The Usage of System Dynamics in Organizational Interventions

Birgitte Snabe The Usage of System Dynamics in Organizational Interventions

WIRTSCHAFTSWISSENSCHAFT

Birgitte Snabe

The Usage of System Dynamics in Organizational Interventions A Participative Modeling Approach Supporting Change Management Efforts

With a foreword by Prof. Dr. Peter Milling

Deutscher Universitäts-Verlag

Bibliografische Information Der Deutschen Nationalbibliothek Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über abrufbar.

Dissertation Universität Mannheim, 2006

D 17

flage Dezember 1997 1. Auflage März 2007 Alle Rechte vorbehalten © Deutscher Universitäts-Verlag | GWV Fachverlage GmbH, Wiesbaden 2007 Lektorat: Brigitte Siegel / Anita Wilke Der Deutsche Universitäts-Verlag ist ein Unternehmen von Springer Science+Business Media. www.duv.de Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung des Verlags unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Die Wiedergabe von Gebrauchsnamen, Handelsnamen, Warenbezeichnungen usw. in diesem Werk berechtigt auch ohne besondere Kennzeichnung nicht zu der Annahme, dass solche Namen im Sinne der Warenzeichen- und Markenschutz-Gesetzgebung als frei zu betrachten wären und daher von jedermann benutzt werden dürften. Umschlaggestaltung: Regine Zimmer, Dipl.-Designerin, Frankfurt/Main Gedruckt auf säurefreiem und chlorfrei gebleichtem Papier Printed in Germany ISBN 978-3-8350-0711-6

Kathrine and Nicolaj

Foreword Internationalization and globalization are major forces for companies to change their organizational structures and processes fundamentally. To master the associated problems, profound and well planned procedures are indispensable, a task which is referred to as change management, and which has to take into account both structural and dynamic aspects. Especially interventions in the area of human resource management lead to manifold repercussions—intended and unintended, enhancing or interfering with the original intentions. Birgitte Snabe investigates in her dissertation if and to what extent System Dynamics can be helpful to design organizational interventions and to examine and evaluate in a next step which particular actions offer adequate problem solutions. In a distinction from the permanent organizational adaptation to a changing environment, the author understands organizational interventions as discrete and fundamental changes to the company’s structures and processes. The management of organizational interventions consists of two interdependent cycles: Problem formulation, analysis and solution (what to do) on the one hand, and the resulting actions to make the change happen (how to do it) on the other. It is a central hypothesis of the investigation that the implementation of solutions to strategic problems often presents larger challenges than the development of the solution itself. The methodological support for effective implementation processes is the core topic of the dissertation. Following an ‘action research’ approach, a delicate and far reaching personnel decision in a large corporation was investigated and is discussed. System Dynamics uses participative model building since about 1990 and offers the prerequisites for mapping contexts which are difficult to quantify. The author presents the development of a complex system model and the implementation of its recommendations both in the practical steps of the concrete case under investigation and in the abstract form necessary for scientific analysis. She demonstrates the power of the selected modelling approach and points out how group dynamics lead to the integration of

VIII

Foreword

initially individual objectives and strategies into a generally accepted process model – i.e. into a “shared mental model”. Mrs. Snabe’s work skillfully combines theoretical considerations and aspects of the practical implementation. The propositions about organizational interventions developed in the conceptual parts of the study are tested in a rigorous – even though not extensive and representative – practical setting. Their viability is shown in the context of top-management decision making.

Peter Milling

Preface The managers below the top executives in large corporate organizations are often placed in the challenging situation of implementing other people’s ideas. Top executives will frequently launch strategic initiatives, and expect the managers at lower levels to act as change leaders even though they have often played only a small or no part at all in the groping strategy forming process where the strategic initiative has its origin. Consequently, there is a need for learning processes that focus on the transfer of the insights and reasoning behind the decision, as well as supporting the refining of implementation plans. Furthermore, the processes should allow iterations with top executives, with the dual objective of adjusting the strategic initiative according to implementation issues and giving the managers responsible for implementation true influence on the entire process. To a great extent, this dissertation addresses the process of transfer of insights and ownership as well as the operationalization of strategic initiatives and other change projects. The main topic is the usage of system dynamics modeling in organizational interventions in general, and specifically the use of system dynamics modeling for the purposes of change management. The first two chapters mainly discuss organizational interventions and the use of modeling in decision-making and policy forming processes, which is the predominant application of system dynamics. The last three chapters concentrate on a rather specific application of system dynamics: modeling in a change management context. Change management dedicated application of system dynamics builds upon the theories and methods of system dynamics in a decision-making and policy-forming context, but aims at the transfer of insights and ownership from decision-makers to implementers, as well as refining and aligning cross-organizational implementation plans. Writing this doctoral dissertation has been an interesting journey for me. I have enjoyed the opportunity to take the time to go into depth with the literature especially from the disciplines of system dynamics and organizational psychology. Coming directly from 10 years of management consulting, the academic experience has been of great personal and

X

Preface

educational value to me. I would therefore like to sincerely thank first of all Professor Dr. Peter Milling for supporting me on my journey. He and assistant Professor Dr. Andreas Größler have guided me with great patience through the learning process – helping me to adjust my normative and solution-oriented worldview from the consulting world to also embrace scientific and academic viewpoints. I would also like to thank all my doctoral colleagues at ‘Industrieseminars’ for the weekly discussions at the doctoral seminars, and Markus Salge and Dr. Nadine Schieritz especially for always volunteering to finding literature and discussing modeling issues. Furthermore, I want to thank my good friend since early childhood, Kirstine Munk, who has been struggling with her own dissertation at the same time as me. Although our subjects are very different, we have had many and interesting discussions on a wide variety of issues including theories of science, social constructivism, aesthetic in consultations and workshops, and using cognitive frameworks (being both models and horoscopes!) to reduce personal barriers for involvement and honesty in discussions. Last but not least, I want to thank my husband, Jim Hagemann Snabe. As well as receiving personal support in many ways, I have also been very privileged to be able to draw on his extensive business experience and conceptual capabilities.

Birgitte Snabe

List of Contents

Foreword .............................................................................................. VII Preface .................................................................................................. IX List of Figures ...................................................................................... XV List of Tables..................................................................................... XVII

A. Organizational Intervention Skills as Corporate Competence........ 1 I. The Need for and the Challenges of Organizational Interventions ......................................................... 1 II. Foundation and Strategies for Planned Change Interventions....... 10 III. The Usage of System Dynamics Modeling in Organizational Interventions ................................................... 15

B. Conceptual Foundation for the Usage of System Dynamics ......... 27 I. The Usage and Utility of Modeling in Decision-Making .............. 27 II. Cognitive and Behavioral Rationale for the Usage of System Dynamics ......................................................... 35 1. Individual Learning and Change of Behavior in a Complex and Dynamic Environment ...................................... 35 2. Establishing Group Consensus by Sharing Mental Models...... 46 3. Enhancing Organizational Learning through System Thinking Experience and Double-Loop Learning ................... 51

XII

List of Contents

III. The Development Process of System Dynamics Models in Corporations ........................................................................... 56 1. Decompositions and Iterations in Model Development ........... 58 a. Problem Definition and System Conceptualization............. 58 b. Model Formulation and Testing......................................... 62 c. Policy Formulation and Implementation ............................ 66 2. Designing System Dynamics Modeling-Based Interventions... 68 a. Experimentation-Based Learning Cycles ........................... 68 b. Knowledge Acquisition in Modeling Projects .................... 71 c. Designing Participative Modeling Interventions ................ 73 d. Facilitation of Participative Modeling Interventions .......... 77

C. A Case Study Using Participative System Dynamics Modeling in the Implementation of a Sensitive Change Project .................... 83 I. Research Considerations for the Case Study Application ............. 83 II. Case Study: Refining and Implementing a Location Strategy ....... 89 1. The Problem and its Context .................................................. 89 2. Intervention Process............................................................... 93 3. The Model and Selected Simulations...................................... 98 III. Evaluation of the Case Study .....................................................108 1. A Framework for Evaluating the Effectiveness and Efficiency of the Case Study .................................................108 2. Conclusions on Case Study Effectiveness and Efficiency.......116

List of Contents

XIII

D. The Usage and Utility of Participative Modeling in Change Management .................................................................119 I. Context Factors Relevant for Deciding on Usage of Modeling in Change Management ...............................120 II. Process Considerations ..............................................................128 1. Business Objectives and Targets Directing and Framing the Intervention .......................................................131 2. Structured Development of Change Leaders ..........................135 3. Designing the Change Process...............................................140 4. Facilitation of modeling and simulation sessions ...................148 III. Outcomes of Participative Modeling Efforts in the Implementation of Change Programs..........................................155 1. Modeling and simulation as a tool for transfering insights and ownership from decision-makers to implementers ...........155 2. Refining and Aligning Implementation Plans Through Scenario Simulation ................................................157 3. Organizational Learning in Change Management Oriented Modeling ................................................................158

E. Targeted Participative Modeling in Change Management...........161

Appendices ...........................................................................................165 Bibliography.........................................................................................201

List of Figures Figure A-1:

The Basic Model of Corporations ...................................... 2

Figure A-2:

Model of stages of problem solving ................................... 5

Figure A-3a:

Diagnostics and decision-making (cycle I) ......................... 6

Figure A-3b: Change management (cycle II) ........................................... 6 Figure A-4:

Goal-seeking system ........................................................ 16

Figure B-1:

Accumulative levels of models in the usage of system dynamics ......................................................................... 30

Figure B-2:

Limited linear perception of system ................................. 36

Figure B-3:

Theory of planned behavior ............................................. 43

Figure B-4:

Different limited linear perceptions of a system ............... 47

Figure B-5:

Mental models as instruments between actual systems and formal models ........................................................... 51

Figure B-6:

The basic structure of organizational learning .................. 54

Figure B-7:

Formal models supporting organizational learning ........... 56

Figure B-8:

The learning cycle for learning labs ................................. 70

Figure B-9:

Mental database and decreasing content of written and numerical databases................................... 71

Figure B-10:

Maps, frameworks and microworlds................................. 74

Figure B-11:

A model of communication .............................................. 79

Figure B-12:

The Wallow Curve at work .............................................. 81

XVI

List of Figures

Figure C-1:

The reinforcing growth loop underlying the intervention ..................................................................... 91

Figure C-2:

Intervention process as communicated in the project ........ 94

Figure C-3:

The location strategy model ............................................. 99

Figure C-4:

Fraction of employees in low-cost countries compared to total number of employees in the division ................. 104

Figure C-5:

Development in cost-index for an average productive unit (e.g. cost for one employee for a fixed period) ....... 105

Figure C-6:

Development of productivity index for an average unit (e.g. output/month/unit) ................................................. 105

Figure D-1:

Conceptual modeling steps in the case study .................. 128

Figure D-2:

Four antecedent processes in organizational interventions.................................................................. 144

Appendix D, Figure 1: The preliminary model in the case study....……. 185 Appendix E, Figure 1: Simulation run of adjusted model (avoiding rate-on-rate modeling)………….……..187 Appendix E, Figure 2: Model without rate-on-rate modeling……...…… 188

List of Tables Table C-1:

Roles and responsibilities as defined in the project........…. 95

Table C-2a: Main sources for evaluation of outcomes on Individual level…………………………………….………... 111 Table C-2b: Main sources for evaluation of outcomes on group level…… 112 Table C-2c: Main sources for evaluation of outcomes on organization level………………………………………… 113 Table C-2d: Main sources for evaluation of system dynamics compared to other approaches……...……………………….. 114 Table C-2e: Main sources for evaluation of the usage of system dynamics in a change management context……………………………. 115 Table C-3:

Questionnaire results……..………………………….……. 118

Table D-1:

Political characteristics of situations in terms of the issues of interest, conflict, and power………………..…... 123

Table D-2:

Generic symptoms of change resistance..………………… 142

Appendix A, Table 1: Parameters relevant to high-cost locations……… 167 Appendix A, Table 2: Parameters relevant to low-cost locations………. 168 Appendix A, Table 3: Parameters mainly relevant for transfer of tasks and build-up of employees in low-cost locations………… 169 .

Appendix B, Table 1: Main equations influencing stock levels………… 171 Appendix B, Table 2: Main equations influencing production/month….. 172

A. Organizational Intervention Skills as Corporate Competence I.

The Need for and the Challenges of Organizational Interventions

Finding ways to lead and develop organizations is a constant quest seeking to ensure competitiveness in a changing and dynamic world, which is well illustrated by Forrester’s words calling change “the essence of the manager’s environment.”1 Furthermore, industries are typically facing shorter changes cycles in new technologies, competition, value chain, environmental factors, and customer demands, resulting in an increased need for effectiveness and efficiency in organizational change processes.2 The changes constitute challenges representing both threats and new opportunities for the individual business organization, putting pressure on its ability to learn and transform.3 Organizations change in various ways: in organic, incremental processes of adapting to changing environments or in more abrupt organizational interventions.4 The latter way is the focus of this dissertation.

1 2

3

4

Forrester, Jay W.: Industrial Dynamics, Cambridge, 1961, p. 1. See Kotter, John P: Leading Change, Boston, 1996, p. 18; Fine, Charles H.: “Clockspeed-based strategies for Supply Chain Design”, Production and Operation Management, Vol. 9, No. 3, 2000, p. 213; Brown, John Seely: “Research That Reinvents the Corporation”, Harvard Business Review, August 2002, p. 105. In de Geus, Arie P.: “Planning as Learning”, Harvard Business Review, Vol. 66, No. 2, March-April 1988, p. 71, it is proposed that ”the ability to learn faster than competitors may be the only sustainable competitive advantage.“ The Japanese concept of Kaizen is an example of a continued process improvement focus, whereas the western world typically is more oriented towards innovation- and result-oriented thinking, see Imai, Masaaki: Kaizen: Der Schlüssel zum Erfolg der Japaner im Wettbewerb, German translation, München, 1992, p. 15.

2

A. Organizational Intervention Skills as Corporate Competence

Discussions of the concepts of ‘organizations’ and ‘interventions’ cover a wide range of literature from the sciences of management, the social sciences and system sciences. Taking a departure from the field of cybernetics, Takahara offers a rather generic definition of an organization as “a complex system of interconnected human and nonliving machines; and it is formed for a purpose, to achieve a certain goal.”5 In line with this definition, Milling has described the basic model of corporations as a goal-seeking input-output feedback system:

Goals

Managing Stratum

Interventions

Resource input

Feedback

Causal Stratum

Achieved output

Figure A-1: The Basic Model of Corporations 6

In this basic model of corporations, two different, but interrelated, conceptual processes form the basic structure: the causal stratum, which is the operative

5

6

Takahara, Yasuhiko: “A Formal Model of Organization”, in Takahashi, Singo, Kyoichi Kijima and Ryo Sato (eds.): Applied General Systems Research on Organizations, Tokyo, 2004, p. 3. Own translation of figure in Milling, Peter: Systemtheoretische Grundlagen zur Plannung der Unternehmenspolitik, Berlin, 1981, p. 17. It should be noted that the original figure uses the German term “Führungsstratum” (translated to Managing Stratum) which is a broader term also encompassing the meaning of leading, steering, controlling. This model is chosen due to its abstraction level suitable to illustrate the concept of interventions. Takahara offers a more detailed basic model of organizations decomposing the operational level (the causal stratum), which inherent has a stronger focus on internal structures and coordination challenges; see Takahara, Yasuhiko: A Formal Model of Organization, in Takahashi, Singo, Kyoichi Kijima and Ryo Sato (eds.): Applied General Systems Research on Organizations, Tokyo, 2004, pp. 10—13.

A. Organizational Intervention Skills as Corporate Competence

3

domain producing the output, and the higher-level managing stratum, which includes information processing to goal compliant forming and controlling of the causal stratum.7 In the social-psychological field of science, Argyris describes organizational interventions from a task point of view stating: “the interventionist’s primary tasks are to generate valid information, to help the client system make informed and responsible choices, and to develop internal commitment to these choices”.8 The terms ‘the interventionist’ and ‘the client’ are often used in intervention literature.9 Although disagreement exists with regards to the importance of independence of the system and the intervener, the contemporary literature focusing on organizational development typically sees organizational interventions as embracing both change processes with and without the use of external interventionists.10 The client system can, in the socialpsychological field of science, be an individual, a group of people or an organization. At these three levels, behavioral changes are largely explained with

7

8

9

10

Milling, Peter: Systemtheoretische Grundlagen zur Plannung der Unternehmenspolitik, Berlin, 1981, p. 18. Milling later decomposes the managing stratum into four levels: the normative level (formulation of long-term goals), the structuring level (determination of the basic structures), the adaptive level (specification of change programs) and the operative level (selection of actions), p. 20. Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Massachusetts, 1970, p. 21. Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, p. 142, describe the term intervention as “sequenced planned actions or events intended to help an organization increase its effectiveness. Interventions purposely disrupt the status quo; they are deliberately attempts to change an organization or sub-unit towards a different and more effective state.” Linguistic, the term ‘intervention’ indicates, that a party is proactively doing something to change the system. This is also seen in fields like economy (state interventions) and foreign affairs (armed conflicts). In Argyris, Chris: Interventions Theory and Method – A behavioural Science View, Reading, Massachusetts, 1970, p. 15, the importance of independency between the client and the interventionist is stressed, whereas in Schein, Edgar H.: Process Consultation, Boston, 2000, part II, p. 35, (collection of work first published in the 1960’s), it is argued that both external consultants as well as managers from within the company can serve the role of the interventionist. Recent textbooks generally support the latter view.

4

A. Organizational Intervention Skills as Corporate Competence

the same mechanisms, having the change of attitudes and intentions of individuals as a central element.11 Offering an additional perspective from the social-psychological field of science, Schein identifies three basic models of organizational interventions as being (1) The Purchase of Expertise Model, (2) The Doctor-Patient Model, and (3) The Process Consultation Model.12 In the Purchase of Expertise Model, the role of the interventionist is to provide recommendations based on expert information and services, whereas the Doctor-Patient Model starts with an investigation of ‘symptoms’ followed by analyses and recommendations made by the interventionist. In both The Purchase of Expertise Model and The DoctorPatient Model, the primary objective is the identification of ‘the solution.’ The third intervention model, The Process Consultation Model, focuses rather on strengthening the organization’s own ability to identify the core problem in general, as well as finding a suitable solution in the specific situation. Furthermore, the Process Consultation Model focuses strongly on stakeholder involvement in the search for sustainable change, as was also seen in Argyris’ discussion of establishment of internal commitment. Theories of organizational interventions are closely linked with those of decision-making and problem solving. The processes of decision-making and problem solving from individual, group and organizational perspectives are extensively discussed in the literature.13

11

12

13

See Chin, Robert and Kenneth D. Benne: “General Strategies for Effecting Changes in Human Systems”, in Bennis, Warren G., Kenneth D. Benne and Robert Chin: The Planning of Change, 4th edition, New York, 1985, p. 24; and Bungard, Walter and Catrin Niethammer: “Psychologische Aspekte des Change Management im interorganisationalen Kontext”, in Walter Bungard, Jürgen Fleischer, Holger Nohr, Dieter Spath and Erich Zahn (eds.), Customer Knowledge Management, Stuttgart, 2003, p. 109. Schein, Edgar H.: Process Consultation, Boston, 2000, part I, pp. 9—12, and part II, pp. 29—35. It should be noted, that this book mainly consists of reprints from his work in the late 1960’s. Schein’s work in general addresses the last intervention model type, the Process Consultation Model. In Akkermanns, Henk: Modelling With Managers, Breda, The Netherlands, 1995, pp. 7—12, a literature overview of decision-making and problem solving is found, covering Operation Management, System Dynamics, Strategic Management, Operations Research/“Soft OR”, Group Decision Support Systems and Organizational Psychology.

A. Organizational Intervention Skills as Corporate Competence

5

4 Action planning

1 Problem formulation

6 Evaluating outcomes

3 Felt need Forecasting consequences, testing proposals

2 Producing proposals for solutions

5 Taking action steps

Figure A-2: Model of stages of problem solving14

Figure A-2 illustrates a problem-solving process of iterative stages with two conceptual cycles succeeding the problem identification (the “felt need” in the center of the figure). Cycle I (figure A-3a) includes the problem formulation, producing proposals for solutions and forecasting consequences and testing proposals. Cycle II (figure A-3b) includes action planning, taking action steps and evaluating outcomes.

14

See Schein, Edgar H.: Process Consultation, Boston, 2000, part I, p. 61. The model is an elaboration of a model originally developed by Richard Wallen for use in sensitivity training programs. The model has strong similarities with Dörner’s “Steps in Planning and Action”, although Dörners model less sharp separate in a planning and an implementation part, see Dörner, Dietrich: The Logic of Failure, New York, 1996, p. 43. Also, the model has similarities with the PDCA-cycle (Plan-Do-Check-Act) as seen in TQM. The PDCA-cycle is a generic version of the Deming Cycle, see Imai, Masaaki: Kaizen: Der Schlüssel zum Erfolg der Japaner im Wettbewerb, german translation, München, 1992, p. 87.

6

A. Organizational Intervention Skills as Corporate Competence

4 Action planning 1 Problem formulation Felt 3 2 Forecasting need Producing consequences, proposals 6 testing for solutions 5 Evaluating proposals Taking outcomes action steps

Figure A-3a: Diagnostics and decision-making (cycle I)

4 Action planning

1 Problem formulation Felt 3 2 Forecasting need Producing consequences, proposals 6 testing for solutions 5 Evaluating proposals Taking outcomes action steps

Figure A-3b: Change management (cycle II)

For organizational interventions addressing the strategic problems of larger organizations, the two cycles of problem solving are typically discussed in

A. Organizational Intervention Skills as Corporate Competence

7

two distinct areas: the area of strategy forming and the area of strategy implementation.15 The latter is described in the change management focused literature from the disciplines of organizational psychology and organizational development (OD). Cycle I of interventions have their primary focus on diagnostics and decision-making, which for strategic organizational interventions can be understood as strategy forming. In the field of organizational psychology, cycle I activities are normally labeled ‘organizational diagnostics’ which is leading to the initiation of the planned change process.16 Cycle II interventions, focusing on planning, carrying out and following up on implementation, comprise what in the fields of organizational psychology and OD are typically categorized as ‘planned change’ interventions. The iterative and recursive nature of the entire problem solving process (both cycle I and cycle II) should not be underestimated, as also discussed in the problem solving system described by Flood, “Total System Intervention”.17 Total System Intervention focuses on creative problem investigation and deliberate selection of methods to solve problems, through an iterative and recursive process of three phases, (1) creativity, (2) choice, (3) implementation. In this context, strategy forming is influenced by implementation considerations and experiences, and the strategy implementation constitutes in itself a new cycle with the need for creative ideas on how best to implement the strategy and the choice of the best methods to achieve the implementation. Research within the area of strategy forming proposes that the way corporations address strategic problems should be considered as groping, interactive processes emphasizing learning, creativity, synthesis, and sharing of mental models among decision-makers.18 In this regard, the understanding of 15

16

17 18

Huff, Anne S. and Rhonda Kay Reger: “A Review of Strategic Process Research”, Journal of Management, Vol. 13, No. 2, 1987, p. 212. It should be noted, that Huff and Reger use the term strategy formulation rather than strategy forming. The term ‘diagnostics’ is often used as heading for activities leading to the planned change interventions, e.g. see the list of contents in Bennis, Warren G., Kenneth D. Benne and Robert Chin: The Planning of Change, 4 th edition, New York, 1985 as well as in Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001. Flood, Robert L.: Solving Problem Solving, Chichester, 1995, p. 32. See Mintzberg, Henry: The Rise and Fall of Strategic Planning, New York, 1994, p. 77; de Geus, Arie P.: “Planning as Learning”, Harvard Business

8

A. Organizational Intervention Skills as Corporate Competence

mental models offers some interesting perspectives on the challenges in the process. The interpretation of mental models varies significantly in the literature from understanding mental models as “pre-compiled” limited conceptual representations to seeing them as implicit, foggy, intuitive system perceptions also involving the subconscious.19 Despite difficulties in the literature to agree on a definition on mental models, it seems that general agreement exists in the understanding, that mental models influence behavior and decisions significantly, and that in information selection and interpretation, people subconsciously seek confirmation of their existing mental models, which can also result in them rejecting or ignoring information that contradicts their beliefs.20 Due to the complexity of social systems, the strategy forming processes must encourage strategies and assumptions to be challenged from inside or outside the problemsolving environment to challenge improper beliefs people may have about causal relations in their mental models.21

19

20

21

Review, March-April 1988, p. 71; Davenport, Thomas H.: Process Innovation, Boston, 1993, pp. 278—279. In Eisenhardt, Kathleen M.: “Strategy as Strategic Decision Making”, Sloan Management Review, Spring 1999, p. 66, this process is described as “building collective intuition.” For a discussion on mental models and literature on mental models, see among others Senge, Peter M.: The Fifth Discipline, New York, 1994, pp. 174—204; Doyle, James K. and David N. Ford: “Mental models concepts for system dynamics research”, System Dynamics Review, Vol. 14, No. 1, Spring 1998, pp. 3—29, Doyle, James K. and David N. Ford: “Mental models concepts revisited: some clarifications and a reply to Lane”, System Dynamics Review, Vol. 14, No. 1, Spring 1998, pp. 3—12. The word ‘subconscious’ is used in this dissertation as “existing or operating in the mind beneath or beyond consciousness”, see Webster’s Encyclopedic Unabridged Dictionary of the English Language, New York, 1989, p. 1414. See Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 21; Senge, Peter M.: The Fifth Discipline, New York, 1994, p. 175; Kampmann, Christian P. E.: Feedback complexity and market adjustment – An experimental approach, Boston, 1992, p. 29; Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993, p. 30; Hogarth, Robin: Judgment and Choice – The Psychology of Decision, 2 nd edition, Chicago, 1987, p. 130. See Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993, p. 30; Argyris, Chris: Reasoning, Learning, and Action – Individual and Organizational, San Francisco, 1982, p. 39.

