, Doctor of Philosophy, The University
of Canberra, Canberra, forthcoming.
[44] Borgatti S., Analytic Technologies, Harvard, Massachusetts, NetDraw:Graph visualization software,
2002.