A. Organizational Intervention Skills as Corporate Competence

9

The challenge of improper beliefs and the improvement and realignment of mental models is not only relevant among the executive decision-makers, but is often critical in larger circles. In many organizations, decentralization and empowerment have resulted in business decisions being made also at the lower levels of the organizational hierarchy. Therefore, new strategies can seldom be implemented only by introducing new guidelines or policies, as it is required that a larger number of employees understand why the organization must change as well as understand the reasoning behind the new strategy. Consequently, the transfer of insights gained by the decision-makers in the strategy forming process is of great importance. In terms of the previously introduced problem solving cycles, this means that the insights gained in Cycle I need to be transferred to the people responsible for the implementation (cycle II). A further argument for this transfer of insights to take place in the intervention process is the trend among business organizations to motivate employees using non-monetary instruments such as involvement and influence, which requires the employees to have indepth understanding of the relevant strategic issues.22 Some researchers, however, are questioning the importance of this value substitution at times of high unemployment rates and strong focus on cost rationalizations.23 A number of major researchers within the field of strategic planning devote much attention to the discussion of problems in the implementation of strategies and policies, as implementations far too often remain unsuccessful.24 A parallel can also be drawn with Repenning and Sterman’s view on improvement programs, which argue that successful implementation of new methods represents a bigger challenge than identifying or learning new improvement methods.25

22

23

24

25

See Schein, Edgar H.: Organisationspsykologi, Danish translation, Herning, 1990, p. 53; Kieser, Alfred: “Human Relations-Bewegung und Organisationspsychologie”, in Kieser, Alfred (ed.): Organisationstheorien, 3 rd edition, Stuttgart, 1999, pp. 101—131. This viewpoint is discussed in Jöns, Ingela: Managementstrategien und Organisationswandel, Mannheim, 1995, p. 156. See Preface, Warren, Kim: Competitive Strategy Dynamics, Chichester, 2002. For a theoretical discussion of the implementation problem, see also McPherson III, L. Fillmore: “Organizational Change: An Industrial Dynamics Approach”, in Edward B. Roberts (ed.): Managerial Applications of System Dynamics, Cambridge, Massachusetts, 1978, pp. 447—449. See Repenning, Nelson P. and John D. Sterman: “Nobody Ever Gets Credit for Fixing Problems that Never Happened: Creating and Sustaining Process

10

A. Organizational Intervention Skills as Corporate Competence

Senge furthermore lists a number of studies documenting significant failure rates in achieving sustainable change in top management-driven change initiatives.26 In other words, the implementation of a solution to a strategic problem can even constitute a bigger challenge than finding the solution.27

II.

Foundation and Strategies for Planned Change Interventions

Lewin is often accredited as being the intellectual founder of ‘planned change’ organizational interventions, and he is still one of the most frequently quoted authors in the social sciences.28 Planned change refers to attempts where change is “conscious, deliberate, and intended, at least on the part of one or more agents related to the change attempt”.29 The theories of planned change are concerned with the complex processes of learning and change necessary to overcome the normal resistance most humans have towards change, even when the goals are apparently highly beneficial.30 Planned change in Lewin’s perspective is also known as the “unfreezing-movement-freezing” process. Social systems seem to have some sort of “inner resistance” to change, indicating that in spite of the

26 27

28

29

30

Improvement”, California Management Review, Vol. 43, No. 4, Summer 2001, p. 65. Senge, Peter M.: The Dance of Change, New York, 1999, pp. 5—6. This does not imply that strategy making is easy. Rethinking strategies and entire business concepts is both a difficult and crucial task in given situations; see Hamel, Gary: Leading the Revolution, Boston, 2000, p. 28; Fine, Charles H.: “Clockspeed-based strategies for Supply Chain Design”, Production and Operation Management, Vol. 9, No. 3, 2000, pp. 213—221. See preface in Gold, Martin: The Complete Social Scientist: A Kurt Lewin Reader, Washington, 1999; Schein, Edgar H.: Organisationspsykologi, Danish translation, Herning, 1990, p. 249. See Chin, Robert and Kenneth D. Benne: “General Strategies for Effecting Changes in Human Systems”, in Bennis, Warren G., Kenneth D. Benne and Robert Chin: The Planning of Change, 4 th edition, New York, 1985, p. 22. The intervention process described by Argyris concerning moving an individual, a group or an organization towards Model II theory-in-use and double-loop learning is an example of an extensive and detailed planned change intervention in action research tradition, see Argyris, Chris: Reasoning, Learning, and Action – Individual and Organizational, San Francisco, 1982, especially pp. 162—162 and pp. 468—474.

A. Organizational Intervention Skills as Corporate Competence

11

application of a force, the social process will not change without an additional force to break the habit, to unfreeze existing customs.31 Unfreezing existing behavior or attitudes can take place through the mechanisms of weakening existing behavior or attitudes, then the establishment of feelings of dislike regarding the present situation, and also establishment of psychological feeling of safeness in the change process.32 The second phase in the planned change process is the actual change part with development of new attitudes and behavior based on new information and cognitive and affective redefinitions. The last phase in the planned change process (freeze) is concerned with how to change behavior in a sustainable way, avoiding it sliding back to its old level in a short time.33 To strive for sustainable and continuous benefit of new attitudes and behavior, useful mechanisms could be testing the congruence between the change and the individual’s own situation, team-building efforts and continuous support, or recognition from both formal and informal leaders in the organization.34 Lewin’s theories of planned change originally focus on cognitive and behavioral change aspects related to a specific change situation, but researchers often also emphasize the broader term organizational learning. Argyris and Schein both place interventions as part of the continuous learning and forming of the organization and its change readiness, Senge emphasizes the importance of improving system thinking skills, and Sterman emphasizes the importance of helping organizations to improve the critical thinking skills necessary to challenge future mental models and biases, opposed to only helping to solve a specific problem.35

31

32 33

34

35

Lewin, Kurt: “Group Decision and Social Change” (first published in Newcomb and Hartley’s Readings in social psychology, 1948, pp. 330—341), in Gold, Martin: The Complete Social Scientist – A Kurt Lewin Reader, Washington, 1999, p. 281. See Schein, Edgar H.: Organisations Psykologi, Herning, 1990, pp. 254—255. Lewin, Kurt: “Group Decision and Social Change”, in Gold, Martin: The Complete Social Scientist – A Kurt Lewin Reader, Washington, 1999, p. 265. The arguments include, that behavior observed in a training program is often not continued when the person goes back to his normal routines. See Schein, Edgar H.: Organisationspsykologi, Herning, 1990, pp. 256—257; Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, pp. 22—30. Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Ma., 1970, chapters 1 & 2; Schein, Edgar H.: Organisations-

12

A. Organizational Intervention Skills as Corporate Competence

The planning of organizational interventions always, more or less deliberately, reflects underlying change strategies.36 Chin and Benne have developed a taxonomy of strategies for effecting changes in human systems, consisting of three types of general strategies:37

36 37

38

•

Empirical-Rational Strategies assume changes to be adopted if they are rationally justified. Examples in business organizations include allocation of funding, personnel replacement and scientific management projects.38

•

Normative-Re-educative Strategies focus the change process on attitudes, values and skills, inspired from the fields of sociology and psychology. These strategies are often less concerned with solving specific problems but rather focus on organizational development and improving the problemsolving capabilities of the organization.

•

Power-Coercive Strategies emphasize political and economical sanctions in the exercise of power, or even playing upon sentiments of guilt and shame. Top-down implementation of new strategies or policies often has implicit elements of use of power.

psykologi, 1990, p. 40; Senge, Peter M.: The Fifth Discipline, New York, 1994, pp. 57—67; Sterman, John D.: “All models are wrong: reflections on becoming a systems scientist”, System Dynamics Review, Vol. 18, No. 4, Winther 2002, p. 526. Borum, Finn: Strategier for organisationsændringer, Copenhagen, 1995, p. 15. Chin, Robert and Kenneth D. Benne: “General Strategies for Effecting Changes in Human Systems”, in Bennis, Warren G., Kenneth D. Benne and Robert Chin: The Planning of Change, 4 th edition, New York, 1985, pp. 22—45. The school of scientific management dates back to the early 20th century with Taylor’s work on rational optimization of work processes. The most well known mechanism element might be the detailed time studies of work procedures, although this should be seen in context with the underlying principles, including focus on science development, scientific basis for selection and development of workmen, and friendly cooperation between the management and the men, see Taylor, Frederick W.: The Principles of Scientific Management, New York, 1967, (first published in 1911), pp. 129—130.

A. Organizational Intervention Skills as Corporate Competence

13

Another taxonomy proposed by Borum includes The Technical-Rational Change Strategy, The Humanistic Change Strategy and The Political Change Strategy.39 The first two types in each of the taxonomies are rather similar, whereas Borum’s political change strategy has less focus on coercive elements and more focus on interpersonal negotiations, handling of personal interests, personal power-bases and interpersonal conflicts. Although using the terminology ‘change mindsets’ rather than ‘change strategies’, Anderson and Anderson compare the ‘industrial mindset’, including power and control, predictability, discrete events, with the ‘emerging mindset’, including participation, uncertainty and self-organization.40 The study of change strategy taxonomies contributes to enabling change agents to deliberately construct intervention strategies based on scientific research, and intervention strategies of large and complex corporate change programs are likely to combine elements from two or more of the general change strategies. A central element in such intervention strategies is the overcoming of human resistance to change.41 Literature discussions on change resistance mostly center on attitudes and intentions towards change, with attitudes being influenced by cognitive, affective and conative elements. Ajzen offers a widely used framework for the study of change of behavior based on the change

39

40

41

In Borum, Finn: Strategier for organisationsændringer, Copenhagen, 1995, the change strategies are discussed thoroughly, and at p. 117, a schematic overview can be found. Furthermore a fourth change strategy regarding network organizations/communities is discussed. Anderson, Linda A. and Dean Anderson: “Awake at the Wheel: Moving beyond Change Management to Conscious Change Leadership”, OD Practitioner, Vol. 33, No. 3, 2001, p. 44. The two mindsets seem to a high extend to correspond with the traditional two views on Man-in-Organization: The Human View vs. The Resource View, see Leavitt, Harold J., William R. Dill, and Henry B. Eyring: The Organizational World – A systematic view of managers and management, New York, 1973, pp. 122—123. Overcoming human resistance to change is among the most discussed topics the literature of strategy implementation and change management, see Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, pp. 154—173; Argyris, Chris: Interventions Theory and Method – A behavioral Science View, Reading, Ma., 1970, p. 70.

14

A. Organizational Intervention Skills as Corporate Competence

of intentions leveraging on attitudes, norms and perceived behavioral control.42 This framework will be discussed further in chapter B. The intervention context, primarily understood as organizational and problem characteristics, is the main factor influencing the intervention strategy and the decisions on intervention mechanisms.43 Context elements include the organization’s management traditions such as authoritarian or democratic decision processes.44 According to Jöns, when strategy implementation also encompasses changes in the organization’s philosophies, it is insufficient that the implementation process addresses qualifications and acceptance; the implementation process must also address the underlying values of the organization and to a higher degree includes employee development, information and participation.45 Different science schools exist in regards to using theory or the real world as the starting point for an academic research approach.46 In the field of planned change, action research belongs to the most recognized approaches. Action research takes the theoretical point of departure, that dynamic systems such as organizations can be examined through carefully planned, theory-based interventions.47 Action research is formative as well as summative, as the 42

43

44

45

46

47

Ajzen, Icek: Attitudes, Personality and Behavior, Chicago, 1988, pp. 20—131. See also Rouwette, Etiënne: Group model building as mutual persuasion, Nijmegen, 2003, pp. 104—111, for a discussion on Ajzen’s framework. See Rouwette, Etiënne: Group model building as mutual persuasion, Nijmegen, 2003, p. 103. See Schein, Edgar H.: Organisationspsykologi, Herning, 1990, p. 142 for a literature overview of management traditions in regards to involvement of subordinates in decisions. Jöns, Ingela: Managementstrategien und Organisationswandel, Mannheim, 1995, p. 157. An extensive literature-based discussion of the relationship between practice and theory in organizations theories in general (not specific related to action research) can be found in Scherer, Andreas G.: “Kritik der Organisation oder Organisation der Kritik? Wissenschaftstheoretische Bemerkungen zum Umgang mit Organisationstheorien”, in Kieser, Alfred (ed.): Organisationstheorien, 3 rd edition, Stuttgart, 1999, pp. 1—37. See Schein, Edgar H.: Organisationspsykologi, Herning, 1990, pp. 249—259; Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, pp. 22—30. Action research has some parallels to the field of cybernetics, where focus is on behavior of systems (what does it do)

A. Organizational Intervention Skills as Corporate Competence

15

interventions are altered if resultant data or changing conditions suggest the appropriateness.48 Action research projects do not involve traditional science evaluation models, e.g. the use of control groups, as the complex social environment is not controllable to a degree that allows isolation of the true variables. Furthermore, action research projects and practical use is supposed to coexist with mutual benefit, also in regards to evaluation. Evaluation of human systems will influence the system, for example as a Hawthorne effect or as expectation settings.49 Consequently, interventions in action research tradition must be designed in a way where the academic evaluation is not counterproductive with regards to the desired results of the intervention.

III.

The Usage of System Dynamics Modeling in Organizational Interventions

On an abstract, conceptual level, modeling takes place in organizations all the time. Every time a decision is made, the decision maker’s cognitive model of the situation will influence the decision. Figure A-4 is a model of decision maker (D) and the process (P), which is the target for goal-seeking decision making. Ideally, rational decision-making should be a function of P, input and output of P, as well as system goals (G) and decision principles (DP).50

48

49

50

rather than on a detailed understanding of the system elements (what is this thing), see Ashby, W. Ross: An Introduction to Cybernetics, paperback version, London, 1964, p. 1. See Gold, Martin: The Complete Social Scientist – A Kurt Lewin Reader, Washington, 1999, p. 253. The Hawthorne effect is widely discussed in various literatures and typically refers to performance improvements among workers participating in experiments, although no theoretical basis exists for the effects. According to Wikipedia, accessed April 2006, Mayo Elton has interpreted the performance improvement among workers (the Hawthorne effect) as: “it was the feeling they were being closely attended to that caused the improvement in performance.” Only external information input (u e) is made explicit, as resource input is viewed as controllable in P. M, Ue, Y, represent the sets of values of the manipulating value, the external input and the outcomes. The decision problem is to find the decision variable (m) in M such that G(m,ue,P(m,u e)) is maximized. For further

16

A. Organizational Intervention Skills as Corporate Competence

G, DP

m

D, P

ue

P

y

Figure A-4: Goal-seeking system51

However, in social-economic systems, due to uncertainty and complexity, only a limited cognitive model of P will be available (Pm) for the decision-making. Furthermore, D will be characterized by perceptions, motivation, and personal values. Figure A-4 illustrates the often-implicit use of modeling in decisionmaking, whereas system dynamics offers an explicit, deliberate use of modeling.52

51 52

description see Takahara, Yasuhiko: “A Formal Model of Organization”, in Takahashi, Singo, Kyoichi Kijima and Ryo Sato (eds.): Applied General Systems Research on Organizations, Tokyo, 2004, pp. 15—21. Takahara, Yasuhiko: “A Formal Model of Organization”, Tokyo, 2004, p. 16. It should be noted, that a model is always only a limited reflection of a real system, representing a given viewpoint on a problem or a system, based on human decisions on parameters and structures to be included in the model. In Sterman, John D.: “All models are wrong: reflections on becoming a systems scientist”, System Dynamics Review, Vol. 18, No. 4, Winter 2002, p. 525, a model is called “a simplification, an abstraction, a selection“ and “inevitably incomplete, incorrect – wrong.” Nevertheless, modeling offers an opportunity to overcome a number of the problems in unsupported decision-making, as discussed later in this chapter.

A. Organizational Intervention Skills as Corporate Competence

17

The system dynamics field has its origin as a primarily analytical and rational oriented problem-investigating and policy-forming discipline.53 Forrester states that the purpose of system dynamics in corporate environments is to aid in the design of improved industrial and economical systems, and system dynamics has over the years contributed significantly to create insight being used in strategic planning and policy design.54 System dynamics offers a complementary opportunity to analyze complex and dynamic problems, as most of the traditional tools offered by the strategic planning field, are largely static, and thereby often insufficient in our present-day environment of complexity and dynamics, consequently resulting in actions often being made based on intuition and experience.55 System dynamics addresses the need for decision makers to learn and understand complex problems and situations. Since Descartes, cognitive science has been interested in how humans learn.56 Human brain processes are eventorientated, which – without long experience or effective learning - makes it 53

54

55

56

In Forrester, Jay W.: Industrial Dynamics, Cambridge, 1961, p. 56, objectives in using mathematical models are described as follows: “A mathematical model of an industrial enterprise should aid in understanding that enterprise. It should be a useful guide to judgment and intuitive decisions. It should help establish desirable policies.” Milling, Peter: “Leitmotive des System-DynamicsAnsatzes”, Wirtshaftswissenschaftliches Studium, Vol. 10, 1984, p. 508, also supports this understanding of system dynamics: “System Dynamics verwendet formale Modelle, um zu einem verbesserten Verständnis des zu studierende Phänomens zu gelangen und um Eingriffe in das System auf ihre Konsequenzen hin zu untersuchen.” See Forrester, Jay W.: Industrial Dynamics, Cambridge, 1961, p. 115. For examples of the practical usage of system dynamics, see the numerous cases published in System Dynamics Review over the years. See Lyneis, James M.: Corporate Planning and Policy Design: A System Dynamics Approach, Massachusetts, 1980, p. 3; and Warren, Kim: Competitive Strategy Dynamics, Chichester, 2002, preface; Mintzberg, Henry: The Rise and Fall of Strategic Planning, New York, 1994, p. 319. In René Descartes first major contribution, 1628, “Regulae ad directionem ingenii,” regarding rules for the use of the human’s cognitive means, a method for acquiring scientific or any other type of rational founded insight is described, see Lübcke, Poul (ed.): Politikens filosofi leksikon, Copenhagen, 2001, pp. 82—87. Wikipedia (accessed April 2006) describes the work as a method for scientific and philosophical thinking and translates the title of the book into “Rules for the Direction of the Mind.”

18

A. Organizational Intervention Skills as Corporate Competence

difficult to comprehend feedback-loops of even relatively simple and small dynamic structures, resulting in problem-solving in the area of strategic, complex, and dynamic problems often not taking unwanted side-effects, delayed reactions and policy resistance into consideration.57 Although companies deal with these types of complex problems every day, they have often not been solved using analytical tools, but have in many respects been managed based on past experiences on what works and what does not – with respect to the existing production facilities, portfolios of customers, products, etc. This experiencebased decision-making is regularly implemented as heuristics, rules of thumb, organizational routines or the use of simplifications and traditions.58 No explicit, formal models underlie this type of decision-making, but the experience-based decision-making is building on mental models of individuals. While intuitive, implicit knowledge such as simple heuristics and experience proves to be helpful in many situations (particularly when dynamics are rather low), it is insufficient in innovative and rapidly changing situations.59 This dependence on erroneous intuitive solutions is, in the view of Forrester, the cause of most misbehavior in corporate systems.60 System dynamics can help decision-makers cope with and understand situations and problems, that would have taken years to understand based on empirical experiences. The accelerated learning is partly due to the structured emphasis on understanding and exploring how behavior is influenced by corporate structures and policies, partly due to the aid of computer modeling to design improved policies and resource allocation and the utilization of the computers’ ability to calculate thousands of iterations, and partly due to the 57

58

59

60

See Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, pp. 10—11; Kampmann, Christian P. E.: Feedback complexity and market adjustment – An experimental approach, Boston, 1992, p. 31; Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993, pp. 29—30; Dörner, Dietrich: The Logic of Failure, New York, 1996, pp. 38—42. Größler, Andreas: “A Content and Process View on Bounded Rationality in System Dynamics”, Systems Research and Behavioral Science, Vol. 21, No. 4, July/August, 2004, p. 320. Bonabeau, Eric: “Don’t Trust Your Guts”, Harvard Business Review, May 2003, pp. 118—119. Forrester, Jay W.: “System Dynamics, System Thinking, and Soft OR”, System Dynamics Review, Vol. 10, No. 2, Summer 1994, p. 249.

A. Organizational Intervention Skills as Corporate Competence

19

creation of improved and shared mental models among the decision-makers across multiple organizational units.61 De Geus gives three main reasons for going through the trouble of making and simulating computer models.62 These main reasons are: (1) that most people only are able to deal with a few variables at a time, and this only in one or two time iterations, (2) the need for separation of cause and effect in time and space, and (3) computer models help to identify what information is most relevant. The first reason in particular, is also supported in Miller’s work on limitations on the amount of information humans are able to receive, process and remember.63 De Geus’ arguments further include statements on computer models, often revealing counter-intuitive behavior, which is a view also supported by Lane.64 Milling furthermore emphasizes the synergy of combining human creativity with the capabilities and the power of high-speed computing.65 The system dynamics literature often points out promising results from the transfer research regarding computer simulations, which according to Bakken “may be attributed to motivational side-effects of the interactive pedagogy”.66 Lastly, the data acquiring process in quantitative modeling processes is in itself valuable for understanding

61

62

63

64

65

66

Lyneis, James M.: Corporate Planning and Policy Design – A System Dynamics Approach, Massachusetts, 1980, p. 9 and p. 15; Lane, David C.: “Should System Dynamics be Described as a ‘Hard’ or ‘Deterministic’ System Approach?” Systems Research and Behavioral Science, Vol. 17, 2000, p. 4. de Geus, Arie P.: “Planning as Learning”, Harvard Business Review, Vol. 66, No. 2, March-April 1988, pp. 70—74. Miller, George A.: “The Magical Number Seven, Plus Minus Two: Some Limits on Our Capacity for Processing Information”, The Psychological Review, Vol. 63, No. 2, March 1956, p. 95. Lane, David C.: “Should System Dynamics be Described as a ‘Hard’ or ‘Deterministic’ System Approach?”, Systems Research and Behavioral Science, Vol. 17, 2000, p. 4. Milling, Peter: “Modeling Innovation Processes for Decision Support and Management Simulation”, System Dynamics Review, Vol. 12, No. 3, 1996, p. 227. Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993, p. 31.

20

A. Organizational Intervention Skills as Corporate Competence

the problem parameters, and stimulates best practice discussions and brings to surface misperceptions among key individuals.67 In the 1990’s a school within the field of system dynamics, oriented towards participative modeling approaches, emerged with increased embracing of softer aspects such as organizational learning, group processes, and the importance of consensus and commitment.68 This development might follow from the change in organizational structures in many organizations. Modern organizations with a high degree of employee empowerment typically have a need for a large number of people to have an understanding for the whole of the organization and its strategy, including the dynamics and the interdependencies, to be able to make the right decisions in their daily work as well as for motivational factors. This to some degree substitutes the “old way” with a few executives directing strategies and policies to be implemented downwards in the organization. The challenge of interventions nowadays is therefore not only to find good solutions to problems or new situations. The solution must also be understood and find acceptance among the many stakeholders, and efforts of establishing internal commitment in organizational interventions are often centered on the creation of awareness, consensus, and confidence regarding the goals and the change process.69 Although system dynamics projects (participative modeling, in particular) are concerned with both cognitive and behavioral aspects relevant for implementation, system dynamics modeling efforts are typically elements in organizations’ strategy forming, with less focus on strategy implementation. However, the value creation of a corporate modeling study is seldom “a new

67

68

69

See Warren, Kim: Competitive Strategy Dynamics, Chichester, 2002, p. 30; Snabe, Birgitte and Andreas Größler: “Targeted Participative Modelling as Organisational Intervention: Concept and Case Study”, Journal of Systems Research and Behavioral Science, Vol. 23, No. 4, in print, 2006, p. 20. Through the introduction of participative model-building methodologies and “planning as learning”, focus has been put on creating conceptual insights, changing mental models of decision-makers and creating consensus and commitment; see Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 97; de Geus, Arie P.: “Planning as Learning”, Harvard Business Review, March-April 1988, p. 70; Lane, David C.: “Modelling as Learning: A consultancy methodology for enhancing learning in management teams”, European Journal of Operational Research, Vol. 59, No.1, 1992, pp. 64—84. See Akkermans, Henk: Modelling With Managers, Breda, p. 20.

A. Organizational Intervention Skills as Corporate Competence

21

policy” to be implemented subsequently, but rather the learning that appears among modeling participants through the exploratory process examining the problem and the system behavior.70 New improved mental models among decision-makers will constitute a part of the solution, as the new insight and new decided policies will influence future operational decisions. In many cases, no formal implementation is needed, as the learning among decision-makers is all that was required to solve the problem.71 Some projects, nevertheless, are taking place in a context, where more formal implementation is needed, and for this reason most of the system dynamics literature and textbooks calls for the implementation challenge to be considered throughout the intervention process, and also considers the iterative nature of the process.72 The purpose of a system dynamics modeling study is likely be articulated as an exploratory exercise addressing a problem and possible solutions, and only seldom to be stated as strategy forming or strategy implementation. Although seen from a strategic organizational intervention viewpoint, addressing the context of the modeling studies, such a modeling study has conceptual orientation towards strategy forming due to the strong Cycle I focus. As explicit support of the change management phase (Cycle II), in the implementation efforts when a strategic decision is already made, almost no tradition exists for using system dynamics in a modeling-oriented way, as most implementation-oriented SD

70

71 72

See Lane, David C.: “Modelling as Learning: A consultancy methodology for enhancing learning in management teams”, European Journal of Operational Research, Vol. 59, No.1, 1992, p. 64; Vennix, Jac A. M.: Group Model Building, Chichester, 1996, pp. 98—99. See Akkermanns, Henk: Modelling With Managers, Breda, 1995, p. 17. Sterman, John D.: Business Dynamics, Boston, 2000, p. 80 and p. 88; Roberts, Edward B.: “Strategies for Effective Implementation of Complex Corporate Models”, in Edward B. Roberts (ed.): Managerial Applications of System Dynamics, Cambridge, 1978, pp. 79—84. A more critical view on system dynamics efforts in organizational interventions can be found in Zock, Alexander: “A critical review of the use of System Dynamics for organizational consulting projects”, at CD-ROM of Proceedings, System Dynamics Conference, System Dynamics Society, 2004, p. 7, where it is argued that not even the participative modeling approaches are sufficient attentive to the overall challenges of change processes.

22

A. Organizational Intervention Skills as Corporate Competence

studies and approaches are concerned with gaming-oriented simulations such as flight simulators and planning games.73 Taking the view of change management versus decision-making or strategy forming, two conceptually different usages of system dynamics modeling in organizational interventions appear: 74 •

Exploratory modeling supporting diagnosing, learning and decisionmaking75

•

Transfer-oriented modeling usage supporting change management

The focal point in this differentiation is the purpose of the usage of system dynamics: exploring a problem versus to transferring existing insights. Modeling used in supporting diagnosing, learning and decision-making is driven by the desire to explore and understand system behavior and to identify and simulate possible new policies addressing a complex problem. Such interventions do not have the same degree of control characteristics as typical implementation projects, for example detailed project plans or thorough stakeholder analyses and 73

74

75

The use of modeling-oriented simulations vs. gaming-oriented simulations comes from the taxonomy proposed by Maier, Frank und Andreas Größler: “What are we talking about? A taxonomy of Computer Simulations to Support Learning”, System Dynamics Review, Vol. 16, No 2, 2000, p. 143. The term ‘modeling-oriented simulations’ does not refer to the context of the project in regards to decision-making or implementation. In this dissertation, strategic formulation is understood to include both strategic planning and policy formulation. In the system dynamics society, the term “policy formulation“ is often used as the aim of modeling projects with policies being rules stating how the day-by-day operating decisions are made, see Forrester, Jay W.: Industrial Dynamics, Cambridge, 1961, p. 93. Strategies are constituted by both corporate goals and corporate policies, and strategic planning is defined as the process of transforming corporate goals into policies, see Lyneis, James M.: Corporate Planning and Policy Design: A System Dynamics Approach, Boston, 1980, p. 19 and p. 3. ‘Exploratory modeling’ should not be mistaken with ‘exploratory models’ as described in Homer, Jack B.: “Why we iterate: scientific modeling in theory and practice”, System Dynamics Review, Vol. 12, No. 1, Spring 1996, p. 1. Homer defines exploratory models as a kind of easy-made, draft models less occupied with validation. In this dissertation ‘exploratory modeling’ refers to the purpose of the project: to explore and understand a given problem, no matter if the model used is less detailed or if it is highly developed and refined with scientific rigor.

A. Organizational Intervention Skills as Corporate Competence

23

communication plans. This is due to the fact, that for exploratory modeling interventions, the organizational change process cannot be defined before the outcome of the modeling process is (at least to some extent) clear.76 Often, a strategy forming modeling intervention will result in changed mental models among decision-makers; frequently, implementation will not take place in an explicit, planned change manner.77 Exploratory modeling can take place as ‘participative modeling’ or as ‘expert modeling’, the difference primarily being the way people are involved. In expert modeling, people – apart from the main decision-makers and a few modelers - are primarily involved for information collection purposes.78 In participative modeling, such as Group Model Building and Modeling for Learning, people representing an extensive array of viewpoints are involved in the modeling process itself, with strong focus on mental model alignment and refinement.79 Transfer-oriented usage of system dynamics modeling belongs to the planned change type of organizational interventions and has some common characteristics with the field of action research, with its parallel focus on the implementation of planned change as well as on continued knowledge 80 development. The modeling project supports implementation processes where a

76

77

78

79 80

In Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 99, it is argued that learning cannot be predicted in the outset of a project. In Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 97, it is argued that insights are conceptual rather than instrumental, and although stressing that the purpose of system dynamics is performance improvement, he also states (p. 99) that “implementation becomes evasive.” Richardson, George P. and Alexander L. Pugh: Introduction to System Dynamics Modeling with DYNAMO, Cambridge, 1981, p. 355, write that “a modeling study usually focuses on what policies will help, not on how those policies ought to be introduced into the system.” See Forrester, Jay W.: “Policies, decisions and information sources for modeling”, European Journal of Operational Research, Vol. 59, No. 1, 1992, pp. 59—60, and Forrester, Jay W.: Industrial Dynamics, Cambridge, 1961, p. 364, where it is recommended to use industrial dynamics in a business company by initiating in a small, exclusive group of people with the right qualities to go in-depth with the dynamics of the company including the “innermost secrets and hopes of the organization.” Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 112. For definitions of action research, see Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, p. 23.

24

A. Organizational Intervention Skills as Corporate Competence

strategy forming process has outlined business objectives and targets, but where the optimization of the strategy or policy is left as a part of the implementation. Therefore a balance is needed between the initial detail level of the objectives 81 and the degree of freedom to make decisions in the implementation process. This could be called “framing the intervention”, giving participants empowerment to explore, decide and act within a given ‘frame’ (how to do), but not to explore, decide and act outside the given ‘frame’ (what to do). Transferoriented usage of modeling could be called instrumental usage of modeling for change management purposes. In respect to transfer of existing learning from one group of people to another group of people, this type of modeling has similarities with gaming-oriented simulations, such as flight simulators or educational games. Gaming-oriented simulations make use of fixed models, with the purpose being to transfer the understanding of the causal relations and the behavior of the system. Compared with gaming-oriented simulations, transfer-oriented modeling to a higher extends aims to transfer commitment in the search for sustainable change. Through involvement and participation in modeling sessions, implementers take part in the refining of the change program and the operationalization of the implementation. In his description of the system dynamics process, Forrester proposes a phase called “Educate and Debate”, in which consensus for implementation is aimed for.82 The phase is placed after the actual exploratory modeling, but is expected to raise questions resulting in repeated analysis in the previous phases. If modeling were to be used in the Educate and Debate phase, it would be an example of change management oriented modeling. Although for transfer-oriented usage of modeling, it may or may not be the case that a model has been developed in an earlier strategy formulation phase.83 The modeling can also be based on a model especially

81

82

83

Borum, Finn: Strategier for organisationsændringer, Copenhagen, 1995, p. 58, discuss the problem of “a free, informed choice as a condition for establishment of commitment” in a change process planned and controlled by consultants. Forrester, Jay W.: “System Dynamics, System Thinking, and Soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, p. 247. An interesting case, where a modeling project was continued into instrumental implementation activities is the well-known “Maintenance Game” described in Repenning, Nelson P. and John D. Sterman: “Nobody Ever Gets Credit for Fixing Problems that Never Happened: Creating and Sustaining Process Improvement”, California Management Review, Vol. 43, No. 4, Summer 2001, pp. 64—88.

A. Organizational Intervention Skills as Corporate Competence

25

drafted for the project. Using such a preliminary model of the problem-system, the process should allow for interactive refining and evaluation of the model itself, and through both model adjustments, model enhancements and model simulations, the modeling approach has the threefold aim of change program refining, transfer of system understanding, and establishment of commitment. The design of participative modeling interventions supporting change management can draw from the normative, prescriptive management literature of planned change with regards to activities such as intervention planning, stakeholder management and implementation planning and review. Intervention planning includes the definition of business objectives and targets, the framing of the intervention, the identification of consultation relationships, roles and responsibilities in the project organization, and time and budget planning. Stakeholder management involves a thorough analysis of all the major interest groups and individuals who have significant influence - directly or indirectly - on the success of the intervention. Focus is on interests and power, importance for solution design and implementation, and relevant means of involvement and 84 communication. Stakeholder analysis is a major input to intervention planning, both to secure relevant parameters to be included in the process, and to secure appropriate involvement and communication with stakeholders and other 85 employees. Implementation planning and review deals with the planning of the implementation, including a communication plan and a clear assignment of responsibilities. The communication plan develops over the course of the intervention and includes elements such as motivating change and the communication of visions, results, implementation plan and successes. Planning activities for implementation should to be understood as something to be done

84

85

See Flood, Robert L. and Michael C. Jackson: Creative Problem Solving – Total Systems Intervention, Chichester, 1991, p. 12; Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Massachusetts, 1970, p. 81; and Borum, Finn: Strategier for organisationsændringer, Copenhagen, 1995, pp. 77—89. For discussions on “Employee Involvement”, see both Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, p. 317; and Thun, Jörn-Henrik, Peter M. Milling, and Uwe Schwellbach: “The Impact of Total Employee Involvement on Time-based Manufacturing”, in Blackmon, Kate, Steve Brown, Paul Cousins, Andrew Graves, Christine Harland, Richard Lamming, and Harvey Maylor (eds.): “What Really Matters in Operations Management”, Bath, 2001, pp. 133—135.

26

A. Organizational Intervention Skills as Corporate Competence

only after a strategy or policy has been designed. Involving the right people in the right way early in the process might be one of the most important criteria for successful implementation later, together with communication strategies including timely information and dialogues. Also, the iterative process continues after implementation activities, as follow-up activities will create learning to be used for further corrective actions. The main differentiator of modeling supporting change management compared to other approaches within the field of system dynamics, is the usage and utility of participative modeling in a new context, namely in strategy implementation rather than strategy formation. The main research objective is to investigate whether change management-oriented participative modeling seems to be an effective approach seeking sustainable change through: •

Transfer of insight from decision-makers to implementers in such a way that not only the decisions but also the underlying arguments are effectively transferred,

•

Allow true involvement of implementers through participative strategy refinement within a given decision ‘frame’ and strategic direction.

The rest of the dissertation is structured as follows: chapter B is a theoretical discussion of the conceptual foundation for the usage of system dynamics in organizational interventions. The discussion is not limited to usage of system dynamics in change management, but rather it aim to investigate the general purposes and methods for using system dynamics, which is traditionally placed in decision-making and strategy and policy forming contexts. Focus on usage of system dynamics in a change management context, begins in chapter C, which describes a field study and an underlying research approach. In Chapter D the insights from the field study are discussed in terms of both the theory basis from chapter B and the normative, prescriptive management literature from the field of organization development and planned change. Chapter E concludes with the discussions from the previous chapters and proposes areas for further research.

B. Conceptual Foundation for the Usage of System Dynamics The description of the conceptual foundation for the usage of system dynamics involves a journey through a variety of disciplines: although it is based in mathematics, physics and engineering, system dynamics also draws on cognitive and social psychology, economics and other social sciences.86 According to Martinez and Richardson, conceptual differences in research designs can be discussed in terms of theory, method and procedure elements.87 The theories underlying the usage of system dynamics (why use system dynamics) are discussed in chapter B.II: Cognitive and Behavioral Rationale for the Usage of System Dynamics. The methods and procedures for the usage of system dynamics in organizational interventions (how system dynamics is used) are described in chapter B.III: The Development Process of System Dynamics Models in Corporations. Before the theory, method and procedure discussions, however, chapter B.I seeks to place the usage of system dynamics within the overall context of decision-making.

I.

The Usage and Utility of Modeling in Decision-Making

In the field of decision-making Baron discusses utility theory as a normative model concerned with elements of (1) the trade-off between the probability of an outcome and its utility, (2) the trade-offs among different goals, (3) maximizing utility over all relevant people as a normative model for moral decisions, and (4)

86

87

Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, pp. 4—5. Martinez, Ignacio J and George P. Richardson: “An Expert View on the System Dynamics Modeling Process: Concurrences and Divergences Searching for Best Practices in System Dynamics Modeling”, at CD-ROM of Proceedings, System Dynamics Conference, System Dynamics Society, 2002, p. 25.

28

B. Conceptual Foundation for the Usage of System Dynamics

handling conflicts among outcomes that occur at different times.88 The rationality based expected-utility decision-making is often seen analyzed in the traditions of operations research, decision trees, game theories, etc. 89 Whereas principles of rational choice are considered as reasonable in abstract form, their implications are often violated in actual choices. In socioeconomic systems it is an illusion to assume perfect rational decision-making due to complexity, uncertainty and human factors. The topic of complexity and complex systems has been of great interest to scientists using terms such as theories of holism, cybernetics, general system theory, chaos theories etc. since World War I.90 Although not undertaking a formal definition of complexity, Simon explains a complex system as “one made up of a large number of parts that have many interactions.”91 In system theory traditions, Senge and Sterman describe complexity as consisting of detail complexity and dynamic complexity.92 Milling further divides detail complexity into three sub-dimensions: number of relevant elements (variety), number of connections between elements (connectivity), and functional relationship between elements (functionality).93

88

89

90

91

92

93

Baron, Jonathan: Thinking and Deciding, 3 rd edition, Cambridge, UK, 2000, pp. 223—243. See Baron, Jonathan: Thinking and Deciding, Cambridge, UK, 2000, p. 227, for a discussion on using game theory to examine expected-utility decision-making. Expected-utility is also a cornerstone in the expected-monetary-value method; see e.g. Tversky, Amos: “Additivity, utility and subjective probability”, in Edwards, Ward and Amos Tversky (eds.): Decision Making, 1967, pp. 208—238. In Simon, Herbert A.: The Science of the Artificial, 3 rd edition, Cambridge, 1996, pp. 169—181, an overall discussion is offered on the major scientific trends in this field. Simon, Herbert A.: The Science of the Artificial, Cambridge, 1996, pp. 183—184. Senge, Peter M.: The Fifth Discipline, New York, 1994, p. 71; Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, p. 21. Milling, Peter: “Kybernetische Überlegungen beim Entscheiden in komplexen Systemen”, in Entscheiden in komplexen Systemen, Wirtschaftskybernetik und Systemanalyse, Band 20, Berlin, 2002, p. 12.

B. Conceptual Foundation for the Usage of System Dynamics

29

Simon’s view on rationality is, that whereas human decision-making is not rational from an economic standpoint, it is still purposeful.94 Simon argues that in real-world context even rational expectationists are retreating from rational utility maximizing to a more realistic scheme of adaptive expectations.95 Decisions are constrained not only by human process capabilities, but also by an incomplete search for information. This only continues until a satisfactory solution is found (contrary to seeking an optimal solution).96 History is full of grave examples of people seeking to solve a problem and actually managing to worsen the situation despite the best intentions. This is often due to what Forrester calls “counterintuitive behavior of social systems”, or “policy resistance” in Sterman’s terminology, where unintended side effects and neglected feedback loops make a system behave differently from the intentions of the intervener.97 The literature offers extensive discussion on these phenomena: the descriptive literature identifies the deficiencies of traditional, unsupported decision making, whereas the prescriptive tradition offers a number of methods and techniques to overcome these limitations.98 When seeking to improve the organizational decision-making, organizations make different types of analyses and models. For problems characterized by feedback loops and delays, organizations can make use of

94

95

96

97

98

Herbert Simons view was expressed in the 1950’s, and discussed in Hogarth, Robin: Judgement and Choice – The Psychology of Decision, 2nd edition, Chicago, 1987, p. 63. Simon, Herbert A.: The Science of the Artificial, 3 rd edition, Cambridge, 1996, p. 39. Miller, George A.: “The Magical Number Seven, Plus Minus Two: Some Limits on Our Capacity for Processing Information”, The Psychological Review, Vol. 63, No. 2, March 1956, p. 95; Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 27. Forrester, Jay W.: “Counterintuitive Behavior of Social Systems”, in Collected Papers of Jay W. Forrester, Cambridge, 1975, p. 216. In Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, pp. 5—9, a larger number of examples of policy resistance are described. In Dörner, Dietrich: The Logic of Failure, New York, 1996, a few, but more detailed examples are discussed throughout the book; including failures in Third World efforts and the Chernobyl disaster. See Rouweette, Etiënne: Group model building as mutual persuasion, Nijmegen, 2003, pp. 19—29, for a description of descriptive and prescriptive view-points in decision-making.

30

B. Conceptual Foundation for the Usage of System Dynamics

explicit conceptual feedback models (system thinking). Examples of such system thinking efforts are soft operations research, cognitive mapping, structured system discussions etc. A particular category of explicit, conceptual feedback models is taking advantage of mathematical computer models and simulation. These are the system dynamics models. Figure B-1 tries to graphically depict the accumulative use of models in system dynamics.

System Dynamics System Thinking Intuition & Experience

Use of Formal, Mathematical Models (Simulation Models)

Use of Explicit, Conceptual Feed-back Models (e.g. CLD or Soft OR)

Use of Implicit, Experience-based Models (Mental Models)

Figure B-1: Accumulative levels of models in the usage of system dynamics

The three levels in figure B-1 are accumulative; i.e. in addition to the use of formal mathematical models, system dynamics also comprises explicit, conceptual feedback models as well as intuitive and experience-based models. This is in accordance with Kampmann, who stresses that intuitive assumptions underlie any type of model.99 The usage of qualitative and quantitative models serve the purpose of changing the mental models of the decision makers, as mental models are seen as a vehicle to change decisions and organizational action.100 The difference between system thinking and system dynamics cannot be 99

100

Kampmann, Christian P. E.: Feedback complexity and market adjustment, Boston, 1992, p. 28. Keough, Mark and Andrew Doman: “The CEO as organization designer – An interview with Professor Jay W. Forrester, the founder of system dynamics”, The

B. Conceptual Foundation for the Usage of System Dynamics

31

seen as reflecting the differences between hard and soft modeling approaches, as system dynamics is located somewhere between the two extremes.101 Hard modeling is a term used for single objective optimization, typically without taking people and organization into account.102 Although system dynamics uses mathematical formulas and relatively rigid model structures, it also encompasses soft modeling fundamentals like focus on generating debate and new insights about the problem at hand.103 The field of system dynamics often has vital debates concerning the advantages of qualitative modeling as seen in the system thinking area vs. the advantages of quantitative modeling as seen in system dynamics. The opinions differ from the one extreme, that only if a model is quantified and simulated, a study can be said to be complete, to the other extreme, that for a complex system with many soft relationships, quantification itself can be damaging.104 Some of the better known qualitative modeling approaches include Checkland’s Soft System Methodology (SSM), which is a “process of enquiry”, Eden’s Strategic Option Development and Analysis (SODA), using cognitive mapping for strategic options development and Senge’s use of Cause-LoopDiagrams (CLD) in building learning organizations.105 The goals of qualitative

101

102

103

104

105

McKinsey Quarterly, No. 2, 1992, p. 5; Kim, Daniel H. and Peter M. Senge: “Putting systems thinking into practice”, System Dynamics Review, Vol. 10, Nos. 2-4, Summer-Fall 1997, p. 280. In Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, pp. 4—5. Maani, Kambiz E. and Robert Y. Cavana: Systems Thinking and Modelling – Understanding Change and Complexity, Auckland, 2000, p. 21. In Forrester, Jay W.: “System dynamics, system thinking, and soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, p. 226, it is stated: “Understanding comes first, but the goal is improvement;“ in Fey, Willard and John Trimble: “The Evaluation and Development of Knowledge Acquisition in System Dynamics Studies”, in Proceedings, System Dynamics Conference, System Dynamics Society, 1992, p. 174, the process orientation of system dynamics is compared to the product (being a model) orientation among hard system developers. Groessler, Andreas, Peter Milling and Graham Winch: “Perspectives on rationality in system dynamics: a workshop report and open research questions”, System Dynamics Review, Vol. 20, No. 1, 2004, p. 84. Checkland, Peter: “Soft System Methodology”, in Rational Analysis, Jonathan Rosenhead (ed.): Chichester, 1989, pp. 71—100; Eden, Colin: “Using cognitive

32

B. Conceptual Foundation for the Usage of System Dynamics

modeling include individual learning, challenging and alignment of mental models, establishment of consensus and the search for ways to improve the system. Although quantitative modeling in organizational interventions also includes simulations, the overall goals seldom differ radically from the goals of qualitative modeling, but differ rather in the means of reaching those goals. Devotees of quantitative models argue that softer models like Cause-LoopDiagrams leave open the risk of different interpretations of the same model by different individuals. This is mainly due to the fact that the qualitative models do not incorporate any test of logic. In mathematical formal models, built-in logical constraints force model builders to have a more precise description and understanding of the model. Forrester states that qualitative studies to a higher degree depend on intuition compared to quantitative studies, where level and rate diagrams discipline the thinking process in model formulation and simulation.106 Forrester gives examples where Harvard Business School graduates arrive to wrong policy recommendations, inconsistent with their own quantitative system description.107 This is in accordance with observations by other researches observing students revealing significant differing interpretation of relatively simple Cause-Loop-Diagrams.108 Whereas the opinion differs with respect to the general applicability of quantitative vs. qualitative studies, there is a broad agreement that determinants for selection of methods include: (1) problem characteristics, (2) how the methods fit with the decision-making context, and (3) purposes and goals of the decision situation.109

106

107 108 109

mapping for strategic options development and analysis”, also in Rosenhead (ed.): Rational Analysis, Chichester, 1989, pp. 21—42; Senge, Peter M.: The Fifth Discipline, New York, 1994. Though it should be noted, that in Forrester, Jay W.: “System Dynamics, System Thinking, and Soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, p. 253 it is argued that Senge’s system archetypes and behavioral descriptions are based upon extensively explored system dynamics models. Forrester, Jay W.: “System Dynamics, System Thinking, and Soft OR”, System Dynamics Review, Vol. 10, No. 2, Summer 1994, p. 252. Forrester, Jay W.: “System Dynamics, System Thinking, and Soft OR”, p. 240. Observations by faculty members at Mannheim University. Milling, Peter: “Kybernetische Überlegungen beim Entscheiden in komplexen Systemen”, in Entscheiden in komplexen Systemen, Wirtschaftskybernetik und Systemanalyse, Band 20, Berlin, 2002, pp. 12—16.

B. Conceptual Foundation for the Usage of System Dynamics

33

The discussion has until now focused on the usage of modeling in decision-making. As a side-remark, it is interesting to note, that system dynamics has also been used to study the concept of decision-making. Based on a study investigating a multiplier-accelerator model of capital investments, Sterman concludes that it appears to be feasible to do experimental exploration of dynamic decision-making strategies in aggregate systems, with the results being directly compared to formal models of behavior.110 In this discussion, it is interesting to take a look at some of the critics of the usage of system dynamics in strategic decision-making. Mintzberg has criticized the utility of system dynamics, being concerned whether the methodology allows sufficient creativity, and he states that analytical thinking can be as wrong as intuitive thinking, especially as he finds that analysis does not seem to encourage creativity.111 It is worth noticing that Mintzberg in his criticism uses the argument that system dynamics focuses on analyses and aggregation and pays little attention to comprehending and synthesizing, which is a standpoint that most system dynamics practitioners have opposed to.112 It is also worth to notice that his criticism was stated more than 20 years ago. The field has developed since then, especially in the “softer” aspects with extensive research within system thinking and participative model-building approaches, which focus on discussions and involvement. Nevertheless, his concerns regarding creativity should still be taken into consideration, as additional focus on how to conceive creative potential new policies to be simulated in system dynamics models could add value.

110

111

112

Sterman, John D.: “Misperceptions of Feedback in Dynamic Decision Making”, in Milling, Peter M. and Erich O.K. Zahn (eds.): Computer-Based Management of Complex Systems, Proceedings of the 1989 International Conference of the System Dynamics Society, 1989, p. 30. In Mintzberg, Henry: The Rise and Fall of Strategic Planning, New York, 1994, pp. 298—299 and pp. 326—328, system dynamics is criticized for being shallow in depth and not embracing creativity and intuition, although on pp. 376—378 in the same book, credits are given to a number of system dynamics case stories. The whole article of Lane, David C.: “Should System Dynamics be Described as a ‘Hard’ or ‘Deterministic’ System Approach?”, Systems Research and Behavioral Science, Vol. 17, 2000, pp. 3—22, is a discussion of the misinterpretations of system dynamics, and also holds the quote “It may seem paradoxical but the results of a quantitative system dynamics study are qualitative insights” (p. 17).

34

B. Conceptual Foundation for the Usage of System Dynamics

Another interesting criticism of system dynamics, or actually of the earlier modeling approaches, is made by de Geus, who states that most managers have resistance to computer models to a degree where a computer-model becomes a barrier for starting up discussions and exploration of mental models.113 For this reason De Geus proposes a system dynamics modeling approach where soft mapping techniques are used to start up the process capturing mental models. Using soft modeling techniques in the beginning of the modeling activities is also incorporated in many other modeling approaches.114 De Geus furthermore warns against having expert modelers transform the soft models to hard models, as subject matter knowledge is needed in this process.115 This also complies with the arguments for participative modeling, namely, that insights are gained primarily through participation in the modeling itself and that insights are difficult to transfer to others, who were not involved in the modeling process.116 Akkemans furthermore agues, that each type of diagram provides a different and useful view of the problem situation, and that synthesis cannot be automated.117

113

De Geus, Arie P.: The Living Company, Boston, 1997, p. 71. Luna-Reyes, L.F. and D. L. Andersen: “Collecting and analyzing qualitative data for system dynamics: methods and models”, System Dynamics Review, Vol. 19, No. 4, 2003, pp. 271—296 give an overview of many qualitative data collections methods to be used not only in the beginning of a modeling process but also in the later stages. Furthermore, Hodgson, A. M.: “Hexagons for system thinking”, European Journal of Operational Research, Vol. 59, 1992, pp. 123—136, introduces a soft modeling technique that is incorporated in many modeling approaches; e.g. in Group Model Building. 115 De Geus, Arie P.: The Living Company, Boston, 1997, p. 72. 116 See Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993, p. 31; Vennix, Jac A. M.: Group Model Building, Chichester, 1996, pp. 97—99. 117 Akkermans, Henk: Modelling With Managers, Breda, 1995, p. 116. 114

B. Conceptual Foundation for the Usage of System Dynamics

II.

35

Cognitive and Behavioral Rationale for the Usage of System Dynamics

Interventions in social systems are typically discussed in terms of impact on an individual level, group levels (formal and informal groups) and organizational levels.118 On the individual level, learning and change of intentions and behavior are often the focuses of interest. On the group level, alignment of mental models and understanding group dynamics are often seen as corner stones, and on the organizational level, interest often focuses on the creation of a learning organization. This will be addressed in the next three subchapters. 1.

Individual Learning and Change of Behavior in a Complex and Dynamic Environment

Human beings have a tendency to think in events or limited linear causal structures and more often than not, they underestimate or ignore complex dynamic processes (illustrated in figure B-2).119 As a consequence, human decision-makers leave out concerns for side effects and self-reinforcing dynamics and fail to adjust their decision strategies to account for delays in the system and expect feedback to arrive before the system can provide such information.120

118

119

120

These three levels for impacts of interventions are widely used in the system dynamics literature, see Rouwette, Etiënne: Group model building as mutual persuasion, Nijmegen, 2003, pp. 21—27 for a discussion on the three levels in the literature of decision-making. In Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Massachusetts, 1970, p. 38, the three levels are listed together with an additional level; called intergroups (formal and informal). In Dörner, Dietrich: The Logic of Failure, New York, 1996, p. 6 cognitive limitations in analytical, serial and visualized thinking are mentioned (as opposed to female, “parallel” or non-western thinking); In De Bono, Edward: Lateral Thinking for Management, England, 1971, pp. 4—9, it is argued that linear vertical thinking being overly dominant in our education system. In Miller, George A.: “The Magical Number Seven, Plus Minus Two: Some Limits on Our Capacity for Processing Information”, The Psychological Review, Vol. 63, No. 2, March 1956, p. 95, limitations on the amount of information humans are able to receive, process and remember are discussed. Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993, pp. 29—30; Kampmann, Christian P. E.: Feedback

36

B. Conceptual Foundation for the Usage of System Dynamics

Several studies within the field of system dynamics confirm these theories: based on a study investigating a multiplier-accelerator model of capital investments, Sterman concludes that many subjects fail to adequately account for the effects of delays, and fail to understand the feedback between their own decisions and the environment – even when provided with perfect information and knowledge of the system structure.121 In gaming environments, Dörner has observed students and professionals carrying out simulations of a variety of systems (small town, eco-system etc.), with the majority of subjects failing to achieve “good results”; failures he believes are primarily due to lack of system understanding and tendencies to focus on short term, immediate effects rather than more long-term and fundamental processes.122

C

E

F

B

A => E => C => F

A D

System

-

Perception

Figure B-2: Limited linear perception of system

“Mental models” is a repeatedly used term for the cognitive structures of individuals; introduced by Johnson-Laird in the late 1970’s as internal, deducted representations complementing the use of logic, affecting the way humans make

121

122

complexity and market adjustment – An experimental approach, Boston, 1992, p. 31. Sterman, John D.: “Misperceptions of Feedback in Dynamic Decision Making”, in Peter M. Milling and Erich O.K. Zahn (eds.): Computer-Based Management of Complex Systems, Proceedings of the 1989 International Conference of the System Dynamics Society, Heidelberg, 1989, p. 30. Dörner, Dietrich: The Logic of Failure, New York, 1996, p. 18.

B. Conceptual Foundation for the Usage of System Dynamics

37

inferences.123 The underlying ideas and concepts can be traced further back in psychology literature, including the work of Craik’s in the mid 1940’s proposing that people construct internal symbolic representations or models of external events.124 Definitions in the fields of psychology, communication and system dynamics vary from seeing mental models as being stable internal models to being reconstructed on an ad hoc basis; from being simple picture-like images to being complex and intuitive theories; from considering people being aware of their own mental models to being intuitive and inaccessible models.125 Doyle and Ford propose a definition of mental models suitable to the field of system dynamics:

“A mental model of a dynamic system is a relatively enduring and accessible but limited internal, conceptual representation of an external system (historical, existing or projected) whose structure is analogous to the perceived structure of that system.” - Doyle and Ford126

123

124

125

126

Johnson-Laird, P. N.: Mental Models, Cambridge, 1983, pp. 144—145. See also Baron, Jonathan: Thinking and Deciding, 3 rd edition, Cambridge, UK, 2000, p. 74. In Doyle, James K. and David N. Ford: “Mental models concepts for system dynamics research”, System Dynamics Review, Vol. 14, No. 1, Spring 1998, p. 8, reference is made to Craik’s (1943) book “The Nature of Explanation”. This reference is also made in Johnson-Laird, P. N.: Mental Models, Cambridge, 1983, p. 2. Doyle, James K. and David N. Ford: “Mental models concepts for system dynamics research”, System Dynamics Review, Vol. 14, No. 1, Spring 1998, p. 4, p. 9 and p. 14. Doyle, James K. and David N. Ford: “Mental models concepts revisited: some clarifications and a reply to Lane”, System Dynamics Review, Vol. 15, No. 4, Winter 1999, p. 414. The definition is a revised version from an earlier article, based on comments in Lane, David C.: Friendly amendment: A commentary on Doyle and Ford’s proposed re-definition of “mental model”, Systems Dynamics Review, Vol. 15, No. 2, Summer 1999, pp. 185—194.

38

B. Conceptual Foundation for the Usage of System Dynamics

With Doyle and Ford’s definition the system dynamics society to some degree has an aligned use of the term mental models. In line with this definition, cause-loop-diagrams or stock-and-flow diagrams could be viewed as externalized mental models.127 Lane recommends not using the term “cognitive map” for these diagrams, as this is a well-established term in management science literature for the representation used in the SODA methodology.128 Doyle and Ford have a broader definition, and argue that cognitive map is a well suited term, due to its intuitive appeal, and arguing that Eden’s use of the term is merely a particular form of cognitive mapping.129 The conceptual application of mental models gives meaning in both everyday life and in more complex decision-making situations. In a normal discussion of a single topic, participants will implicitly employ different mental models, with different underlying assumptions, and with different goals. Underlying relationships, assumptions and goals are typically left unstated and never brought into the open. With this in mind, Forrester argues that it is easy to understand why compromises often takes long and often fail in their objectives or produce new difficulties.130 A different view could be, that the “slack” in understanding does not make compromising more difficult, but allows everyone to have a different understanding of the compromise. For important and complex decision-making it consequently makes sense to invest in the efforts of making the models explicit. Forrester further argues for not only making mental models explicit, but also using computer analyses to investigate the problem, as even models correct in structure and assumptions due to the fact that limited

127

128

129

130

Lane, David C.: Friendly amendment: “A commentary on Doyle and Ford’s proposed re-definition of ‘mental model’”, Systems Dynamics Review, Vol. 15, No. 2, Summer 1999, p. 186. Lane, David C.: “A commentary on Doyle and Ford’s proposed re-definition of ‘mental model’”, Systems Dynamics Review, Vol. 15, No. 2, Summer 1999, p. 186. Doyle, James K. and David N. Ford: “Mental models concepts revisited: some clarifications and a reply to Lane”, System Dynamics Review, Vol. 14, No. 1, Spring 1998, pp. 412—413. Forrester, Jay W.: “Counterintuitive Behavior of Social Systems”, in Collected Papers of Jay W. Forrester: Foreword by Gordon S. Brown, Cambridge, 1975, p. 213.

B. Conceptual Foundation for the Usage of System Dynamics

39

human brain capacities often yield wrong conclusions (individually or as group consensus).131 An interesting aspect of mental models is that people seek confirmation for their theories rather than questioning them, and as a consequence they are often stuck in severely sub-optimal decision and problem-solving strategies.132 This is an effect of prior beliefs: people tend to find in data what they expect to find.133 A classical study, replicated several times, is the Rosenthal experiment illustrating the self-fulfilling prophecy that is the result of people seeking confirmation of their mental models and in their mind rejecting information contradicting their beliefs.134 It is a double-blind experiment where a large number of teachers, before meeting their new classes, are handed names of those 20% of the students, who based on a certain test would be expected to make above average intellectual progress. In reality the students were picked randomly, but at the end of the school year the randomly picked students showed a real above-average performance increase as well as being rated by the teachers as showing distinguished intellectual curiosity. Another classic example of selffulfilling prophecy is a waiter having a mental model being that well-dressed people tip better. Subconsciously, the waiter will give well-dressed people better service and thereby create a positive loop confirming his or her prior beliefs.

131

132

133

134

Forrester, Jay W.: “Counterintuitive Behavior of Social Systems”, in Collected Papers of Jay W. Forrester, Cambridge, 1975, p. 214. Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993. pp. 29—30. An often mentioned example: Newell, Allen and Herbert A. Simon: Human Problem Solving, Englewood Cliffs, New Jersey, 1972, pp. 90—91, describe the well known ‘Nine Dot Problem’ (with nine dots arranged in a 3 by 3 square array), where each subject is “directed to draw four straight lines, without raising his pencil from the paper, that pass through all nine dots.” Most subjects subconsciously assume, that the lines may not continue outside the boundaries of the square, which make the problem unsolvable. Baron, Jonathan: Thinking and Deciding, 3 rd edition, Cambridge, UK, 2000, p. 182. The Rosenthal experiment was carried out in the 1960’s, and a description can be found in Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 20. Other experiments showing conservatism in mental models include gaming environments where participants have a tendency to stick to their initial understandings of the system even when data would suggest otherwise, see Dörner, Dietrich: The Logic of Failure, New York, 1996, p. 17.

40

B. Conceptual Foundation for the Usage of System Dynamics

Prior beliefs can also result in illusory correlations, e.g. if a teacher thinks a child is intelligent, he or she will often tend to overestimate how well behaved the child is.135 Before discussing how to bring about changes in individuals’ cognitive structures and mental models, it is relevant to take a look at the theories of judgment. Kahneman and Tversky argue, that human judgment relies on heuristics rather than calculus of chance or statistical based prediction.136 Human judgment is affected by a number of elements: contextual effects, the extent to which cues are available, the order in which information is presented, whether comparative judgments involve similar or dissimilar information, qualitative and quantitative data, and other factors.137 Hogarth especially points out the importance of availability of information in the forming of judgments. Individuals tend to put more emphasis on clues available to them, and furthermore, some information from the past has stuck better, and therefore influences judgments to higher degree.138 The contextual effects influencing perceptions include the absolute variation in size, as experiments have shown distorted perceptions due to absolute sizes.139 Another contextual effect shows in the study of scenario planning where scenarios that include both a possible cause and an outcome seems to appear more probable than scenarios merely involving the outcome.140 An interesting hypothesis stated by Enhager, is that the intensity of emotions activated (being ether positive or negative) strongly influences how well humans remember a given situation.141 He argues, that people typically 135

136

137

138

139 140 141

Baron, Jonathan: Thinking and Deciding, 3 rd edition, Cambridge, UK, 2000, p. 183. Kahnemann, Daniel and Amos Tversky: “On the Psychology of Prediction”, Psychological Review, Vol. 80, No. 4, July 1973, p. 237. Hogarth, Robin: Judgment and Choice: The Psychology of Decision, 2 nd edition, Chicago, 1987, p. 55. Hogarth: Judgment and Choice, Chicago, 1987, p. 53. Hogarth mentions work by Tversky and Kahnemann in much of his argumentation. Hogarth: Judgment and Choice, 1987, p. 52. Hogarth: Judgment and Choice, 1987, p. 49. Kjell Enhager is a known speaker on the psychology of sports, and he has used his theories in the training of the female national golf team in Sweden, e.g. by having the women forcing a positive feeling with a smile and a happy exclamation for each good swing and trying to ignore the bad ones.

B. Conceptual Foundation for the Usage of System Dynamics

41

remember especially positive or especially negative situations from their pasts, and by forcing positive feelings by different means, effectiveness in learning situations can dramatically improve. This is in accordance with the line of reasoning for storytelling: that activating feelings make a message memorable.142 Baron discusses dedicated scientific research on how human learning “sticks” from experiences, and is biased by attention factors, such as giving more attention to easily available information or being more or less open-minded.143 With regards to learning from experiences, it seems that positive feedback is weighted more heavily in memory than negative feedback, and furthermore indications exist, that explicit focus on identifying underlying rules lower the attentional bias.144 In handling judgment, anchoring is a commonly used strategy. The judgment process is based on a starting point (the anchor) from where predicted adjustments are made – and often this type of judgment is believed to be the foundation for much intuitive anticipation.145 Even randomly selected anchors that have nothing to do with the judgment situation, or the information available, have shown to influence judgment processes.146 For decision-making under complexity and uncertainty this means anchoring outcome of choices as deviations (losses and gains) from reference points, although often weighting losses larger that gain reflecting a negative attitude towards risk.147 Cognitive learning is often a necessary prerequisite in organizational interventions; although not adequate for creating changes in behavior. Thinking is rooted in the total process of psychic activity, and is linked with emotions,

142

143

144

145

146

147

McKee, Robert: “Storytelling That Moves People”, Harvard Business Review, June 2003, p. 52. Baron, Jonathan: Thinking and Deciding, 3 rd edition, Cambridge, UK, 2000, pp. 176—181. See Hogarth, Robin: Judgment and Choice: The Psychology of Decision, Chicago, 1987, p. 130; Baron, Jonathan: Thinking and Deciding, 3 rd edition, Cambridge, UK, 2000, p. 182. See discussions in Hogarth, Robin: Judgment and Choice, Chicago, p. 54; Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 28. Vennix: Group Model Building, 1996, p. 28, offers a number of examples from psychology literature. Hogarth: Judgment and Choice, 1987, p. 109.

42

B. Conceptual Foundation for the Usage of System Dynamics

values and motivations.148 Ajzen and Fishbein have developed a theory of planned behavior building upon research within attitudes and personality, as well as behavioral, normative and control beliefs (see figure B-3). The framework is based upon their early research implying that changes in attitudes due to persuasive communication is insufficient to produce behavioral changes.149 Consequently, the model includes elements such as group dynamics, social norms, and control beliefs. Attitudes towards behavior include cognitive, affective and conative dimensions, and are determined by behavioral beliefs reflecting the person’s understanding of the likely outcomes given certain behavior, weighted with the person’s individual evaluation of the possible outcomes.150 This tripartite model is also used in the field of consumer behavior, where Kotler discusses a number of different response hierarchy models with regards to not only the elements included, but also how different buying situations call for different order sequences.151

148

149

150

151

Dörner, Dietrich: The Logic of Failure, New York, 196, p. 8. An example of personal motives and perceived behavioral control influencing behavior is given in Berger, Ulrike and Isolde Bernhard-Mehlich: “Die Verhaltenswissenschaftliche Entscheidungstheorie”, in Kieser, Alfred (ed.): Organisationstheorien, 3 rd edition, Stuttgart, 1999, p. 155. A practical example of how to involve emotions in change processes is the use of storytelling, see McKee, Robert: “Storytelling That Moves People”, Harvard Business Review, June 2003, p. 52. Ajzen, Icek and Martin Fishbein: “The Prediction of Behavior from Attitudinal and Normative Variables”, Journal of Experimental Social Psychology, No. 6, 1970, p. 483. See Ajzen, Icek: Attitudes, Personality and Behavior, 1988, Chicago, pp. 20—23, and pp. 120—121. Ajzen is using the three components of attitude (cognitive, affective, and conative), arguing that this tripartite model of attitude has served as the starting point of most behavioral analyses since the 1960’s. See Kotler, Philip: Marketing Management, 7th edition, New Jersey, 1991, pp. 573—575 where different buying situations are described in terms of models embracing cognitive (learning), affective (feeling) and behavior (doing) stages. The AIDA-model (awareness-interest-desire-action) is probably the most known model. Kotler also argues, that high-involvement purchases with perceived high differentiation calls for the “learn-feel-do“ sequence of marketing.

43

B. Conceptual Foundation for the Usage of System Dynamics

Attitude towards the behavior

Subjective norm

Intention

Behavior

Perceived behavioral control

Figure B-3: Theory of planned behavior 152

For research within organizational interventions, understanding the three attitude components (the cognitive, affective, and conative elements) and personality traits are important in establishing the “general laws” of human action, as well as concepts of the dispositional nature of human behavior.153 Although many researchers argue that the evaluations between the three attitude components can differ, Ajzen also discuss that many researchers find it

152

153

Taken from the figure in Ajzen, Icek: Attitudes, Personality and Behavior, Chicago, 1988, p. 133. Ajzen, Icek: Attitudes, Personality and Behavior, Chicago, 1988, p. 46. See also Rao, Abhijit: “Recognition of Conative and Affective Behavior in Web Learning using Digital Gestures”, North America Web-Based Learning Conference, Online Proceedings, New Brunswick, 2001, p. 1, on conative and affective elements improving learning experiences.

44

B. Conceptual Foundation for the Usage of System Dynamics

empirically difficult to distinguish between cognition and affect, as they often are highly correlated.154 The subjective norms are assumed to be a function of the person’s beliefs that specific individuals or groups approve or disapprove of performing this behavior, and deal with both social norms and group dynamics aspects.155 The importance of subjective norms is often stressed since Elton Mayo in 1945, published a set of social assumptions regarding human nature. This focuses on the social need of employees and recognizes the strong influence high status individuals have on peer employees independently of formal power.156 Schein states that norms within a group are maybe the most important influential factors for the everyday behavior of the employees in a company.157 A related view on social norms is what Gladwell calls the Power of Context: how seemingly small changes in context can influence behavior. A known example is the Broken Window theory, proposing that if small context flaws are tolerated (e.g. broken windows), soon larger context flaws will occur (e.g. more windows will be broken, which will start a process towards more serous crimes).158 Together the attitudes towards behavior and the subjective norms influence intended behavior in a way that can be considered as goals to be pursued.159 The intention to pursue these goals are affected by the person’s perceived control over a given behavior determined by internal factors including information, skills, abilities and control of emotions and compulsion, and external factors like opportunity and dependence on others.160 The broken arrow

154 155

156 157 158

159 160

See Ajzen, Icek: Attitudes, Personality and Behavior, 1988, p. 21. Ajzen, Icek: Attitudes, Personality and Behavior, 1988, p. 121; Ajzen, Icek and Martin Fishbein: “The Prediction of Behavior from Attitudinal and Normative Variables”, Journal of Experimental Social Psychology, No. 6, 1970, p. 483. See Schein, Edgar H.: Organisationspsykology, Herning, 1990, p. 67. Schein, Edgar H.: Organisationspsykology, 1990, p. 37. See Gladwell, Malcolm: The Tipping Point, paperback edition, New York, 2002, p. 141, giving credit of the Broken Window theory to the criminologists James Q. Wilson and George Kelling. At pp. 142—145 an example of interventions partly based on this theory is described: the fight against crime in New York in the 1990’s, where fight against graffiti and fare-beating in the subway was the starting point Ajzen: Attitudes, Personality and Behavior, Chicago, 1988, p. 128. See Ajzen: Attitudes, Personality and Behavior, 1988, pp. 128—131.

B. Conceptual Foundation for the Usage of System Dynamics

45

in figure B-3 indicates that the influence will only emerge when there is some agreement between a person’s actual control and the person’s perceived control of the behavior.161 Furthermore, perceived control does not only influence a person’s current behavior. It also has long-term perspectives, as there is a general agreement that perceived control results in positive feelings of one’s own competence and worth, and likewise that lack of or loss of perceived control influences both the physical and psychological well-being negatively.162 Understanding mental models, judgment processes, and behavior theories is essential in understanding how people learn and react. People learn primarily on the basis of what they can observe.163 A fundamental attribute of system dynamics offers the opportunity to model and simulate the behavior of dynamic structures and through observing the behavior reaching an understanding of a complex problems in a few days or weeks that would have taken years to understand by “normal” experience. Simulations have shown to result in better transfer than case studies and lectures and this may also be attributed to motivational side-effects of the interactive pedagogy.164 It is often difficult to change the truly fundamental assumptions and beliefs underlying one’s thoughts and actions, and especially when they are vague and poorly understood are they difficult to escape from.165 People are typically unaware of their own inconsistencies leading to self-fulfilling prophecies, and self-sealing processes.166 Modeling and simulating system structures is a way to bring forward underlying

161 162

163

164

165

166

Ajzen, Icek: Attitudes, Personality and Behavior, 1988, p. 134. See Bungard, Walter: “Zur Implementierungsproblematik bei BusinessReengineering Projekten”, in Perlitz, Manfred, Andreas Offinger, Michael Reinhardt and Klaus Schug (eds.): Reengineering zwischen Anspruch und Wirklichkeit, Wiesbaden, 1996, pp. 260—261. Also Kieser, Alfred: Organisationstheorien, 3 rd edition, Stuttgart, 1999, pp. 129—131, describes “Humanisierung der Arbeit” and discusses the benefits of the individual’s influencing own work situation. Hogarth, Robin: Judgment and Choice – The Psychology of Decision, Chicago, 1987, p. 130. Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993, p. 31. Morgan, Gareth: Creative Organization Theory, Newbury Park, California, 1989, p. 28. See Argyris, Chris: Reasoning, Learning, and Action – Individual and Organizational, San Francisco, 1982, p. 39.

46

B. Conceptual Foundation for the Usage of System Dynamics

assumptions and inconsistencies, being the first step in a process to changing them.167 Mental models are the “product” that “modelers take from students and clients, disassemble, reconfigure, add to, subtract from, and return with value added”.168

2.

Establishing Group Consensus by Sharing Mental Models

Development of group consensus and individual learning are strongly interrelated processes. Alignment between individuals in a group necessarily requires individual learning, and likewise individual learning is affected by the context of a group setting. A decision regarding individual goals made by an individual in a group setting seems to have a significantly more enduring behavioral effect compared to settings with 1-on-1 lecturing.169 Furthermore, it is easier to change the ideology and social practice of a small group handled together than of single

167

168

169

In chapter A, learning was discussed in terms of single-loop and double-loop learning. The discussions on double-loop learning and mental model refinement are closely related. Both types of learning require impact of belief systems and value systems, with beliefs being an understanding of causality and values being a network where one value is supported by the other values. The belief system especially impacts the internal process of “making sense” and defining the situation, whereas the value system subsequent impacts the problem definition, see Eden, Colin: “Cognitive mapping and problem structuring for system dynamics model building”, System Dynamics Review, Vol. 10, Nos. 2-3, Summer-Fall 1994, p. 263. Doyle, James K. and David N. Ford: “Mental models concepts revisited: some clarifications and a reply to Lane”, System Dynamics Review, Vol. 14, No. 1, Spring 1998, p. 4. Lewin, Kurt: “Group Decision and Social Change” (first published in Newcomb and Hartley’s Readings in social psychology, 1948, pp. 330—341), in Gold, Martin: The Complete Social Scientist – A Kurt Lewin Reader, Washington, 1999, pp. 276—279. Note that Lewin has a different terminology use compared to most SD research, as he calls individual decision made in a group setting for “group decisions“ (see p. 274).

47

B. Conceptual Foundation for the Usage of System Dynamics

individuals partly due to unwillingness of the individual to depart too much from group standard of the group they belong to or wish to belong to.170 In organizational settings, where teams work together to make the business run, and where decisions often are made in teams involving the main stakeholders, the concept of mental model offers insight into many of the typical problem solving challenges. Each individual has a limited linear understanding of the main problems, and these will differ with regards to which causal relationships are perceived as the most influential ones (see figure B-4).

A => E => C => F

C

E

F B

A => E => C => -D

A D

System

A => B => D

-

Perception

Figure B-4: Different limited linear perceptions of a system

The differing mental models among individuals in team settings often result in reduced communication effectiveness, as the mental models reflect different perceptions of the topics being discussed, typically without these differences are being investigated. The diverse mental models or system perceptions partly explain some of the classic conflicts in organizations. An example being conflicts between the production and the marketing department, 170

Lewin, Kurt: “Group Decision and Social Change”, in Gold, Martin: The Complete Social Scientist – A Kurt Lewin Reader, Washington, 1999, p. 273 and p. 281.

48

B. Conceptual Foundation for the Usage of System Dynamics

although not only differing mental models, but also conflicting interests based on conflicting goal-settings play a role. Baron discusses the concept of cognitive dissonance resolution, as a way humans eliminate conflicts among beliefs.171 An example is, that after a decision is made, it seems that increased weight is put on the reasons in favor of the decision, and less value is given to the arguments in favor for paths that were not selected. In organizational settings, where decision-makers typically have had a significant role in the forming of the present situation, and have often had different viewpoints in earlier decision-making processes, cognitive dissonance resolution must be expected to contribute to the creation of biases in their individual mental models. Janis has put forward the concept of Groupthink, which is partly related to cognitive dissonance resolution on a group level, and the concept is based on extensive case studies on group decision situations, including major US failures such as the Bay of Pigs decision, the escalation of the Vietnam War and the failure to predict the attack on Pearl Harbor.172 The Groupthink phenomenon refers to the lack of critical thinking that can occur in groups characterized by group cohesiveness and consensus. It is in line with social-psychological research findings, showing powerful social pressures if a dissident in a cohesive group reveals objections to group norms or consensus. Although the term Groupthink primarily refers to self-censorship in the critical thought process to avoid group disunity. Dörner discuss the Chernobyl explosion in terms of Groupthink and overconfidence, arguing that the team responsible for the accident felt they were so experienced that they did not have to follow safety rules.173 Although the team was in fact very experienced they made basic errors of interaction with dynamic

171

172

173

Baron, Jonathan: Thinking and Deciding, 3 rd edition, Cambridge, UK, 2000, p. 208. Leavitt, Harold J.: Top Down – Why Hierarchies Are Here to Stay and How to Manage Them More Effectively, Boston, 2005, p. 130, discuss the same concept, stating that “we humans don’t just do what we believe. We also believe what we do.” Janis, Irving L.: “Groupthink: The Problems of Conformity” (original printed in Psychology Today, Nov. 1971, pp. 271—279), in Morgan, Gareth: Creative Organization Theory, Newbury Park, California, 1989, pp. 224—228. Dörner, Dietrich: The Logic of Failure, New York, 1996, p. 34.

B. Conceptual Foundation for the Usage of System Dynamics

49

systems; they over-steered (regulation of the situation rather than the process), they ignored side effects, and they interpreted system indicators erroneously.174 Groupthink is normally discussed is the context of smaller groups, often being management teams, but some of the same mechanisms also apply to a people’s philosophy of life as expressed in the below quote of Nietzsche:

“What is truth? A moving army of metaphors, metonymies and anthropomorphisms, in short a summa of human relationships that are being poetically and rhetorically sublimated, transposed, and beautified until after long and repeated use, a people considers them as solid, canonical, and unavoidable.” - Nietzsche175

Methods for overcoming Groupthink include establishing teams with cognitive diversity, ensuring that teams to embrace different interests and viewpoints in the organization, as well as applying different facilitation approaches in decision-making processes.176 Examples of facilitation approaches include the concept of “Think-hats” (where group members play different roles during the meetings, including playing roles like out-of-the-box-thinking and being the-devils-advocate), and Beer’s concept of Team Syntegrity. Team

174

175

176

Dörner, Dietrich: The Logic of Failure, New York, 1996, pp. 30—33; Salge, Markus and Peter Milling: “Who is to blame, the operator or the designer? Two stages of human failure in the Chernobyl accident”, System Dynamics Review, Vol. 22, in print, 2006. Quoted in Morgan, Gareth: Creative Organization Theory, Newbury Park, California, 1989, p. 22. For a definition on cognitive diversity, see Tilebein, Meike: “Eine strukturwissenschaftliche Betrachtung von Diversity Management”, Tagungsband, GWS-Tagung, Greifswald, in print, 2006, p. 1.

50

B. Conceptual Foundation for the Usage of System Dynamics

Syntegrity is a process to establish shared meaning of a problem universe as well as to create shared action plans among a larger group of people.177 This process includes giving the participants the roles of critics and observers in different team settings.178 Team effectiveness problems, which might be more frequently seen than Groupthink, are the problems originating from teams characterized by politics and differing perceptions. Especially if the problem reflects an unpleasant situation, where the company performs unsatisfying, a normal human reaction is to mistrust and blame others.179 Participative model building can be an approach for overcoming Groupthink as well as for handling politics or differing perceptions. Through a group model building process, mental models can be challenged and the process can be seen as a vehicle for facilitating negotiation or mutual persuasion.180 Figure B-5 is a simplified illustration of how mental models influence formal models, and how the formal models in turn also influence the mental models. This is the loop where a shared formal model through an iterative model-building process will facilitate an alignment between individual mental models. The new mental model will result in new behavior and therefore indirectly affect the system, and more directly the new mental models might result in new policies and in this way result in system changes. Meyer offers a somewhat similar, but more elaborated model, where the outcome of the process of formulating and

177

178

179

180

Pearson, Alan: “You Drive for the Show but you Putt for the Dough”, appendix in Beer, Stafford: Beyond Dispute – The Invention of Team Syntegrity, 1994, p. 321. See Beer, Stafford: Beyond Dispute – The Invention of Team Syntegrity, 1994, p. 59 and p. 102. Kanter, Rosabeth Moss: “Leadership and the Psychology of Turnarounds“, Harvard Business Review, June 2003, p. 61. Eden, Colin: “Cognitive mapping and problem structuring for system dynamics model building”, System Dynamics Review, Vol. 10, Nos. 2-3, Summer-Fall 1994, p. 259; Rouwette, Etiënne: Group model building as mutual persuasion, Nijmegen, 2003, p. 251. Also the alignment of mental models has some similarities with building collective intuition, see Eisenhardt, Kathleen M.: “Strategy as Strategic Decision Making”, Sloan Management Review, Spring 1999, pp. 66—67.

51

B. Conceptual Foundation for the Usage of System Dynamics

simulating formal models results in selection and implementation of new policies, which is how the feed-back is primarily given back to the actual system.181

Actual System

Mental Model

Formal Model

Figure B-5: Mental models as instruments between actual systems and formal models

3.

Enhancing Organizational Learning through System Thinking Experience and Double-Loop Learning

System modeling and simulation projects in organizations not only address specific problems, but also influence ongoing organizational learning efforts, since the individual learning and improvement and alignment of mental models described in the two previous subchapters are important instruments in organizational learning.182 The term organizational learning has been subject to a large variety of definitions in the literature. A few examples from these definitions include the understanding of organizational learning as the sum of individual learning, the understanding that it takes place, when learning is accumulated into rules or routines that guide decisions and behavior, and also

181

182

Maier, Frank: Die Integration wissens- und modellbasierter Konzepte zur Entscheidungsunterstützung im Innovationsmanagement, Berlin, 1995, p. 217. In Kim, Daniel H. and Peter M. Senge: “Putting systems thinking into practice”, System Dynamics Review, Vol. 10, Nos. 2-4, Summer-Fall 1997, p. 279, the concept of mental models is called the transfer mechanism between individual learning and organizational learning. In Senge, Peter M.: The Fifth Discipline, New York, 1994, mental models and system thinking are together with team learning, shared vision and personal mastery, described as the main elements in organizational learning.

52

B. Conceptual Foundation for the Usage of System Dynamics

include the understanding that organizational learning is a fundamentally different process than individual learning.183 In the context of this dissertation, and in accordance with Senge and Sterman, organizational learning is understood as the improvement of system thinking skills necessary to challenge future mental models and biases, opposed to only learning to solve a specific problem.184 This understanding embrace the concept of Double-Loop Learning, proposed by Argyris and Schön, focusing on the difficult learning issues that cannot be solved unless underlying individual and organizational values and assumptions are reexamined.185 Organizational learning deals with the ability to learn and on a continuing basis to become better at double-loop learning.186 For organizational interventions solving business problems or improving business processes, organizational learning only takes place if the underlying problem solving capability of the organization is not ignored.187 This is also in agreement with the process consultation view of Schein, focusing on the organization’s general ability to solve problems, rather than on just solving a present problem.188 Taking this view on organizational learning in SD projects means that not only should a modeling study focus on exploring the problem and suitable solutions, it should also increase the participants ability to think in terms of systems and causal loops in later problem solving situations. 183

184

185

186

187

188

For literature overview and discussions of organizational learning, see Argyris, Chris: On Organisational Learning, 2 nd edition, Oxford, 1999, pp. 7—14; and Kieser, Alfred and Ulrich Koch: Organizational Learning through Rule Adaptation: From the Behavioral Theory to Transactive Organizational Learning, Mannheim, 2000, pp. 2—26. Senge, Peter M.: The Fifth Discipline, New York, 1994, p. 69; Sterman, John D.: “All models are wrong: reflections on becoming a systems scientist”, System Dynamics Review, Vol. 18, No. 4, Winter 2002, p. 526. Argyris, Chris and Donald Schön: “Organizational Learning: A theory of Action Perspective” (first printed as part of book with the same title in 1978), reprint in Morgan, Gareth: Creative Organization Theory, Newbury Park, California, 1989, p. 140; Argyris, Chris: Reasoning, Learning, and Action – Individual and Organizational, San Francisco, 1982, p. 160. Argyris, Chris and Donald Schön: “Organizational Learning: A theory of Action Perspective”, reprint in Morgan, Gareth: Creative Organization Theory, Newbury Park, Ca., 1989, p. 142. Argyris, Chris: On Organisational Learning, 2nd edition, Oxford, 1999, pp. 230—238. Schein, Edgar H.: Process Consultation, Boston, 2000, part I, p. 194.

B. Conceptual Foundation for the Usage of System Dynamics

53

Organizational learning can also be discussed in terms of company business planning. De Geus is known for the quote: “At Shell, planning means changing minds, not making plans.” 189 He describes how the normal decisionmaking processes in organizations can be viewed as learning processes, where mental models are improved and aligned through the ongoing dialogue, and where he perceives one of the main problems to be the speed of learning of such groping processes. De Geus furthermore recommends striving for acceleration of the institutional learning through using business planning as a tool to change the mental models of the decision makers, recognizing the mutual influence between planning and mental models. Based on cybernetic tradition, Milling offers a conceptual illustration of organizational learning (figure B-6) as a number of interrelated second-order individual learning processes, combining a feedback-loop based on observations of actual system behavior (adaptive learning), and a feedback-loop representing individuals modification of mental models (double-loop-learning).190 Due to the interaction between the individual learning cycles, the organizational learning is more than a mere summation of individual learning processes. Kim and Senge have proposed a somewhat similar illustration, of what they call the Organizational Learning Cycle, where changes of mental models also take place as second-order learning (double-loop learning). 191

189

190

191

De Geus Aire P: “Planning as Learning”, Harvard Business Review, MarchApril 1988, p. 70. Milling, Peter: “Organisationales Lernen und seine Unterstützung durch Managementsimulatoren”, in: Zeitschrift fåür Bestriebswirtschaft, 65 Jg. Lernende Unternehmen (Sonderausgabe), 1995, pp. 98—100. Kim, Daniel H. and Peter M. Senge: “Putting systems thinking into practice”, System Dynamics Review, Vol. 10, Nos. 2-4, Summer-Fall 1997, pp. 280—281. This framework is more comprehensive, and includes a number of different types of learning, but is as a consequence less intuitively understandable compared to the framework of Milling.

54

B. Conceptual Foundation for the Usage of System Dynamics

Observed Behavior

Comparison Comparison Statements Comparison Expectations

Comparison Comparison Comparison Comparison

Comparison Comparison Mental Comparison Models

Real World

Comparison Comparison Comparison Reflection

Decision Action

Figure B-6: The basic structure of organizational learning 192

This framework more explicitly operates with an organization’s shared mental model, and when changes in individual mental models result in changes in the organization’s shared mental models, it is called Organizational Double-Loop Learning. Due to the significant time-delays in learning based on behavior of real systems and the often seen breakdowns in the learning cycle involving real life actions, Kim and Senge argue for the importance of Managerial Practice Fields. Managerial Practice Fields are settings where teams who need to take action together can learn together and benefit from virtual worlds allowing in a short time to learn long-term systemic consequences of decisions (without real-life consequences of mistakes) and as a by-product improve their learning

192

Own translation of figure in Milling, Peter: “Organisationales Lernen und seine Unterstützung durch Managementsimulatoren”, in: Zeitschrift für Betriebswirtschaft, Ergänzungsheft 3/95, Lernende Unternehmen (Sonderausgabe), 1995, p. 100.

B. Conceptual Foundation for the Usage of System Dynamics

55

capabilities.193 Examples of virtual worlds or simulation games used in business context are Senge and Sterman’s Claims Learning Lab, Repenning and Sterman’s Maintenance Game and the often-described Beer Game.194 The opportunity to significantly speed up organizational learning through the usage of management simulators is illustrated by a “short-cut” in the system behavior feedback-loop in Milling’s framework (figure B-7). The arrows from Mental Models to Formal Models reflect that a formal model should represent a transformation of the mental models of the decision-makers. If the figure should illustrate a participative model building process, rather than a management simulator, one could argue that the arrows should go both ways due to the challenge and alignment of mental models seen in the modeling processes itself. This type of learning would be complementary to the learning achieved by simulating the model.

193

194

Kim, Daniel H. and Peter M. Senge: “Putting systems thinking into practice”, System Dynamics Review, Vol. 10, Nos. 2-4, Summer-Fall 1997, pp. 278—279 and p. 286. Senge, Peter M. and John D. Sterman: “Systems thinking and organizational learning: Acting locally and thinking globally in the organization of the future”, European Journal of Operational Research, Vol. 59, No. 1, 1992, pp. 146—148; Repenning, Nelson P., Sterman, John D.: “Nobody Ever Gets Credit for Fixing Problems that Never Happened: Creating and Sustaining Process Improvement”, California Management Review, Vol. 43, No. 4, Summer 2001, pp. 64—88; and Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, pp. 130—132.

56

B. Conceptual Foundation for the Usage of System Dynamics

Observed Behavior

Comparison Comparison Statements Comparison Expectations

Comparison Comparison Comparison Comparison

Comparison Comparison Mental Comparison Models

Formal Model

Real World

Comparison Comparison Comparison Reflection

Decision Action

Figure B-7: Formal models supporting organizational learning 195

III.

The Development Process of System Dynamics Models in Corporations

Luna-Reyes and Andersen offer a discussion of the system dynamics modeling process across five selected representatives of the classic system dynamics literature.196 The described processes all divide the modeling process into a number of iterative phases, varying from three to seven phases. No dedicated

195

196

Own translation of figure in Milling, Peter: “Organisationales Lernen und seine Unterstützung durch Managementsimulatoren”, in: Zeitschrift für Betriebswirtschaft, Ergänzungsheft 3/95, Lernende Unternehmen (Sonderausgabe), 1995, p. 105. In Luna-Reyes, Luis Felipe and Deborah Lines Andersen: “Collecting and analysing qualitative data for system dynamics: methods and models”, System Dynamics Review, Vol. 19, No. 4, 2003, pp. 274—279, primarily based on the following mentioned sources: Randers, Richardson and Pugh, Roberts et al., Wolstenholme and Sterman.

B. Conceptual Foundation for the Usage of System Dynamics

57

participative model-building approach is represented in the Luna-Reyes and Andersen overview, but such descriptions can be found in the work of Vennix, Andersen, Richardson and Rohrbaugh.197 Among other additional interesting modeling process descriptions, Forrester offers a generic process with six steps from problem symptoms to improvement.198 The following subchapter, Decompositions and Iterations in Model Development and Use, consists of a general discussion of the modeling process used in the context of organizational interventions, and is divided into three main sections: a) Problem Definition and System Conceptualization b) Model Formulation and Testing c) Policy Formulation and Implementation The second subchapter, Designing System Dynamics Modeling-Based Interventions, primarily consists of discussions on the design of experimentationbased learning cycles, knowledge acquisition, and intervention design considerations with special focus on participative modeling processes. The practical application and the special requirements for the development process of management simulators, micro-worlds and simulation-based games will not be discussed in this dissertation, due to its focus on participative modeling processes.199

197

198

199

See Vennix, Jac A. M.: Group Model Building, Chichester, 1996, chapter 2 and 3; and Vennix, Jac A. M; David F. Andersen; George P. Richardson and John Rohrbaugh: “Model-building for group decision support: Issues and alternatives in knowledge elicitation”, European Journal of Operational Research, Vol. 59, 1992, pp. 28—41. Forrester, Jay W.: “System dynamics, system thinking, and soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, p. 244. For a description and discussion of management simulators, see Größler, Andreas: Entwicklungsprozess und Evaluation von Unternehmenssimulation für lernende Unternehmen, Frankfurt am Main, 2000.

58

B. Conceptual Foundation for the Usage of System Dynamics

1.

Decompositions and Iterations in Model Development

Modeling is an iterative process embedded in the dynamics of the system.200 The establishment of a modeling project in itself is a new part and influencer of the real system. Although the process naturally needs to start at a point and address first things first, any step can yield insights making adjustments to an earlier step necessary. Model revisions usually continue almost to the end of the project, and even in implementing results or following up on results, insights can occur changing the very foundation for the process with a new focus of the problem.201 The path towards a useful model is always to some degree a process of trial and error, and the goal of effective procedures for model construction is to achieve reasonable consistency rather than to eliminate all iterations.202 For practical reasons, all descriptions of modeling processes group tasks of the modeling process into modeling stages or phases.

a.

Problem Definition and System Conceptualization

Defining the problem and conceptualizing the system is not only the most critical part of a modeling process; it is often also the most difficult one.203 The word problem is used without attachment of value; it might as well reflect a positive opportunity as a negative problem. Problem-orientation is central in most system dynamics literature, and the importance of modeling a problem, as opposed to a system, is often stressed as a mean to develop focused, purposeful and relevant

200

201

202

203

See Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, pp. 87—88. Homer, Jack B.: “Why we iterate: scientific modeling in theory and practice”, System Dynamics Review, Vol. 12, No. 1, Spring 1996, p. 16. Randers, Jørgen: “Guidelines for Model Conceptualization”, in Randers, Jørgen (ed.): Elements of the System Dynamics Method, Cambridge, Connecticut, 1980, p. 118. In Forrester, Jay W.: “System dynamics, system thinking, and soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, p. 253, it is stated that to describe the system is “the most important and the least straightforward of the stages in system improvement;” and Schein, Edgar H.: Process Consultation – MAOM Capstone Course for the University of Phoenix, Boston, 2000, part I, p. 62 it says: “In my own experience in solving problems and watching others solve them, by far the most difficult step is the first one – defining the problem.”

B. Conceptual Foundation for the Usage of System Dynamics

59

models.204 Different people will have different views on the importance of competing problems, and context and relevance are related to whom the modeling is for.205 Problem focus is a way to think in needed organizational changes, and it is linked to creating implementation orientation from the very outset of the project, as well as acknowledging that the project is a part of the ongoing organizational learning process of addressing problems in the organization.206 Eden argues that the most common defaults to proper problem definition arrive from not seeing or ignoring the complexity, or from the politics of “the truth” being defined by a powerful or otherwise dominant person or group.207 To characterize the problem dynamically and to avoid a short-term eventoriented worldview it is often recommended to establish a reference mode, with graphs or data illustrating how the problem has developed in the past (and how it might evolve further on).208 A reference mode will often be looked back at

204

205

206

207

208

Richardson, George P. and Alexander L. Pugh: Introduction to System Dynamics Modeling with DYNAMO, Cambridge, 1981, p. 18, Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, p. 89. An exception from the problem-orientated viewpoint is the Strategy Dynamics approach, where a company’s strategic architecture is modeled rather than a problem, see Warren, Kim: Competitive Strategy Dynamics, Chichester, 2002, pp. 89—113. Eden, Colin: “Cognitive mapping and problem structuring for system dynamics model building”, System Dynamics Review, Vol. 10, Nos. 2-3, Summer-Fall 1994, p. 261. Roberts, Edward B.: “Strategies for Effective Implementation of Complex Corporate Models”, in Edward B. Roberts (ed.): Managerial Applications of System Dynamics, Cambridge, Massachusetts, 1978, pp. 83—89; Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, pp. 87—88. Eden, Colin: “Cognitive mapping and problem structuring for system dynamics model building”, System Dynamics Review, Vol. 10, Nos. 2-3, Summer-Fall 1994, pp. 261—262. Eden furthermore argues for the usage of cognitive mapping in the problem definition phase, to overcome the problems of handling complexity and politics (pp. 263—268). Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, p. 90; Randers, Jørgen: “Guidelines for Model Conceptualization”, in Randers, Jørgen (ed.): Elements of the System Dynamics Method, Cambridge, Connecticut, 1980, pp. 121—122.

60

B. Conceptual Foundation for the Usage of System Dynamics

(referred to) in many stages of the project, and timescale should be appropriate to the time frame where significant changes and feedbacks are likely to occur.209 The literature in general furthermore recommends formulating a dynamic hypothesis characterizing the problem in terms of underlying feedback, stock and flow structures of the system.210 The dynamic hypothesis guides modeling efforts by focusing on certain system structures. In the system conceptualization, the first explicit structures of the model will typically take form. Decisions will be made with regards to what parameters and casual relations to include in the model, focus on feedback loops etc.211 In general, the system dynamics literature agree on the value of softer techniques in problem identification, whereas a vital discussion exists with regards to the usage of softer techniques in the starting point of the conceptualization and model formulation. The discussions especially focus on the advantages of causal-loopdiagrams vs. stock-and flow diagrams in the search for identifying the most important structures of the system. Sterman, Richardson, Pugh and de Geus often recommend the combination of use of quantitative modeling and qualitative modeling, and having the qualitative modeling serve the purpose of providing good understanding of the problem, it’s main parameters, and it’s boundaries before actually outlining the stock and flow structures in the quantitative model.212 The discussion on whether to include soft modeling techniques in the

209 210

211

212

Warren, Kim: Competitive Strategy Dynamics, Chichester, 2002, p. 27. Forrester, Jay W.: “System dynamics, system thinking, and soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, p. 246; Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, p. 95; Luna-Reyes, L.F. and D. L. Andersen, “Collecting and analyzing qualitative data for system dynamics: methods and models”, System Dynamics Review, Vol. 19, No. 4, 2003, p. 275. In Sterman, John D.: Business Dynamics, Boston, 2000, p. 97, it is recommended to do a formal model boundary chart explicitly grouping endogenous, exogenous and excluded parameters. Sterman, John D.: Business Dynamics, Boston, 2000, p. 102; Richardson, George P. and Alexander L. Pugh: Introduction to System Dynamics Modeling with DYNAMO, Cambridge, 1981, pp. 25—26; de Geus, Arie P.: The Living Company, Boston, 1997, pp. 70—73; Luna-Reyes, L.F. and D. L. Andersen: “Collecting and analyzing qualitative data for system dynamics: methods and models”, System Dynamics Review, Vol. 19, No. 4, 2003, pp. 271—296; Hodgson, A. M.: “Hexagons for system thinking”, European Journal of Operational Research, Vol. 59, 1992, pp. 123—136.

B. Conceptual Foundation for the Usage of System Dynamics

61

system conceptualization in system dynamics projects centers around how to ensure the best quality of the conceptualization; if a softer discussion will allow for more holistic and creative thinking, or if the more structured stock and flow approach will ensure more rigidity and help to not ignore important parts. Forrester represents the latter, recommending using stocks and flows as a starting point.213 Warren, with his Strategy Dynamics approach, also stresses the advantages of a direct mathematical approach, using fundamental resources of a company as starting point for identifying the stock-flow structure of the model.214 This way, the resources and “harder” discussion around reference modes starts off the modeling process, opposed to the use of softer techniques and models. Although it should be mentioned that another resource-oriented approach, Wolstenholme’s Step-wise Approach proposes first to identify the major feedback loops as a starting point for identification of the major resources, and thereby the core stock-flow structure of the model.215 Vennix argued in a discussion at a course in Group Model Building, that if a modeling project is going to include a simulation model, the soft modeling efforts at the beginning of the projects should be of only limited character. If significant time were invested in building a soft model, e.g. a larger Cause Loop Diagram model, he would seldom recommend a later formulation of a quantitative model.216 This might also be a natural consequence of the significant time pressure most organizations experience. If an organization invests significant time in softer modeling exercises, the decision-makers might perceive that the 80% insights have been identified, and consequently that further time to be spent in quantitative modeling would not being worthwhile.

213

214 215

216

Forrester, Jay W.: “System dynamics, system thinking, and soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, p. 252. Warren, Kim: Competitive Strategy Dynamics, Chichester, 2002. Wolstenholme, Eric F.: “The definition and application of a stepwise approach to model conceptualisation and analysis”, European Journal of Operational Research, Vol. 59, 1992, p. 128. Wolstenholme describes the approach as a combination between a feedback loop approach to model construction and a modular approach to model construction (p. 136). Jac Vennix, Professor at Nijmegen University, in a Course in Group Model Buillding, Nijmegen, May, 2004.

62

B. Conceptual Foundation for the Usage of System Dynamics

The problem identification and system conceptualization phase does not only structure and focus the intervention. Through the first analysis of the problem and the microstructure, important insights are likely to occur.217

b.

Model Formulation and Testing

The concept of stocks and flows are the basic building blocks of system dynamics models. In 1968, Forrester formulated the basic principles for model constructions, which are the foundation for most system dynamics modeling work.218 Sterman offers in his comprehensive text book Business Dynamics a thorough discussion and detailed explanations on how to construct, test, and analyze models based on stock and flow structures, feedback-loops, delays, decision-points, parameter-estimations etc.219 A basic understanding behind the usage of system dynamics is that system behavior is a consequence of the system structure. Richardson and Pugh emphasize the importance of having the model contain all significant components relevant to the problem, in order to have the model generate the problem behavior.220 This way the model offers endogenous explanations for system behavior, as opposed to exogenous explanations, where external events cause the behavior.221 Recognizing that a model is only a limited reflection of a real system, the art is to identify the smallest possible set of truly significant components necessary for the predictive ability of the model.222

217

218 219

220

221

222

Flood, Robert L. and Michael C. Jackson: Creative Problem Solving – Total Systems Intervention, Chichester, 1991, p. 74. Forrester, Jay W.: Principles of Systems, Cambridge, 1968. Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000. Richardson, George P. and Alexander L. Pugh: Introduction to System Dynamics Modeling with DYNAMO, Cambridge, 1981, p. 63. Sterman, John D.: Business Dynamics, Boston, 2000, p. 95, argues: “system dynamic models seek endogenous explanations for phenomena.” Luna-Reyes, L.F. and D. L. Andersen, “Collecting and analyzing qualitative data for system dynamics: methods and models”, System Dynamics Review, Vol. 19, No. 4, 2003, p. 285.

B. Conceptual Foundation for the Usage of System Dynamics

63

Striving for high quality and efficiency in model formulation, the field of system dynamics has dedicated much effort to the development of generic models and model elements, which can serve as a starting point in a model building process. Sterman offers a significant number of ‘example-models’ in his extensive textbook, and the literature often discusses generic models, archetypes, as well as microstructures to be used as building blocks in model formulation.223 In the field of group model building, Anderson and Richardson discuss scripts supporting the entire modeling process.224 An important step in the modeling process, and closely linked to the model formulation, is the model testing. Forrester and Senge identify three categories of tests: tests of model structure, tests of model behavior, and tests of policy implications.225 Tests of model structure include structure and parameter verification (in comparison with knowledge of the real system), tests of extreme situations, boundary-adequacy tests, and test of dimensions.226 Tests of model behavior include behavior reproduction based on historical data, behavior prediction tests, behavior-anomaly tests (study of anomaly features of the model behavior), family-member test (comparisons to related models), surprisebehavior tests (behavior which has gone unrecognized in the real system),

223

224

225

226

Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000; Wolstenholme, Eric: “Using generic system archetypes to support thinking and modelling”, System Dynamics Review, Vol. 20, No. 4, Winter 2004, pp. 341—356. Andersen, David F and George P. Richardson: “Scripts for group model building”, System Dynamics Review, Vol. 13, No. 2, 1997, pp. 107—129. Forrester, Jay W. and Peter Senge: “Tests for Building Confidence in System Dynamics Models”, in Legasto, Augusto A., Jay W. Forrester und James M. Lyneis (eds.), TIMS Studies in the Management Sciences, Vol. 14, Amsterdam, 1980, pp. 211—226. A discussion of many of these tests (although differently labeled), as well as additional tests, can be found following a schematically overview of formal model validation tests in Barlas, Yaman: “Formal aspects of model validity and validation in system dynamics”, System Dynamics Review, Vol. 12, No. 3, 1996, p. 189. Also Sterman, John D.: Business Dynamics, Boston, 2000, pp. 845—901, offers an extensive discussion on validation of models. Forrester, Jay W. and Peter Senge: “Tests for Building Confidence in System Dynamics Models”, in Legasto, Augusto A., Jay W. Forrester und James M. Lyneis (eds.), TIMS Studies in the Management Sciences, Vol. 14, Amsterdam, 1980, pp. 211—216.

64

B. Conceptual Foundation for the Usage of System Dynamics

extreme policy tests, as well as behavior tests for boundary-adequacy and sensitivity.227 The last category, tests of policy implications, focus on attempts to verify that response of a policy change corresponds between the real system and the model, and to examine the robustness of policy implications. The policy implication tests include system improvement tests (identifying policies leading to improvements), changed-behavior-predictions tests, and tests for the boundary-adequacy and sensitivity of examined policies.228 Model testing is a part of the gradual process of establishing confidence in the soundness and usefulness of a system dynamics model.229 The entire modeling process seeks to establish trust in the model, as trust is prerequisite to gaining learning from the model and establishing confidence in the policy implications derived from the model. Barlas furthermore discusses validity of a model as usefulness with respect to purpose, which also involves a discussion of the validity of the purpose itself.230 When discussing model adequacy and achieving confidence in the model, Forrester acknowledges that a model is a compromise between adequacy and time and costs of further improvements.231 A model is always a limited representation of the real world reflecting a certain worldview, and for social systems there is no such thing as a perfect and correct model.232 Größler discusses

227

228

229

230

231

232

Forrester, Jay W. and Peter Senge: “Tests for Building Confidence in System Dynamics Models”, in Legasto, Augusto A., Jay W. Forrester und James M. Lyneis (eds.), TIMS Studies in the Management Sciences, Vol. 14, Amsterdam, 1980, pp. 217—223. Forrester and Senge: “Tests for Building Confidence in System Dynamics Models”, in Legasto, Forrester, und Lyneis (eds.), TIMS Studies in the Management Sciences, Vol. 14, Amsterdam, 1980, pp. 224—225. See Barlas, Yaman: “Formal aspects of model validity and validation in system dynamics”, System Dynamics Review, Vol. 12, No. 3, 1996, p. 188; Forrester and Senge: “Tests for Building Confidence in System Dynamics Models”, in Legasto, Forrester, und Lyneis (eds.), TIMS Studies in the Management Sciences, Vol. 14, Amsterdam, 1980, p. 210. Barlas: “Formal aspects of model validity and validation in system dynamics”, System Dynamics Review, Vol. 12, No. 3, 1996, p. 184. Forrester, Jay W.: “System dynamics, system thinking, and soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, p. 226. See Sterman, John D.: “All models are wrong: reflections on becoming a systems scientist”, System Dynamics Review, Vol. 18, No. 4, Winter 2002, p. 522;

B. Conceptual Foundation for the Usage of System Dynamics

65

bounded rationality in the process view and in the content view.233 In the system dynamics modeling part of an organizational intervention, this can mainly be understood as limitations in the modeling process skills and in problem content skills. Modeling process skills regards such elements as knowledge acquisition, system modeling, methods and tools, testing principles, model analysis, challenging assumptions, fostering discussions and creativity, and theories of human learning. Also general intervention design skills such as stakeholder involvement, ensuring agreement in the process, placing the intervention in a broader organizational development view, and focus on implementation will influence the rationality in the process view. Problem content skills regarding elements as the understanding of the problem at hand, understanding of goal-structures, identification and estimation of the important parameters, establishment of cause-effect-relationships, etc. If both modeling skills and problem content skills are high (i.e. relatively low degree of bounded rationality in both views), it should be expected that the developing process should create fruitful discussions and that the quality and utility of the model should be high. If both modeling skills and problem content skills are low, the risk should be high for both off-track discussions and wrong models. On the other hand, if modeling skills are high but problem content skills are low, the modeling process should hopefully open discussions that challenge some of the wrong perceptions, but the end quality of the model could be expected to be low. And lastly if problem content skills are high, whereas modeling skills are low, then it is likely, that the problem would have been solved better using alternative way of addressing it. Barlas discusses practical differences for model validity with respect to different uses of models, although the same general philosophy applies for the

233

Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, p. 846; Barlas, Yaman: “Formal aspects of model validity and validation in system dynamics”, System Dynamics Review, Vol. 12, No. 3, 1996, p. 187. Größler, Andreas: “A Content and Process View on Bounded Rationality in System Dynamics”, Systems Research and Behavioral Science, Vol. 21, No. 4, July/August, 2004, pp. 319—330.

B. Conceptual Foundation for the Usage of System Dynamics

66

different modes of usage.234 The differences include the limited emphasis on behavior accuracy testing for models used in theory testing in theoretical research, and an increased focus on testing against real data in practical applications focusing on improving real systems. For interactive simulation games, including management flight simulators, an added step of validity must be added regarding usage and learning elements, and for participative modeling processes, testing of the often rather small models is likely to take place partly as an integral part of the model-building discussions.235 For models addressing regulators of systems, Ashby has proposed The Law of Requisite Variety, that only variety can destroy variety, arguing that one of the validity criterion for this type of model is that the model reflects the complexity and details of the system.236

c.

Policy Formulation and Implementation

Policy formulation includes the creation of entirely new strategies, structures, and decision rules.237 In Forrester’s six-step modeling process, the three last steps address policy formulation and implementation, these steps being: ‘Design Alternative Policies and Structures’, ‘Educate and Debate’, and ‘Implement Changes in Policies and Structures.’238 The design of alternative policies and structure is determined based on simulation of alternatives. The alternatives to be simulated may occur from intuitive insights gained in the previous modeling phases, as suggestions from the experienced analyst or people from the operating system, or as a result of a systematic change of parameter-settings. Simulations 234

235

236

237

238

Barlas, Yaman: “Formal aspects of model validity and validation in system dynamics”, System Dynamics Review, Vol. 12, No. 3, 1996, pp. 199—200. Barlas, Yaman: “Formal aspects of model validity and validation in system dynamics”, 1996, pp. 200—201. Ashby, W. Ross: An Introduction to Cybernetics, paperback version, London, 1964, pp. 206—207. A regulator of the air traffic flows around New York is given as example. See also Conant, Roger C. and W. Ross Ashby: “Every Good Regulator of a System Must Be a Model of that System”, International Journal of System Sciences, Vol. 1, No. 2, 1970, pp. 89—97. Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, p. 104. Forrester, Jay W.: “System dynamics, system thinking, and soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, pp. 245—247.

B. Conceptual Foundation for the Usage of System Dynamics

67

are stressed throughout system dynamics literature as being an essential part of the modeling process. Complementary to the many advantages of making mental models explicit in the system conceptualization and model formulation phases (discussed earlier in this chapter), the simulation phase offers two essential rewards: (1) overcoming the limited cognitive ability in handling a larger number of parameters, feedback-loops and delays, and (2) experimentation is a very effective way of learning.239 ‘Educate and Debate’ is a step Forrester proposes aiming at establishing consensus for implementation, addressing both active and passive change resistance, and is likely to yield questions requiring repeated work in the previous modeling steps. The last step in Forrester’s six step modeling process, implementation, is strongly dependent on the success of the activities in the ‘educate and debate’ step, and besides normal implementation activities, it is also recommended to ensure later evaluation of the policy changes. The implementation of the new policies consists of two conceptually different elements: implementation elements incorporated in the modeling process, and implementation elements belonging to traditional change implementation roll-out activities. The implementation elements incorporated in the modeling process are especially strong in participative modeling approaches, such as group model building, focusing on involvement and creating improved and aligned mental models among decision-makers.240 The traditional change implementation roll-out activities are seldom described in the system dynamics literature, but are extensively discussed in the literature of organizational psychology and the normative organization development literature.

239

240

Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993, p. 31; Sterman, John D.: “All models are wrong: reflections on becoming a systems scientist”, System Dynamics Review, Vol. 18, No. 4, Winter 2002, p. 522. See Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 97. Although, the importance for designing the modeling process from the very start with implementation in view is also emphasized in more traditional system dynamics articles, see Roberts, Edward B.: “Strategies for Effective Implementation of Complex Corporate Models”, in Edward B. Roberts (ed.): Managerial Applications of System Dynamics, Cambridge, Massachusetts, 1978, p. 83

68

B. Conceptual Foundation for the Usage of System Dynamics

2.

Designing System Dynamics Modeling-Based Interventions a.

Experimentation-Based Learning Cycles

A modeling process with a model well focused on the core of the problem at hand, constructed and simulated according to best practices within the field, and yielding highly interesting and relevant insights does not necessary result in organizational change. The implementation challenge has many times proven to be of such magnitude that otherwise successful modeling studies result in only very limited change. Based on a discussion of the implementation problems, Weil offered a list of critical success factors for implementation of system dynamics projects, which included factors regarding focus and urgency of the problem as well as a number of factors regarding client involvement.241 Roberts addressed the problem with the statement that “Organizational changes (or decisions or policies) do not instantly flow from evidence, deductive logic, and mathematic optimization.”242 This is, as described earlier in this chapter, in accordance with most behavioral research emphasizing many other elements beside cognition, such as emotions, social norm, control beliefs, and group dynamic aspects. A central characteristic of the implementation challenges in system dynamics context is that modeling results and insights are difficult to transfer to others.243 Two conceptually different ‘movements’ have crystallized out of the system dynamics field addressing the difficulties of transferring modeling results: (1) Participative system dynamics modeling, and (2) Simulation-based knowledge transfer. Participative system dynamics modeling is mainly described in the Group Model Building and Modeling for Learning literature. Participative modeling seeks to bypass the difficulties of transferring modeling insights from model builders to decision-makers by involving the decision-makers in the actual 241

242

243

Weil, Henry B.: “The Evolution of an Approach for Achieving Implemented Results from System Dynamics Projects”, in Jørgen Randers (ed.): Elements of the System Dynamics Method, Cambridge, Connecticut, 1980, p. 290. Roberts, Edward B.: “Strategies for Effective Implementation of Complex Corporate Models”, in Edward B. Roberts (ed.): Managerial Applications of System Dynamics, Cambridge, Massachusetts, 1978, p. 77. See Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993, p. 31; Vennix, Jac A. M.: Group Model Building, Chichester, 1996, pp. 97—99.

B. Conceptual Foundation for the Usage of System Dynamics

69

model-building activities. In research settings, active participation in the modeling or simulation activities has indicated significantly more enduring changes in people’s mental models compared to traditional transfer of insights.244 Enduring mental model changes are central to the implementation of sustainable organizational change. Simulation-based knowledge transfer includes concepts such as management simulators, simulation games, flight simulators, learning laboratories and managerial practice fields. Here, a user-friendly version of a model is used as an instrument in a learning context, not as involvement in modeling activities, but as involvement in simulation activities. The simulation-based knowledge transfer supports investigation of possible actions without the participant having to face the consequences of unintended side effects, and accelerating understanding of the behavior of the underlying system model, which equals getting years of experiences in a few days training. The effectiveness of the simulations-based knowledge transfer does not only depend on the quality and relevance of the underlying model, but also on the process of how it is applied; e.g. structured formulation of hypotheses (mapping of mental models), as well as creation of open and creative discussions. Participative system dynamics modeling typically has a strong exploratory focus aimed at getting new insight on system behavior of a problem, whereas simulation-based knowledge transfer is mainly an instrument for transferring insights from the original model-builders to another group of people. Nevertheless, both approaches build on the understanding that experimentation is an effective way of learning, as formulated in The Learning Cycle by Maani and Cavana (figure B-8).

244

See Vennix, Jac A. M.: Group Model Building, Chichester, 1996, pp. 97—99.

70

B. Conceptual Foundation for the Usage of System Dynamics

Conceptualization (intellectual learning)

Reflection (emotional learning)

Experimentation (action learning)

Figure B-8: The learning cycle for learning labs 245

Maani and Cavana describe Conceptualization as cognitive learning, where people develop theories and hypotheses regarding system behavior, as well as the formalization of learning outcomes from experimentation and reflection.246 Experimentation is described as a learning experience much like real-life learning, although in an accelerated and risk-free way (“learning by doing” in a laboratory). Reflection is described as emotional learning allowing participants to engage with affective elements such as assumptions, attitudes, biases, resentments etc. Consequently, the learning lab process illustrates the usage of applying experimentation to move beyond cognitive learning, which according to the literature as described in subchapter B.II.2. is significant for effective learning and influencing individual’s attitudes towards behavior.

245

246

Taken from the figure in Maani, Kambiz E. and Robert Y. Cavana: Systems Thinking and Modeling – Understanding Change and Complexity, Auckland, 2000, p. 112. In Maani and Cavana: Systems Thinking and Modeling – Understanding Change and Complexity, 2000, pp. 112—113, the three elements of the learning cycle are discussed.

B. Conceptual Foundation for the Usage of System Dynamics

b.

71

Knowledge Acquisition in Modeling Projects

Knowledge acquisition is an important aspect of designing system dynamics interventions. Forrester point out three types of information sources for modeling projects: mental database, the written database and the numerical database:247

Mental Database Observation Experience

Written Database

Numerical Database Figure B-9: Mental database and decreasing content of written and numerical databases 248

By ‘mental database’, Forrester means the non-documented information in people’s minds, based on observations and experience. Forrester stresses the importance of this type of information, arguing that it is not adequately appreciated in the management and social sciences.249 The written database and 247

248

249

Forrester, Jay W.: “Policies, decisions and information sources for modeling”, European Journal of Operational Research, Vol. 59, No. 1, 1992, pp. 55—58. Slight modification of figure in Forrester, Jay W.: “Policies, decisions and information sources for modeling”, European Journal of Operational Research, Vol. 59, No. 1, 1992, p. 56. Forrester’s figure is also depicted in Luna-Reyes, L.F. and D. L. Andersen: “Collecting and analysing qualitative data for system dynamics: methods and models”, System Dynamics Review, Vol. 19, No. 4, 2003, p. 280, along with a discussion on the three types of databases. Forrester, Jay W.: “Policies, decisions and information sources for modeling”, European Journal of Operational Research, Vol. 59, No. 1, 1992, p. 56.

72

B. Conceptual Foundation for the Usage of System Dynamics

the numerical database naturally also provide important information to the modeling process, although caution must be taken as the information has already been filtered by the writer’s purpose.250 Although for the numerical database in particular, Warren argues, that companies are often surprised when existing data are seen in a new perspective due to the modeling efforts, and this actually often produces new insights in itself.251 In retrieving the mental databases, a number of different techniques are described in the system dynamics literature. Luna-Reyes and Andersen provide an extensive overview of methods for collecting and analyzing qualitative data from the perspective of the different modeling phases.252 Fey and Trimble as well as Vennix also discuss a larger number of these techniques, including the different interview forms (fully structured, semistructured or unstructured), questionnaires, the Delphi method, and different versions of the Normative Group Technique.253 Information gathering includes the opportunity - and the risk - of influencing the persons providing the information. Especially in participative modeling projects, information gathering is a balance between structured (and maybe even anonymous) data and knowledge elicitation from experts and individuals and a team effort among participators in the modeling process. In the formulation of a quantitative model, more often than not, the estimation of some parameters will suffer from lack of explicit or precise data. Most literature agrees that building on a qualified guess is nearly always better

250

251

252

253

Forrester.: “Policies, decisions and information sources for modeling”, European Journal of Operational Research, Vol. 59, No. 1, 1992, p. 57. Own notes from Kim Warren’s presentation at the System Dynamics Conference in Oxford in 2004. Luna-Reyes, L.F. and D. L. Andersen: “Collecting and analysing qualitative data for system dynamics: methods and models”, System Dynamics Review, Vol. 19, No. 4, 2003, pp. 271—296. Besides the qualitative methods often mentioned in the system dynamics literature like interviews, focus and Delphi groups, observations, etc., the article also include analysis methods like hermeneutics, discourse analysis, grounded theory, ethnographic decision models and content analysis. Fey, Willard and John Trimble: “The Evaluation and Development of Knowledge Acquisition in System Dynamics Studies”, in Proceedings, System Dynamics Conference, System Dynamics Society, 1992, pp. 173—182; Vennix, Jac A. M.: Group Model Building, Chichester, 1996, pp. 175 and pp. 187—188.

B. Conceptual Foundation for the Usage of System Dynamics

73

than leaving out an important parameter in a model.254 Warren argues that in the everyday business, decisions are made implicit based on judgment and assumptions of the parameters, where data is missing.255 Asking for qualified guesses is just a way to make such judgments and assumptions explicit.

c.

Designing Participative Modeling Interventions

A system dynamics study will often be part of a larger package addressing an important business problem.256 Apart from the seldom-used expert models developed ‘behind the scenes’, most system dynamics projects in organizational interventions will include some elements of participative modeling. Participative system dynamics modeling is primarily described in terms of Group Model Building and Modeling for Learning, and the participative modeling literature in general position the modeling process as an integral part of management discussions.257 Vennix, Andersen, and Richardson discuss the goals of Group Model Building with regards to: (1) creating a climate for team learning, problem understanding and mental model alignment, (2) fostering consensus, and (3) creating acceptance and commitment with the decisions.258 In Modeling as

254

255 256

257

258

Milling, Peter: “Organisationales Lernen und seine Unterstützung durch Managementsimulatoren”, in: Zeitschrift für Betriebswirtschaft, Ergänzungsheft 3/95, Lernende Unternehmen (Sonderausgabe), 1995, p. 104. Warren, Kim: Competitive Strategy Dynamics, Chichester, 2002, p. 25. Weil, Henry B. and Kenneth P. Veit: “Corporate Strategic Thinking: The role of System Dynamics” in Peter M. Milling and Erich O.K. Zahn (Eds.): ComputerBased Management of Complex Systems, Proceedings of the 1989 International Conference on the System Dynamics Society, Stuttgart, 1989, p. 67, discuss how system dynamics is often part of a larger package addressing a problem, e.g. in a strategy consulting project. See Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 4; Lane, p. 70; Richmond, Barry: “The Strategic Forum: aligning objectives, strategy and process”, System Dynamics Review, Vol. 13, No. 2, 1997, p. 131; Morecroft, J. D. W.: “Executive knowledge, models and learning”, European Journal of Operational Research, Vol. 59, 1992, p. 13. See Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 6; Andersen, David F., George P. Richardson and Jac A. M. Vennix: “Group model building: adding more science to the craft”, System Dynamics Review, Vol. 13, No. 2, 1997, p. 191.

74

B. Conceptual Foundation for the Usage of System Dynamics

Learning, Lane strongly stresses the learning objectives, such as changing mental models, improving intuition, allowing risk-free experimentation and revealing systemic complexity; as vehicles to business improvement.259 Morecroft discusses the usage of system dynamics in the context of executive dialogue and debate:

changing business environment

recognised strategic issue

Knowledge base & mental models

executive debate and dialogue

action plans and change facilitation

MICROWORLD

MAPS words, diagrams friendly algebra

Simulation models & gaming simulators

FRAMEWORK

Concepts and theory plus facilitation

Figure B-10: Maps, frameworks and microworlds 260

259

260

Lane, David C.: “Modelling as Learning: A consultancy methodology for enhancing learning in management teams”, European Journal of Operational Research, Vol. 59, No.1, 1992, p. 72. Taken from the figure in Morecroft, J. D. W.: “Executive knowledge, models and learning”, European Journal of Operational Research, Vol. 59, 1992, p. 14. The full article is also printed in the book: Morecroft, John D. W. and John D. Sterman (eds.): Modeling for Learning Organizations, Portland, Oregon, 1994, pp. 3—28 (the book is in principle a reprint of the special issue of European Journal of Operational Research, Vol. 59, No. 1, 1992).

B. Conceptual Foundation for the Usage of System Dynamics

75

The upper-right rectangles in figure B-10 represent a stack of individual knowledge bases and mental models, and these will always implicitly be active in executives recognizing and discussing strategic issues. The lower-right rectangle represents explicit application of models in the dialogue, with the usage of maps (examples are written lists, or causal diagrams) and a framework combining maps with concepts and theory.261 The term Microworlds encompasses in Morecroft’s understanding all the labels and the connections in figure B-10, and is used for the whole learning environment. Figure B-10 does not necessarily imply a dedicated group model building process, but could insure involvement through individual interviews and model discussions (to make sure the model reflects each individual’s mental model), and subsequent usage of the model as a management simulator. In general, though, most group model building and modeling for learning literature emphasize that insights are gained during a model-building process, rather than after the modelbuilding process.262 Andersen, Richardson and Vennix argue for the value of a group modeling structure, where top management and “doers” are together for an extended period. Argyris argue, for interventions in general, that the more difficult an organizational intervention, and the more internal commitment is necessary for effectiveness, the more the client needs to be directly involved in the design, execution and monitoring of the changes with an open and experimental attitude.263 This also contributes to placing the intervention in the continuous learning and shaping of the organization and its change readiness.264 With the importance of creating shared mental models and establishing consensus and commitment for decisions, one of the important design decisions is who to involve in the modeling. If the modeling is used as a tool in strategic planning, top management is likely to be relevant to involved. If the modeling is used to address a specific problem, the main stakeholders should be involved. Also, a project can operate with a structure, that allows for involvement of a large

261

262 263

264

Morecroft, J. D. W.: “Executive knowledge, models and learning”, European Journal of Operational Research, Vol. 59, No. 1, 1992, p. 13. See Vennix, Jac A. M.: Group Model Building, Chichester, 1996, pp. 97—99. Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Massachusetts, 1970, p. 83 Schein, Edgar H.: Organisationspsykologi, Danish translation, Herning, 1990, p. 40; Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Massachusetts, 1970, chapter 1 and 2.

76

B. Conceptual Foundation for the Usage of System Dynamics

number of people, using a setup with steering committee, a core group, reference groups etc. 265 The involvement will often be designed with the dual purpose of acquiring information and ideas, balancing differing perspectives and interests, and establishing internal commitment. A way to coordinate the stream of information in a group model building process is the usage of a Workbook. Vennix describes a Workbook as a document consisting of text and diagrams, and presenting material as well as asking questions.266 The Workbook can be seen as an instrument to summarizing workshops and preparing for the next workshops and is a way to formalize the information stream in a modeling project. Akkermans discusses the concept of The Workshop-Workbook Cycle, being a structured, iterative process involving a number of in-house discussions (among consultants) and client consultations, as well as Workbook distribution and feedback, between all workshops.267 When designing a modeling intervention, one of the decisions is whether or not to employ a preliminary model. The advantages include speeding up the process and reducing participants time investment, but the major risk is to infringe the degree of the group’s sense of model ownership, and therefore Vennix stresses the importance of not being defensive about a preliminary model when discussing and refining the model. 268 Vennix furthermore offers a schematic overview of factors indicating in which situations it is appropriate to use a preliminary model, as well as other design decisions such as communication pattern, usage of soft methods to initiate the process, and facilitator preparation.269

265

266 267 268 269

See Akkermans, Henk: Modelling With Managers, Breda, The Netherlands, 1995, p. 87; Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 114. Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 128. Akkermans, Henk: Modelling With Managers, Breda, 1995, p. 89. Vennix, Jac A. M.: Group Model Building, 1996, p. 113. Vennix: Group Model Building, 1996, p. 132.

B. Conceptual Foundation for the Usage of System Dynamics

d.

77

Facilitation of Participative Modeling Interventions

Effectiveness of using modeling in the learning cycle of an organizational intervention not only depends on intervention design and frameworks, but also on facilitation skills in knowledge elicitation and the handling of the discussions of options and consequences.270 One often encounters discussions on how to reduce dependency of a facilitator through the usage of effective scripts, subcomponents and rigor approaches, but nevertheless most group model building literature stresses the importance of the quality of the facilitator role in system dynamics projects. 271 In group model building literature, the facilitation role is frequently called the role of the modeling team, indicating the usefulness of having more than one facilitator. Vennix argues that a minimum of two people should guide the process, a facilitator and a recorder, however the recorder could be from the client organization.272 Richardson and Andersen suggest that for supporting the modeling process for large groups addressing weighty problems, five roles that are best handled by five individuals are: The Facilitator (focus on group processes and knowledge elicitation), The Modeler/Reflector (focus on model formulation, and sometimes feeding own sketches back to the group), The Process Coach (focus on the dynamics of individuals and subgroups), The Recorder (writing down and sketching group proceedings), and The Gatekeeper (often a person from the client organization bridging the modeling process and the client problem).273

270

271

272 273

Morecroft, J. D. W.: “Executive knowledge, models and learning”, European Journal of Operational Research, Vol. 59, No. 1, 1992, p. 15. See Andersen, David F and George P. Richardson: “Scripts for group model building”, System Dynamics Review, Vol. 13, No. 2, 1997, p. 108. See also Andersen, David F., George P. Richardson and Jac A. M. Vennix: “Group model building: adding more science to the craft”, System Dynamics Review, Vol. 13, No. 2, 1997, p. 195, where furthermore The Gifted Practitioner hypothesis is stated, indicating that some facilitators are skilled to a degree where choice of intervention method is less important – and also the other way around - if facilitation skills are lacking, no methods and tools can assure an effective intervention. Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 134. Richardson, George P. and David F. Andersen: “Teamwork in group model building”, System Dynamics Review, Vol. 11, No. 2, 1995, pp. 114—115.

78

B. Conceptual Foundation for the Usage of System Dynamics

Effective learning in teams requires focus on interpersonal relations and group dynamics.274 The facilitator must strive to create a climate allowing for open-mindedness, learning and creativity. One aspect of ensuring openmindedness is to strive for acceptance of the model by ensuring that the participants understand all model elements and that the model encompasses all major elements from each participant’s mental model. Participant acceptance and understanding of the model is critical for gaining trust in the model, which is a prerequisite for learning. Vennix discuss the equation, E = Q x A (Effectiveness= Quality x Acceptance), implying that as long as a model is correct in the larger perspectives, participant’s acceptance must be strived for, even when it means compromising on less significant modeling issues.275 Especially in the problem definition phase, the facilitator has to manage not only idiosyncratic views on the problem held by the individual decision makers, but also the problem of the decision makers “speaking different languages.”276 The different use of language, often discussed in terms of codes, is a critical problem in communication and interpretation processes. Codes are important in determination of meaning in communication, and creating shared meaning requires use of similar codes, which is why Scheper recommends giving considerable attention to codes in the determination of the problem in the early phase of a group decision process.277 Eco has proposed a widely used model of communication (figure B-11).

274

275 276

277

Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Massachusetts, 1970, p. 83. Vennix, Jac A. M.: Group Model Building, Chichester, 1996, p. 6. Scheper, Willem J.: Group Decision Support Systems, Tilburg/Utrecht, 1991, p. 23. Scheper.: Group Decision Support Systems, 1991, p. 143.

79

B. Conceptual Foundation for the Usage of System Dynamics

sender

coded meassage

channel

codes

message as expression

addressee

context, circumstance

codes

interpreted message as content

Figure B-11: A model of communication 278

Spending time on the modeling process to agree on definitions and the usage of language in understanding problems and issues contributes not only to align codes between individuals, but should also contribute to the usage of larger “chunks” of codes. Miller uses an example of a novice in radiotelegraphic code having to decode every “dit and dah”, whereas more experience will allow him to deal with each letter as a chunk, and later even operate with words as large chunks.279 In participative modeling processes, this would mean usage of larger chunks to allow discussions to build on accumulative insights. The facilitator role also deals with reducing information filters in information acquisition. Morecroft discusses the following five different information filters in decision-making processes: 280 (1) people’s cognitive limitations, (2) operating goals rewards and incentives, (3) information, measurement, and communication systems, (4) organizational and geographical structure, and (5) tradition, culture, folklore, and leadership.

278

279

280

Slight simplification of the model in Eco, Umberto: A Theory of Semiotics, Bloomington, 1976, p. 141. The model has originally also focus on sub-codes, which is especially relevant in the discussion of undercoding and overcoding. The simplification is mainly inspired by the use in Scheper, Willem J.: Group Decision Support Systems, Dissertation, Tilburg/Utrecht, 1991, p. 143. Miller, George A.: “The Magical Number Seven, Plus Minus Two: Some Limits on Our Capacity for Processing Information”, The Psychological Review, Vol. 63, No. 2, March 1956, p. 93. Morecroft, J. D. W.: “Executive knowledge, models and learning”, European Journal of Operational Research, 59, 1992, p. 18.

80

B. Conceptual Foundation for the Usage of System Dynamics

Although the facilitator is also subject to these filters, conscious considerations of how to reduce the information filtering is expected to improve the modeling process. The last facilitator role to be discussed in this subchapter is the role of ensuring momentum in the process, which includes handling attitudes to the project and internal conflicts.281 The Wallow Curve, figure B-12, illustrates an often-experienced tendency of a peak of irritation and conflict in the project, after the first few workshops, when the initial period of getting to know the problem and conceptualize the issues is over.282 The irritation and conflict often reflects uncertainty regarding the quality and usefulness of the model in the early part of the design phase, and also that certain dominant project members are trying to push through their own ideas. It is important to note that the facilitator should not necessary aim at minimizing the level of cognitive conflicts, as the highest decision quality often is found in situations where groups have some degree of conflict, with the reasoning being that disagreement fosters thorough investigation, information processing and searching for alternatives.283 Furthermore, disagreement and vivid discussions can be a stimulating element in fostering innovation and creativity.284 Only the extremes should nearly always be avoided: too much cognitive conflict between group members will result in ineffective communication, and too little conflict will result in Groupthink.285

281

282

283 284 285

For further descriptions on facilitation of group model sessions, see Vennix, Jac A. M.: Group Model Building, Chichester, 1996, pp. 140—171. See Akkermans, Henk: Modelling With Managers, Breda, The Netherlands, 1995, pp. 91—92, where credit for the term The Wallow Curve is given to McKinsey. Vennix, Jac A. M.: Group Model Building, 1996, p. 154. Numerious examples are seen in the art and music industries. Decision quality as a function of conflict level is often seen depicted as a reversed U-curve.

B. Conceptual Foundation for the Usage of System Dynamics

81

problem understanding

irritatoin & conflict level

inter- WS1 views

WS2

WS3

WS4

WS5

Figure B-12: The Wallow Curve at work 286

286

Taken from the figure in Akkermans, Henk: Modelling With Managers, Breda, 1995, p. 91.

C. A Case Study Using Participative System Dynamics Modeling in the Implementation of a Sensitive Change Project I.

Research Considerations for the Case Study Application

The application of system dynamics modeling has proven useful in exploratory organizational interventions because it supports learning, alignment of mental models, group decision-making and the creation of commitment. Consequently, system dynamics has contributed to the formulation of a number of strategic initiatives in corporate settings. The initial question being investigated in this dissertation is: is it purposeful to apply system dynamics modeling in the implementation of strategic initiatives?287 The research is mainly based on discussions of literature, but is also complemented by a case study in the action research tradition with dual focus on the implementation of planned change as well as knowledge development.288 Voss, Tsikriktsis, and Frohlich categorize case research based on purpose: exploration, theory building, theory testing, or theory extension/refinement.289 In these terms, the purpose of the present case study is theory building, since no particular existing theory regarding the use of participative modeling in a planned change organizational intervention viewpoint supports the research. Furthermore,

287

288

289

An alternative scientific approach could have been to investigate the challenges of bridging the two problem-solving cycles in organizational change projects, and to evaluate and compare the usage of a number of different methods and tools in such a setting. The arguments for the applied scientific approach include the decision to make one in-depth case study in action research tradition, rather than to spread focus and efforts over a number of field studies. For definitions of action research, see Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, p. 23. Voss, Chris, Nikos Tsikriktsis, and Mark Frohlich: “Case research in operations management”, International Journal of Operations & Production Management, Vol. 22, No. 2, 2002, p. 198.

84

C. Case Study

the research approach involving a single site case study does not have the design characteristics necessary to qualify for testing a theory. 290 Discussing whether to use one or more case studies in a research project, Stake argues for the benefit of maximizing knowledge through focusing on a single, in-depth case study, rather than to disperse focus over a number of case studies.291 Other researchers, however, champion the viewpoint that theory building should preferably be based on insights from a larger number of case studies.292 The main argument for the decision to use only one case study in the present research is the iterative seekand-learn relationship that has taken place between the case study and the theoretical discussion on how best to apply modeling in change management.293 Furthermore, a survey approach was not considered a possible alternative, due to lack of known cases utilizing system dynamics modeling in change management context. As an alternative to a multiple case study approach or a survey approach, the single-site case study approach puts the spotlight on one instance to be investigated in more detail and thereby concentrates efforts rather than trying to

290

291

292

293

This fits well with the quote from Andersen, David F.; George P. Richardson and Jac A. M. Vennix: “Group model building: adding more science to the craft”, System Dynamics Review, Vol. 13, No. 2, 1997, p. 196: “case studies are only suitable to generate hypotheses, not to test them rigorously.” Stake, Robert E.: The Art of Case Study Research, Thousand Oaks, 1995, pp. 4—5, argues that a good instrumental case study not necessarily needs to examine a typical case, as an unusual case helps illustrate matters often overlooked in typical cases. Furthermore, he argues that even when designing a collection of cases, representation is often difficult to defend. See Eisenhardt, Kathleen M.: “Building Theories from Case Study Research”, Academy of Management Review, Vol. 14, No. 4, 1989, pp. 532—550; and Leonard-Barton, Dorothy: “A Dual Methodology for Case Studies: Synergistic use of a Longitudinal Single Site with Replicated Multiple Sites”, Organizational Science, Vol. 1, No. 3, 1990, pp. 248—266. It should be noted, that even though the present dissertation includes only one case study, it indirectly benefits from a larger number of case studies from the group model building literature, as many of these have elements of a change management focus. See also discussion on exploratory case studies, Yin, Robert K.: Case study Research, 3 rd edition, Thousand Oaks, 2003, p. 23.

C. Case Study

85

cover a large number of instances. The aim is to “illuminate the general by looking at the particular”.294 The design of the case study is based on literature studies as well as the external facilitator’s experiences of organizational interventions due to 10 years of management consulting experience in business process reengineering and change management.295 The literature studies formed the first basis for the theory development on how to benefit from system dynamics modeling in change management context.296 During the case study, the practical insights gained served in parallel as inspiration on where and how to focus further literature studies. The literature studies include primarily literature from the field of system dynamics, the field of organizational psychology, and descriptive management literature regarding implementation of strategic initiatives, often called change management literature. The literature of system dynamics contributes in general to the case study with a dynamic perspective on problem understanding and individual learning, with the emphasis on model-building based on stock and flow structures, and with computer simulations of aggregate time-series. Group Model Building literature contributes in particular with focus on modeling as a framework for discussions facilitating the sharing and aligning of mental models. The core of the model used in the case study consists of two aging chains, and the issues of delays constitute a central part of the analysis. Respecting the scientific foundation and the large number of proven successful system dynamics interventions, the case study does not intend to change or challenge the system dynamics modeling process as described in the leading system dynamics literature, but rather the case study seeks to place the modeling process in a different context: applying system dynamics modeling in planned change organizational intervention.

294

295

296

Denscombe, Martyn: The Good Research Guide for small-scale social research projects, Philadelphia, 2 nd edition, 2003, p. 30. The external facilitator (the author of this dissertation) has most of her management consulting experience from IBM Management Consulting and Deloitte & Touche Consulting Group (now split into Deloitte Consulting and Braxton). Yin, Robert K.: Case study Research, 3 rd edition, Thousand Oaks, 2003, p. 28, emphasizes the importance of theory development as a part of case study design, regardless of whether the purpose of the case study is to develop or test theories.

C. Case Study

86

The literature of organizational psychology contributes in particular to theories of action research, individual learning processes and group psychology. Action research takes the theoretical point of departure that dynamic systems such as organizations can best be examined using carefully planned, conceptbased interventions. Thus, the evaluation cannot involve traditional evaluation schemes, such as the use of control groups, as the complex social environment is not controllable to a degree that allows isolation of the true variables. Furthermore, all evaluation processes of human systems influence the system, and must be designed in such a way that academic evaluation is not counterproductive with regards to the desired results of the intervention. This is in accordance with a premise of action research, namely that action research projects should always be adding value to a given organizational intervention, and therefore not differ from the ethics underlying consulting and counseling projects.297 The theories regarding individual learning and group processes are mainly included as a focus on creation of changes in attitudes and intentions, which also play a major part in the applied evaluation framework. Change management literature offers recommendations on creating the foundation for the rollout of the strategic initiative, including thoughts on the development of change leaders.298 The case study also benefits from the more traditional change management virtues such as intervention planning, stakeholder management, and implementation planning and review:299

•

297

298

299

Intervention planning includes the definition of business objectives and targets, the framing of the intervention, the identification of consultation relationships, roles and responsibilities in the project organization, and

Schein, Edgar H.: Organisationspsykologi, Danish translation, Herning, 1990, p. 253. Anderson, Linda A. and Dean Anderson: “Awake at the Wheel: Moving beyond Change Management to Conscious Change Leadership”, OD Practitioner, Vol. 33, No. 3, 2001, p. 45. See also Roberto, Michael and Lynne Levesque: “The Art of Making Change Stick”, MIT Sloan Management Review, Summer 2005, Vol. 46, No. 4, Summer 2005, p. 56, where modeling is relevant with regards to what they call chartering and learning. Intervention planning, stakeholder management and implementation planning and review are three conceptually different, but also strongly interdependent tasks.

C. Case Study

87

time and budget planning.300 Thus, this activity - as the other two activities discussed next - contains both system dynamics and nonsystem dynamics parts, for instance the coordination with related projects and activities.

300

301

302

303

•

Stakeholder management involves a thorough analysis of all the major interest groups and individuals who have significant influence—directly or indirectly—on the success of the intervention. Focus is on the stakeholders’ interests and power, their importance for solution design and implementation, and on relevant means of involvement and communication.301 Stakeholder analysis is vital to intervention planning, both to secure relevant parameters to be included in the process, and to secure appropriate involvement of and communication with stakeholders, including employees.302

•

Implementation planning and review deals with the planning of the implementation, including an action plan, communication plan and a clear delegation of responsibilities.303 The communication plan develops

Roberto, Michael and Lynne Levesque: “The Art of Making Change Stick”, MIT Sloan Management Review, Summer 2005, Vol. 46, No. 4, Summer 2005, pp. 55—56, describe intervention planning (although calling it chartering). In Koningswieset, Roswita and Alexander Exner: Systemische Intervention, Stuttgart, 1998, it is throughout the book discussed how intervention planning can be seen as creating an overall architecture and design of the intervention, including selection of relevant techniques. Flood, Robert L.: Solving Problem Solving, 1995, p. 20, places ‘organizational politics’ as one of the four key dimensions to take into consideration in a whole organizational system view, the three others being organizational processes, design and culture. Borum, Finn: Strategier for organisationsændringer, Copenhagen, 1995, pp. 79—92, offers a discussion on politics in change processes. For discussions on “Employee Involvement”, see both Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, p. 317; and Thun, Jörn-Henrik, Peter M. Milling, and Uwe Schwellbach: “The Impact of Total Employee Involvement on Time-based Manufacturing“, in “What Really Matters in Operations Management“, Proceedings of the European Operations Management Association, 8th International Annual Conference, 2001, pp. 133—135. For implementation actions, focus is typical on the practical adjustments of business processes and organizational design with single-loop orientation (result

88

C. Case Study

over the course of the intervention and includes elements such as motivating change and the communication of visions, results, implementation plan and successes. This activity furthermore includes the establishment of procedures for reviews and corrective actions.

The case study addresses problems and issues in bridging between the topmanagement launching a strategic initiative and the sustainable rollout of the organizational change. The main arguments for selecting this specific case include: (1) that the change initiative represented important and classical change management challenges such as significant change resistance, strong emotional attitudes, and that the decision to launch the initiative was made without involving the key managers responsible for the implementation, and (2) the change project had characteristics making it likely to gain from the insights from a system dynamics model, due to the importance of understanding the influence of delays in an aging chain. Furthermore, the writer of this dissertation had extensive knowledge of the company and easy access to the decision-makers, due to previous engagements with the company as well as due to personal relationships. The existence of personal relationships allowed for increased top executive involvement in the design of the modeling process, and also allowed for extensive theoretical discussions with the project owners on the practical implications of the usage of system dynamics modeling in corporate settings.304 However, personal relationships also create biases in terms of design opportunities as well as participant perceptions of the researcher’s objectivity, which has to be taken into account especially in the evaluation of the case study. As a conclusion on the research considerations: the main research idea has been to investigate the usefulness of participative system dynamics modeling in

304

orientation) rather than double-loop orientation (process orientation), see Borum, Finn: Strategier for organisationsændringer, Copenhagen, 1995, p. 21 and p. 37. Although, having the implementation plan including procedures for review and corrective actions, the scene will be set for later double-loop learning. The implementation plan must take into consideration all four organizational dimensions (processes, design, culture and politics) as described Flood, Robert L.: Solving Problem Solving, 1995, p. 20. Stake, Robert E.: The Art of Case Study Research, Thousand Oaks, 1995, p. 4, recommends to “pick cases which are easy to get to and hospitable to our inquiry.”

C. Case Study

89

the bridging of the two problem-solving cycles: the cycle of diagnostics and decision-making, and the cycle of change management.305 In other words, the dissertation investigates the utilization of proven, successful system dynamics methods and tools in a new, specific context. An alternative scientific approach could have been to investigate the challenges of bridging the two problemsolving cycles in organizational change projects, and to evaluate and compare the usage of a number of different methods and tools in such a setting. The arguments for the applied scientific approach include the decision to make one in-depth case study in action research tradition, rather than to spread focus and efforts over a number of field studies. A survey approach was not considered a possible alternative, due to lack of known cases utilizing system dynamics modeling in change management context. Regarding the use of the case study, it is stressed that the purpose primarily is to generate inspiration to the focus of the research due to the seek-and-learn nature of the case study, seeking for theory building in the action research tradition. Additionally, the case study serves as a practical illustration in chapter D, in the discussion of the usage of system dynamics modeling in the implementation of strategic initiatives. The next subchapter describes the case itself: the problem context, the intervention planning, the model, and its results. Following the case description comes a subchapter outlining the evaluation framework for the case study as well as the actual evaluation.

II. 1.

Case Study: Refining and Implementing a Location Strategy The Problem and its Context

The Case Study Company is a major, international company, which is a market leader in its main product area. Research and development (R&D) is a large and critical part of the company’s sustainable competitive advantage, reflected by the fact that approximately every third employee works in R&D. The company has a strong tradition of employee empowerment and is a relatively un-hierarchical

305

For a discussion on the two cycles of the problem-solving process, see chapter A.I.

90

C. Case Study

organization. The case study was carried out in one of the major R&D divisions, consisting of a number of rather different R&D business units. Strategically, the company works towards creation of a truly global company in terms of the entire value chain distribution, and as part of these efforts the board of directors launched a strategic initiative regarding a balanced location strategy.306 For R&D, specific targets were decided pertaining to the distribution of the number of R&D employees placed in high-cost countries (e.g. the USA and Western Europe) vs. the number of R&D employees placed in lowcost countries (e.g. India, China, Eastern Europe). At the time of the launch of the initiative, fall 2004, the company had significantly more R&D employees in high-cost countries with the consequence of relatively high development costs. The cost of an R&D employee in a high-cost country was at this time approximately four times the cost of an R&D employee in a low-cost country. It was a sensitive issue due to the fear among employees in high-cost countries that the future could bring reductions of staff in high-cost locations as seen in many other companies in the USA and Western Europe. The situation at the case study company, however, included strong growth expectations and the company did not intend to weaken existing high-cost locations in terms of the absolute numbers of employees. The expansion of low-cost R&D locations should reflect a worldwide growth of the R&D to gain speed in time-to-market and an extended product portfolio.307 The motivation was not only to increase capacity and cost-efficiency, but also to strengthen the local presence in growing markets such as China and India with a view to increased future sales. New employees in low-cost locations were primarily supposed to take over existing tasks from high-cost locations.

306

307

Roberto, Michael and Lynne Levesque: “The Art of Making Change Stick”, MIT Sloan Management Review, Summer 2005, Vol. 46, No. 4, Summer 2005, p. 53, define strategic initiatives as “corporate programs aimed at creating new business processes or transforming existing ones to accomplish major goals, such as enhancing productivity or improving customer service.” The new location strategy has major impact on both R&D processes, delivery processes, and sales support processes as well as the company’s risk profile, currency spread, and organizational culture elements, which is why it was perceived as a major strategic initiative in the company. Perlitz, Manfred: Internationales Management, 5 th edition, Stuttgart, 2004, p. 72, discusses factors for trade between high-cost and low-cost countries. These factors are also relevant for decisions on distribution of employment.

91

C. Case Study

This way the company wanted to free capacity of experienced senior R&D employees in high-cost countries to be used in new, challenging R&D projects. The objective of the organizational intervention consisted of both ‘harder’ and ‘softer’ elements. The softer elements included the creation of acceptance and commitment to the launched strategic initiative among key implementers, and the harder elements included understanding the most influential parameters in building up capacity in low-cost locations, with special focus on productivity and costs. The process should seek to define the ‘ideal’ implementation approach, balancing board expectations regarding a reduced cost/capacity ratio with an effective and realistic implementation plan. The strategic initiative was intended to stimulate a reinforcing growth loop as depicted in Figure C-1. As such, the project context was not to explore and identify potential new strategies or policies, but to implement a given decision in the best possible way and to gain commitment and understanding from the key implementers. Thus, the intervention was more of a change management project than a policy formulation project.

strengthening low-cost locations without reducing high-cost locations

+ company growth

+

+

+

improved reduced time-to- unit-price of R&D market

+ improved competitiveness

+ maintained motivation in high-cost locations

+ +

Figure C-1: The reinforcing growth loop underlying the intervention 308

308

Causal-loop diagrams are normally constructed in such a way that the variable names do not indicate the direction of change. In Figure C-1, the variable names include the direction, e.g. company growth, reduced unit-price etc. This was done for communication purposes, emphasizing how the reinforcing growth loop was intended to work.

92

C. Case Study

The board of directors launched the strategic initiative as an integral part of the yearly business planning and budgeting process. Although, before incorporating the R&D location strategy in the budget, board approval was needed of the detailed implementation plan. The plan should be detailed with regards to the R&D demand (tasks and project to be carried out in each location), the planned staffing in each location, and planning of the transformation (hiring, transfer of tasks, training, process changes and communication strategy). Consequently, the people responsible for making the changes happen needed to be thoroughly involved in the planning. The board also gave specific targets for the planning: at year-end 2007, 25% of the division’s R&D personal should be based in low-cost locations (as opposed to 10% at year-end 2004), and the new strategy should not negatively influence the existing preliminary 3 years cost budget. Also, negative productivity effects in 2005 should be avoided. The board recognized the significant change management challenges following the launch of the initiative. Transfer of jobs to low-cost locations was an issue often discussed at that time in the high-cost society in general, and also at the company’s high-cost locations. For this reason, a decision was made and clearly stated and repeated in corporate communication: “there will be no layoff’s due to the location strategy.” Naturally, this was only one piece in the change management puzzle, as many concerns remained amongst the employees: What will happen in the long run? What will the consequences be for my job situation in the short run? Can “they” deliver the same high quality “we” deliver? Is it effective to “spread” tasks? These questions, and many more, would have to be addressed in the change process. Furthermore, for the middle management responsible for carrying out the changes a number of political challenges existed with regards to individual power bases and interests.309

309

Snabe, Birgitte, Andreas Größler, and Peter M. Milling: “Policies and Politics of Establishing R&D Capacity in Low-Cost Locations”, Tagungsband, GWSTagung, Greifswald, in print, 2006, focus especially on the political challenges in the case study.

C. Case Study

2.

93

Intervention Process

The location strategy project was first initiated with a project team in each of the five business units in the division, but due to lack of consistency and efficiency, and lack of structure in the coordination and communication between the teams, it was decided to develop a shared, formal model on an abstract and highly aggregated level, aiming towards:

• creating a structured and ‘objective’ frame for the rather emotional and diverse discussions; • establishing a forum for exchange of experiences and best practices; • refining and making operational the strategy outlined by the board; • improving the change process effectiveness. The change imperative was stated as: “right now is the right time to hire people in the low-cost locations, because due to company growth, it can be done right now without staff reduction at high-cost locations, and the expected results are improved competitiveness and further company growth, also securing jobs at high-cost locations in the future.” This was a difficult message to communicate because of the fear of jobs moving from high-cost to low-cost locations. An external facilitator from Mannheim University (the author of this dissertation) was brought in as process coach and modeling facilitator, based on a participative modeling approach. For system dynamics modeling, the Vensim modeling environment was used. Targeting and planning the intervention initially involved a discussion with the project owners about the problem, the business objectives, the intervention objectives, the dynamic hypotheses, and the suitability of applying system dynamics to the problem. A preliminary model was built to frame the modeling project, and also for the project owners to feel confident that major results from the model were consistent with the directions of the strategy expressed in the objectives and targets. This preliminary model was the basis for the decision to move forward with modeling.

C. Case Study

BUS. AREA LEVEL

DIVISION LEVEL

94

1. 1. Vision Vision & Modeling Meta Modeling

3. 3. Consolidate, Consolidate, Balance Balance & & Coordinate

Agree on objectives & guiding principles

Reduce limitations/barriers

Agreement on shared, - formal model

Set division targets

Set and agree critical parameters

Balance between areas

Define ideal strategy

Identify critical success factors

Agree on communication strategy

2. 2. Area Area Strategy Strategy

4. 4. Detailed Detailed Planning Planning & Execution Execution

Evaluate ideal strategy against reality

Detailed planning

Define Area strategy based on ideal strategy

Align with budget process 2005

Identify limitations/barriers

Build 3 year plan (quarterly breakdown)

Business cases based on agreed parameters

Detailed communication plan

Identify short term wins

Figure C-2: Intervention process as communicated in the project

The modeling intervention was coordinated with other projects; most importantly the business planning and budgeting process, in which detailed Excel-models existed in each business unit. Figure C-2 illustrates the change process with visioning and modeling on division level combined with project and implementation ownership on business area level. The major stakeholders were identified early in the process, and the senior vice presidents responsible for the five business units were consulted about the plan for the system intervention in order to get their input as well as their commitment. The process was agreed upon, and the project organization was established. The major stakeholders are the management team of the division (Senior Vice Presidents each responsible for a R&D business unit), the project owners (division chief controller and the COO of the division), the core project team chosen to find the proper strategy (called location champions), corporate controlling, corporate management (the board of directors), the world-wide corporate location strategy responsible, and all the day-to-day managers with high influence on the implementation of the strategy. Furthermore, all employees of the company are stakeholders in a communication point of view. Project planning included the establishment of roles and responsibilities in the project; see table C-1 for an overview:

95

C. Case Study

Project Participants

Roles and Responsibilities

Division COO & Division Controller

Project owners - responsible for total system intervention - establishment of business objectives and targets - establishment of preliminary model - facilitators of all modeling sessions and meetings

Leaders of the Business Units (Senior Vice Presidents)

Decision-makers on division and business unit level - acceptance of the process, objectives and principles - approval of the parameter-settings - approval of the business cases and implementation plans

Business Unit project participants (Location champions)

Project core team - improvement of preliminary model and parametersetting - identification of most important parameters - establishment of detailed business cases - detailed planning of implementation

Managers with high influence on implementation

Change leaders in the implementation phase - involvement in implementation planning - execution of implementation

Location Strategy Corporate Coordinator

Corporate location strategy coordinator - coordination of company-wide parameter-settings - discussion on company-wide learning

Corporate Controlling & Management Board

Decision-makers on corporate level - corporate goals and strategies - approval of business cases

Table C-1: Roles and responsibilities as defined in the project

The solution design activities consisted of a larger number of meetings and workshops with a variety of agendas around the problem. Only approximately half of the activities directly involved modeling or simulation, but all meetings had some kind of impact on the model, its parameters and/or the process of implementing modeling results. The intervention facilitation role was split between the modeling facilitator and a discussion facilitator, who was one of the project owners, with in-dept knowledge of the company, the problem and of the intervention goals.

96

C. Case Study

The modeling and simulation process with the core project team served as a cognitive framework for objective discussions of the problem – challenging preconceived perceptions and aiming towards reducing the tendency often known from budgeting and business planning processes, that each stakeholder to some degree primarily looks after his or her business unit’s interests rather than corporate objectives. Using a shared model on a highly aggregated level moved the focus and discussion towards a holistic view. In this process some important new aspects – including one additional stock and a number of additional parameters – were added to the preliminary model. Also some parameters and relationships, which through simulations showed only little impact, were removed in order to simplify the model. For communication purposes, the modeling process was focused on developing a relatively simple model that could provide a picture of the overall behavior of the problem-system without including too many details, as overview was considered more important than detailed correctness. The focus on keeping the model simple was also due to the role of the system dynamics model role as a ‘meta-model’, creating cross-unit overview, and being complemented by a detailed excel-model in each business unit with the format needed in the budgeting and business planning process. The parameter setting was a cornerstone in the change-process, as these agreed parameters formed the basis for each business unit’s business case in phase 2 of the intervention process. The discussion of the parameters often resulted in a discussion on how the strategy could and should be executed, as the parameter setting reflected implementation decisions; e.g. the logistics and cost model of traveling, how to structure knowledge transfer, etc. Model testing was done partly behind the scenes by the modeling facilitator using some of the most respected sources as guides and checklists”,310 partly during the modeling and simulation efforts, as “validation is the process of establishing confidence in the soundness and usefulness of a model“.311 Also

310

311

Barlas, Yaman: “Formal aspects of model validity and validation in system dynamics”, System Dynamics Review, Vol. 12, No. 3, 1996, pp. 183—210; and Forrester, Jay W. and Peter Senge: “Tests for Building Confidence in System Dynamics Models”, in Legasto, Augusto A., Jay W. Forrester und James M. Lyneis (eds.), TIMS Studies in the Management Sciences, Vol. 14, Amsterdam, 1980, pp. 209—228. Forrester, Jay W. and Peter Senge: “Tests for Building Confidence in System Dynamics Models, 1980, p. 210.

C. Case Study

97

bearing in mind Forrester’s view on model validity: “Model validity is a relative matter. The usefulness of a mathematical simulation model should be judged in comparison with the mental image or other abstract model which could be used instead.”312 Through the simulation of different scenarios the discussion focused on the most influential parameters and causal relations of the problem, and possible improvement options. Based on the modeling and simulation, a presentation with the most important learning was developed to document the major insight reached by the core project team. This presentation, together with the Vensim model and the business unit excel models, was used in communicating with the other stakeholders: to both communicate modeling results and to receive the stakeholders input on the model, the parameters and the insights gained. In the last workshop of the case study project, the core team established a set of critical success factors for the implementation. The concept of critical success factors was originally introduced by Rockart as a method for aligning business and IT strategy, but has developed to a broader usage on both strategic levels and in organizational interventions.313 The critical success factors served as feedback to the Board. One of the critical success factors emphasized the need for cost-orientation rather than head-count orientation in the company governance model. Another critical success factor concerned the crossorganizational coordination of the detailed implementation plans per business (also involving managers from outside of the division). The detailed implementation planning included determination of R&D tasks to be moved from each business unit in high-cost countries and furthermore was explicit and specific about the future R&D tasks of the affected employees. The detailed implementation plan, and a detailed communication strategy and plan, were considered to be the cornerstones in securing motivation and morale among employees. The implementation plan furthermore included clear delegation of responsibility for improvement of the parameters identified to be the most

312 313

Forrester, Jay W.: Principles of Systems, Cambridge, 1968, chapter 3, p. 4. Rockart, John F.: “Chief executives define their own data needs”, Harvard Business Review, March-April 1979, pp. 81—93. For a discussion on the usage of critical success factors, see also Snabe, Birgitte and Jakob F. Ølgaard: Informationsstrategi, Master Thesis at The Institute of Mathematical Statistics and Operations Research, Technical University of Denmark, Lyngby, 1992, pp. 25—30.

C. Case Study

98

influential ones. Reviews and follow-ups were integrated in the existing performance measurement system.

3.

The Model and Selected Simulations

The modeling purpose was to find an effective and realistic plan for reducing the cost/capacity ratio under board guidelines of cost growth only due to inflation in high-cost and low-cost countries, and an increase from 10% to 25% of low-cost headcount compared to the total headcount of the division. The basic structure of the model is based on two separate aging chains (see figure C-3), each being an extended version of Sterman’s “two-level promotion chain”.314 The aging chain on the right represents the high-cost locations (abbreviated HC); the aging chain on the left represents the low-cost locations (abbreviated LC). In each aging chain this model had originally three basic stages: “New hired FTE”, who are newly hired employees (FTE = Full Time Equivalents) spending their time in class-room training learning the development tools of the company; “Rookie FTE”, who are employees working on development projects with reduced productivity, and then finally “Productive FTE”, who are fully productive employees.315 In the low-cost aging chain, an additional stock was added: a stock containing Rookies spending time on taking

314

315

See Sterman, John D.: Business Dynamics, Boston, 2000, p. 491. Sterman operates with only two levels in his promotion chain, with employees leaving both levels. The location strategy model operates with 3 and 4 levels, with employees only leaving the latest stage as this reflects the historical data well. Other examples of system dynamics models of aging chains include Martinez, Ignacio J. and Luis F Luna: “The Dynamics of Best Practices: A Structural Approach”, at CD-ROM of Proceedings, System Dynamics Conference, System Dynamics Society, 2001, which operates with three levels of practitioners: junior, intermediate, and advanced. In the model as well as the case description, the abbreviation FTE (Full Time Equivalents) is used to indicate number of full-time employees. The abbreviation is an established term in the case study company, used to adjust budget numbers for employees working part-time.

LC 2a

HAND-OVER FRAKTION OF LC ROOKIE TIME

LC TOTAL ROOKIE TIME

handover capacity reduction

Rookie FTE LC

LC ONGOING TRAVEL COST

quit Productive FTE LC

LC

LC QUIT FRACTION

handover travel costs

Figure C-3: The location strategy model



cost per month

HC TRAINING COST

HC PERSON COST INCREASE RATE

New hired HC FTE HC trained

HC hire



REPLACEMENT IN HC VS. LC

HC TRAINING TIME

total FTE HC

job-trained

Rookie Productive FTE HC HC FTE HC

HC ROOKIE TIME

HC ROOKIE PRODUCTIVITY

HC CAPACITY USE ON HAND-OVER

LC person cost HC person cost

LC HAND-OVER TRAVEL COST

LC TRAINING COST

quit

HC

HC QUIT FRACTION

production per month

LC PERSON COST INCREASE RATE

LC 2b

LC 1c

LC ROOKIE PRODUCTIVITY

LC PRODUCTIVITY REDUCTION

total FTE LC

H-O-R LC 1b FTE LC



LC TRAINING TIME



LC repl. hire

LC flow input

LC

New hired 1a FTE LC

LC new hire

ADDITIONAL GROWTH

LC replacing HC quit



C. Case Study 99

C. Case Study

100

over tasks from high-cost countries, which will be the case for all new employees in low-cost countries who are not merely replacing people who have left a position at a low-cost location. This stock is called H-O-R FTE (Hand-OverRookies). These employees have zero productivity and, as they are spending time physically with those HC employees, whose tasks they are taking over, they furthermore tax time from fully productive employees in high-cost countries in the process of knowledge transfer. The two main decision points influencing the changes in stock levels in the model are: 1.

The rate “HC hire”, where only a share of the employees leaving highcost locations will be replaced at a high-cost location, depending on the factor called REPLACEMENT OF HC VS. LC in the model. Those headcounts not being replaced in HC will be replaced in LC.316

2.

The rate “LC new hire”, consisting partly of replacement of employees leaving the stock of productive LC employees, partly of headcounts not being replaced in HC, and partly of the additional new employees joining the division (called ADDITIONAL GROWTH in the model). All additional new employees are allocated to low-cost countries, based on a growth factor relative to the stock of experienced employees in lowcost. 317

It is important to note that tasks will be moved from high-cost to low-cost locations in bulk. Employees in high-cost locations, who have transferred their tasks to low-cost locations, will either take over tasks from a person leaving the division or take part in new R&D projects within the division. Tasks moved to low-cost locations enable new and innovative tasks in high-cost locations; i.e. does not lower productivity in high-cost locations.

316

317

Due to the non-replacements, the number of employees can be reduced in some locations. However, the total number of employees in high-cost locations will stay roughly stable due to necessary hiring in other functions in the division. In Sterman, John D.: Business Dynamics, Boston, 2000, p. 491, the “two-level promotion chain” has a growth factor related to the total number of employees, but in the location strategy model it makes more sense to base the growth on fully productive FTE’s, due to ramp-up limitations (ratio between experienced staff and new staff).

C. Case Study

101

Based on parameter settings, the model will calculate the development in stock levels as well as the two central model outputs: cost per month and production per month. For a detailed description of model parameters, see appendix A. For model equations, see appendix B. The stocks in the model are initialized in equilibrium (hire rate = quit rate for both low-cost and high-cost locations) through a distribution of the total number of employees for both low-cost and high-cost countries to their respective stocks of newly hired, rookies and experienced FTE’s. The number of newly hired employees, plus rookie employees, plus experienced employees, equals the total number of employees. The distribution into these three categories is a calculation based on quit rate, training time and rookie time. For further description on stock initializations, see appendix C. In the modeling process, the facilitation and communication function was prioritized over model correctness; especially in terms of using parameters directly useful in the budgeting process, and also in keeping the model as small as possible to avoid a “black box” effect. The preliminary model (see appendix D) had “set the stage” for the main trends of the model, and the model refinement process had the dual objective of getting acceptance in the model as well as adjusting and improving the model. Model-wise, the process included adding 1 new stock (the hand-over-rookies), a few new causal relations (centered on productivity reductions due to handover tasks as well as low experience level in low-cost locations), and removing a number of parameters less significant for investigating the decision parameters. The removing of the parameters was done to simplify the model.318 A number of design decisions had to be made in the modeling process. The growth rate ADDITIONAL GROWTH could have been modeled reflecting a goal-seeking structure based on the discrepancy between the actual value and the goal-value for the fraction of LC employees out of total number employee.319

318

319

Examples of removed parameters: the number of productive days is higher in LC is than in HC, but these parameters were removed from the model (and added in a comment instead), as it was decided to let the higher productivity in HC equals out with additional coordination overhead. Also the cost of hire was removed from the model, as they proved to be insignificant in simulations. In this case, it should be used in combination with a MIN-function in the LC new hire-rate to secure that the number of newly hired employees did not exceed

102

C. Case Study

This would have made the policy of how many additional employees to hire endogenous.320 However, in order to keep the model simple and with focus on the few, most important parameters, the growth rate is simply implemented as a constant (fixed for the first period, then gradually decreasing to zero after the 36th month). For the same reasons, as well as to allow some degree of cost overrun, it was decided not to handle the cost restrictions endogenously, but to incorporate cost as an explicit auxiliary. In Sterman’s Two-Level Promotion Chain, the rate of employees moving from a ‘rookie’ state to an ‘experienced’ state was modeled as a fraction of the number of Rookies321. In the location strategy model the rates between stages are calculated as delay-functions of the inflow-rates. For the training period, a highorder delay was used to imitate a pipeline delay, as this is a fixed period of time for each employee. For the period of being a rookie, a lower order delay was used, to reflect the variability in the learning curve for individuals. The variability in the difficulties of the task areas is not explicit in the model, but is considered in stipulating the average time for employees being rookies. Having outflow-rates as a function of inflow-rates (rate-on-rate modeling) places the case study model in a slight conflict with the tendency among system dynamic practitioners to avoid rate-on-rate modeling, and rather calculate outflow-rates as a fraction of stock-level. The reason for the tendency to avoid rate-on-rate modeling can perhaps be seen partly as motivated by prevailing attention to macro-trends of systems with continuous parameter development (rather than systems with steps in inputs), and partly motivated by a historical tradition based on Forrester’s “Principles of System”, which can be dated back to a time where delay-functions constituted a seriously challenging workload for computers.322 For the location strategy model, the new strategy with a step in the hiring rate at the very beginning is being investigated, and the use of the delay-function better reflects the true pattern of employee-flows in the start-up period. Due to the use of rate-on-rate modeling, a fixed period of time will pass before the first

320

321

322

a realistic level compared to number of experienced employees (the ramp-up limitations). See chapter B.III.1 for a discussion on endogenous and exogenous explanations for system behavior. Sterman, John D.: Business Dynamics: Systems Thinking and Modeling for a Complex World, Boston, 2000, p. 491. See Forrester, Jay W.: Principles of Systems, Cambridge, 1968, chapter 5, p. 10.

C. Case Study

103

additional new employees flow into the rookie state (e.g. 3 months) and an even longer time will pass (e.g. 9 months) before the number of experienced employees is beginning to increase. This is intuitively more acceptable for group modeling participants, as they are not trained in system dynamics modeling, and therefore could have trouble abstracting from the number of experienced employees beginning to increase as early as the very first month. Interestingly enough, though, the major trends and learning – even regarding year 1 - are the same with both ways of modeling the delays, see appendix E. The figures C-4, C-5 and C-6 show the behavior of selected variables for four simulation runs. Naturally, a large number of simulations took place in the workshops to investigate the importance of the influence of the different parameter settings and the consequences of different scenarios, but the selected simulation runs give an impression of some of the key learning from the model. The selected simulation runs are labeled: “INI”, “Base run”, “40% HC replacement”, and “Faster training and hand-over”. Below is a short description of the parameter setting in each of the simulations:

•

INI: In this run, no changes ot the number of employees in low-cost or high-cost take place (hire rate = quit rate in both aging chains). The simulation run therefore reflects the development in costs and production as could be expected in a zero-growth situation, based on stipulation of the cost and production related input parameters. The production naturally stays constant, whereas the costs increase due to the expected increases in employee costs.

•

Base run: This simulation run reflects a parameter setting with temporary, strongly reduced replacement hiring in high-cost, in order to finance the build-up of resources in low-cost locations. Compared to INI, two principal changes are made: The REPLACEMENT IN HC VS. LC is set to 0.2 in the first 36 months; where after it is set to 1, indicating full replacement of employees leaving. Furthermore, the ADDITIONAL GROWTH in low-cost locations is set to = 3% per month for the first year (equals more than 40% ramp-up in year 1), then linear decreasing to 0 after 36 months. The efforts not being used to train new people in highcost allow the total new hiring in low-cost to be higher than the additional growth (due to high-cost employees job-training new low-cost hires rather than new high-cost hires).

C. Case Study

104

•

40% HC replacement: The parameter setting in this run is only slightly changed compared to the INI simulation run. The only change is that the REPLACEMENT IN HC VS. LC is set to 0.4 in the first 36 months; as opposed to 0.2. This change is made to investigate the consequences of allowing more external people to be hired in high-cost locations compared to the base run.

•

Faster training and hand-over: The parameter setting in this run is also very like the parameter setting in the Base run. The only difference is that the classroom training is reduced from three months to two months for low-cost new hires (requiring a more intensive training program), as well as reducing the fraction of the rookie time to be spent 1-on-1 to handover tasks with high-cost employees. This fraction is reduced from 50% (equaling 3 months) to 33% (equaling 2 months), and this will require that high-cost employees also have to work more intensively in the hand-over period and it also calls for an improved process for the hand-over of tasks.

The following three figures show major results of the four selected runs:

LC FTE fraction of total FTE 0.4 0.2

4 2 3 4 2 3 1 2 3 1 1

1

1

18

24

2

2 3 4

2 3 4

23 4

4

1

1

3

2

4

3

2

4

1

3

2

4

1

3

2

4

1

1

0 0

6

12

30 36 Months

42

48

1 1 1 1 1 1 1 INI 2 2 2 2 2 2 2 Base run 40% HC replacement 3 3 3 3 3 3 Faster training and hand-over 4 4 4 4 4

54 Dmnl Dmnl Dmnl Dmnl

Figure C-4: Fraction of employees in low-cost countries compared to total number of employees in the division

60

105

C. Case Study

Index Cost per Production 1.2 1.15 1.1

1 3 23 2 3 2 3 2 3 2 34 3 1 3 1 2 3 23 2 4 4 3 2 3 4 4 2 4 4 1 4 4 4 4 2

1.05 1

1

1

1

1

1

1

6

0

12

18

24

30 36 Months

42

1

1

1

1

2 34

34 234 2

48

23

54

1 1 1 1 1 1 1 1 1 1 1 INI 2 2 2 2 2 2 2 2 2 2 2 Base run 3 3 3 3 3 3 3 3 3 40% HC replacement Faster training and hand-over 4 4 4 4 4 4 4 4

60

Dmnl Dmnl Dmnl Dmnl

Figure C-5: Development in cost-index for an average productive unit (e.g. cost for one employee for a fixed period)

Index Production per Month 1.1 4

4 2 3

4 4 3 2

4 3

4 4

4

1.05

4

4 4

4

1

4 3 2

1 2

2

3

2

3

2

3

2 1

1

1

1

24

30 Months

1

1

1

1

1

1

1

1

4

0

6

12

18

36

42

48

54

INI 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Base run 2 2 2 2 2 2 2 2 2 2 2 2 2 2 40% HC replacement 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 Faster training and hand-over

Figure C-6: Development of productivity index for an average unit (e.g. output/month/unit)

Dmnl Dmnl Dmnl Dmnl

60

106

C. Case Study

To some extent, interpreting model results is always a subjective exercise. This is especially true of the scenario simulation in this case, as the stipulation of how much each input-parameter can realistically vary is a matter of argumentation and consensus within the group of participants. The following is a description of the core group’s understanding of key model results:

•

Hand-over efforts have a strong “worse-before-better” effect on productivity, and action must be taken to optimize this process—even if this results in higher travel costs. Although this insight did not have much ‘newness’ value to the business units, it was very valuable to have a model that distinctly and clearly ‘proved’ the matter.323 A workshop was arranged with corporate controlling, to make the point clear that even though the division receives a relatively large number of additional head-counts in year 1, the division will have no additional productivity in year 1, but rather a slightly reduced productivity due to hand-over efforts.

•

Not replacing all employees in high-cost locations for a three-year period is the only practical way to build up resources at low-cost locations in a (rather) cost-neutral scenario. This was a politically sensitive discussion. The model helped to make the discussion more objective compared to the very emotional and unstructured discussions taking place regarding this topic before the modeling part of the project was introduced.

•

Reducing training time has an accumulative productivity effect. With the large number of new employees, investments should be made to optimize their training—even if this results in higher training and travel costs.

The list above concentrates on the insights directly supported by model simulation. Additionally, learning and changing of attitudes and intentions also

323

The “worse-before-better“ effect is widely recognized in change management literature. The effect is also illustrated by the means of system dynamics models, e.g. Repenning, Nelson P. and John D. Sterman: Nobody Ever Gets Credit for Fixing Problems that Never Happened: Creating and Sustaining Process Improvement”, California Management Review, Vol.43, No.4, Summer 2001, p. 74.

C. Case Study

107

took place in a more indirect way due to discussions during model refinement and parameter setting. Examples are:324

324

•

The alignment of actual numbers to be used in the business cases (also called the Excel-models). The business cases were often large ‘blackbox’ models. By having a system dynamics model on a highly aggregated level, the most important parameters were made transparent. An example of this is the employee turnover rate in low cost countries that in the different business cases was set to be in the range of 10% to 25% based on personal assumptions, although historical figures showed to be approximately 7%.

•

The strategy could be accelerated by a shift in the entire company from headcount orientation towards cost orientation.

•

Identification of detailed transfer planning as a critical success factor. The after-transfer solution for involved high-cost location employees was perceived to be especially important in order to secure motivation and morale. This included planning for replacements in high-cost locations being made by colleagues who have handed over tasks to lowcost locations.

•

A number of best practice experiences were exchanged between the business units because some units already had more experience in building up capacity in low-cost locations. The modeling sessions also served as a forum for the discussion of topics such as bridgeheads, hiring strategies, and practical aspects of making new employees productive as quickly as possible.

Further examples are discussed in chapter D.

108

C. Case Study

III. 1.

Evaluation of the Case Study A Framework for Evaluating the Effectiveness and Efficiency of the Case Study

Being a real-time single site case, with no test group for comparison, means that even if the intervention yields good results, it is impossible to know whether other intervention mechanisms would have yielded even better results. The evaluation framework takes the theoretical point of departure in the evaluation frameworks developed by Huz and Rouvette; and is adapted to the research focus and the data collection conditions of the case study.325 The evaluation is structured in two sections: evaluation of outcomes, as well as evaluation of method and comparative conditions that may explain the intervention’s effectiveness. Evaluation of outcome:

• • •

Evaluation of outcome on an individual level Evaluation of outcome on a group level Evaluation of outcome on an organizational level

Evaluation of method and comparative conditions:

•

Method evaluation of the use of system dynamics modeling compared to other approaches • Method evaluation of the specifics of modeling for strategy implementation compared to other system dynamics modeling purposes • Comparative conditions that may explain the intervention’s effectiveness (both context comparative conditions and mechanism comparative conditions) The evaluation of both the outcome and method is relevant to identifying the benefits of the intervention. It gives some guidance for the effectiveness of the applied method, but - because of the nature of a seek-and-learn case study

325

Huz, Steven, David F. Andersen, George P. Richardsen and Roger Boothroyd: “A framework for evaluating systems thinking interventions: an experimental approach to mental health system change”, System Dynamics Review, Vol. 13, No. 2, 1997, pp. 149—169; Rouwette, Etiënne: Group model building as mutual persuasion, The Netherlands, 2003, pp. 68—95. The framework of Rouwette partly is based upon the work of Huz.

C. Case Study

109

with no test group - it will not give data to support or reject hypotheses about relative effectiveness compared to alternative mechanisms. The value of the evaluation is therefore strongly connected with the search for explanations and a focus on the question of why certain outcomes might happen. Data on the evaluation variables was collected from three sources: (1) observations by the facilitator, (2) semi-structured interviews, and (3) questionnaires among both core team members and non-core project participants. The usefulness of the sources differs between the measurement variables, but each source adds some information to all variables regarding outcome and method. In the analysis, the appropriateness of the different sources for each measure variable is taken into account. For observations and interviews, focus was especially on potential manipulation and biases.326 Significant bias must be expected to occur due to the mental model of the observer and interviewer (the author of this dissertation).327 The usage of a structured evaluation framework lower the bias, but the mental model will inevitable be a “filter” in the selection and interpretation of observations and interview results. However, as the purpose of the case study is theory building rather than theory testing, the bias is less crucial for the research value. The questionnaires are closely linked to the evaluation framework, and are designed to be answered anonymously. They are furthermore short (1 page), in order to increase response rate, and they aim to provide unambiguous and objective questions about participants opinions regarding both intervention output and mechanism on a 1 to 7 scale (from strongly disagree to strongly agree, 4 being neutral). The questions refer directly to the measure variables in the evaluation framework, to reduce bias due to the researcher’s own pre-coded view of the research. In addition, the use of checklists for how to use and design questionnaires was applied in an attempt to reduced bias.328 The questionnaires 326

327

328

Yin, Robert K.: Case study Research, 3 rd edition, Thousand Oaks, 2003, pp. 93—96, discusses the problems of participant observations. Kvale, Steinar: InterViews: An Introduction to Qualitative Research Interviewing, Thousand Oaks, 1996, pp. 235—252, discusses the reliability and validity of interviews. For discussions on self-reinforcing of mental models, see Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, 1993. pp. 29—30; Argyris, Chris: Reasoning, Learning, and Action – Individual and Organizational, San Francisco, 1982, p. 39. Denscombe, Martyn: The Good Research Guide for small-scale social research projects, Philadelphia, 2 nd edition, 2003, chapter 9.

110

C. Case Study

were pre-tested with both a project owner and a research colleague. The questionnaire was given to the five core-team members, to the three most involved steering committee members, and to three other participants, who were not as involved in the project, but had only been exposed to the model in one or two meetings. The latter group received a reduced version of the questionnaire, as they did not have an overview of the complete project. The interviews are structured around the same measure variables as the questionnaires, but with open questions, and the interviews were conducted with the two project owners. The facilitator’s own observations are also decoded in the same structure as the questionnaires and the interviews. The following tables (C-2a to C-2e) illustrate the applied framework’s sources for data collection.

111

C. Case Study

Measure variables

Questionnaire self assessments

Personal reactions “I believe it was to the modeling useful to include the process model in the project” (see also under method)

Interviews with project owners

Observations

Project owners’ perception of participants attitude towards the process

Attitudes in modeling sessions and other meetings

Learning gained, and changes in goal structures and mental models

“I gained interesting learning from the model”

Project owners’ perception of insights gained by individuals

Changes in positions in the discussions (and pre/post tests of change ambassadors)

Commitment to the outcome of the modeling sessions

“I agree with the recommendations derived from the model – and will act accordingly”

Project owners’ perception of the commitment among participants

Whether participants in subsequent meetings actively argued for the results

Changes in behavior

“The modeling affected some of my decisions”

Project owners’ perception of the business case’s alignment with the modeling results

Observations regarding behavior

Table C-2a: Main sources for evaluation of outcomes on individual level

112

C. Case Study

Measure variables

Questionnaire self assessments

Interviews with project owners

Observations

Group communication

“The meta-model was a useful framework facilitating discussions”

Project owners’ perception of the communication during modeling sessions

Whether the modeling process created open discussions and exchange of views

Consensus

“The modeling process helped building a shared view of the location strategy”

Project owners’ perception of group consensus established through the modeling process

Whether the group seemed to get closer in opinions regarding the strategy

Common language “The modeling efforts helped creating a common language for the location strategy”

Project owners’ perception of creation of a common language through the modeling sessions

Agreement on using the same terms – also outside the modeling sessions

Transfer of insights

Project owners’ perception of the usefulness of the model in transfer of insights

Effectiveness in transfer of insights to noncore project participants

“The meta-model was a useful tool in the presentation of the ideal location strategy”

Table C-2b: Main sources for evaluation of outcomes on group level

113

C. Case Study

Interviews with project owners

Observations

System changes “I believe the recommendations from the modeling process will be implemented”

Project owners’ perception of the boards reaction to the recommendations

If decision is included in budgets and overall business plans

Results

Project owners’ expectations regarding business benefits

Business results

Measure variables

Questionnaire self assessments

“I believe the recommendations will have positive business impact”

Table C-2c: Main sources for evaluation of outcomes on organization level

114

C. Case Study

Observations

Measure variables

Questionnaire self assessments

Interviews with project owners

Efficiency (compared to normal project execution in the case company)

“The use of modeling increased the efficiency of the project process”

Project owners’ n.a. perception of the efficiency of the process in general

Efficiency (compared to other approaches or methods)

“Using modeling in this case was more efficient compared to other approaches”

Project owners’ perception of the efficiency – compared to if other approaches had been used

Project progress compared to other types of consulting approaches (highly subjective)

Quality in results

“Using modeling in this case created higher quality results compared to other approaches”

Project owners’ perception of the quality of the results compared to if other approaches had been used

The importance of insights gained (highly subjective)

Project owners’ perception of the general trust in the model

Later actual system dynamics projects

Further use of SD “I intend to use modeling in other relevant change projects”

Table C-2d: Main sources for evaluation of system dynamics compared to other approaches

115

C. Case Study

Interviews with project owners

Observations

Intervention driven (not included) by business objectives and targets

Project owners’ perception of the importance of initial business objectives and targets

How the initially established objectives and targets influenced the process

Project framing (effectiveness)

“It was useful to start with a 1st draft of the model to kick off the process”

Project owners’ perception of the usage of a preliminary model

Possible conflicts concerning model boundaries

Project framing (consequences for model trust and ownership)

“I believe the model reflects the core of the problem”

Project owners’ perception of the project participants’ and own trust & ownership

Attitudes in modeling sessions and other meetings

Structured involvement of implementers

(not included)

Project owners’ perception of the importance

Implementers influence on the intervention

Measure variables

Questionnaire self assessments

Table C-2e: Main sources for evaluation of the usage of system dynamics in a change management context

Evaluation of comparative conditions that may explain intervention effectiveness (both context comparative conditions and mechanism comparative conditions) will only depend on observation, as neither questionnaire nor interviews included this.

116

2.

C. Case Study

Conclusions on Case Study Effectiveness and Efficiency

A schematic overview of observations made during the intervention as well as results from the interviews are to be found in appendix F. The case study focused on the role of a system dynamics modeling process used in bridging between the top-management launching a new location strategy, and the planning and implementation of the new strategy. A highly aggregated system dynamics model proved useful for this purpose, and participants accepted it as an abstract representation of the subject matter. This model was preliminary established by the project owners (representatives for the decision-makers), and later elaborated and improved during the participative modeling process. It has to be clarified that in the case study, the actual strategy decision was made without using system dynamics modeling and the model was built post hoc in order to secure that model outcomes are broadly consistent with these decisions. The modeling process allowed participants to exchange their experiences and ideas in a relatively objective way, i.e. without relying too much on gut feelings, and supplemented the discussions on the change imperative, reducing most of the participants’ initial fears concerning the location strategy. Simultaneously, a common ground for discussions was laid by recurring on a formal model that was open for inspection, critique and change. Within the broad strategic objectives set by the board, participants of the modeling process could refine and alter the actual policies that resulted from the strategy. Participants’ reactions in the evaluation of the case intervention emphasized that they acknowledged this possibility. Another remark was that participants appreciated the rather ‘politics-free’ atmosphere that was achieved by working with a system dynamics model. This can be interpreted as a hint that members of organizations are willing to accept strategies even when their results seem to contradict their initial personal wishes as long as the process is transparent and some sort of involvement is secured. Especially due to the sensitive nature of the topic, it was interesting to observe how the modeling and simulation efforts helped to direct the discussions and facilitate individual learning. However, especially in the beginning of the project, there was a tendency among core team members to think of modeling (including quantification) as an additional task, increasing the workload in an already stressed period. Project participants clearly experienced single-loop learning, and indications of double loop learning included changes in attitudes towards the strategy and underlying assumptions. From the case company perspective, it

C. Case Study

117

seems fair to conclude that the modeling process was an effective and efficient way of refining a strategy that fulfils board objectives as well as preparing the grounds for sustainable implementation. The most visible indication for the individual’s commitment to changes as well as changes in actual behavior might be the fact, that the year results for year 2005 even show an over-performance with regards to the movement of tasks from high-cost locations to low cost locations. This despite of the fact, that shortly before the modeling project, the people responsible for the hiring had showed significant resistance to the process. The project owners in particular emphasize how the project progressed satisfactorily, changing focus from ‘seeing only problems’ to discussing sustainable and fair execution, and how it resulted in the expected deliveries based on structured and also vivid and honest discussions. As a conclusion based on interviews and observations, the intervention leaves a positive impression with regards to both effectiveness and efficiency, although it is not possible to compare the intervention with a similar intervention using an alternative approach. On a detailed level, it is not relevant to even try to draw conclusions, as this cannot be justified based on the research methodology. The purpose of the case study in the first place was to serve as an inspiration for literature studies as well as for the observations to serve as illustrations in the discussion in chapter D. Table C-3 reports on results from the questionnaire source of the evaluation. In general, the questionnaire confirms the positive impression based on observations and interviews. The three most positive responses, on average, were if it was believed useful to include the model in the project, the usefulness in facilitating discussions and the usefulness of starting the project with a preliminary model. In addition, individual learning and the building of a shared view scored relatively high. Questions about expected implementations scored relatively low, which in follow-up discussions in a later workshop and informal talks were explained with a lack of trust in the assumptions of the new strategy. The strategy reflects a new paradigm with focus on cost rather than headcount, which is a change from the previous situation with rather strict headcount control. Furthermore, the intended use of modeling in other projects scored relatively low, which can be explained by the project’s lack of explicit focus on transfer of system dynamics skills and understanding, as learning efforts were concentrated on the location strategy issues.

C. Case Study

118

Question I believe it was useful to include the model in the project I gained interesting learning from the model I agree with the recommendations derived from the model and will act accordingly The modeling affected some of my decisions The model was a useful framework facilitating discussions The modeling process helped building a shared view of the location strategy The modeling efforts helped creating a common language for the location strategy The model was a useful tool in the presentation of the ideal location strategy I believe the recommendations from the modeling process will be implemented I believe the recommendations will have positive business impact The use of modeling increased the efficiency of the project process Using modeling in this case was more efficient compared to other approaches Using modeling in this case created higher quality results compared to other approaches I intend to use modeling in other change projects I believe the model reflects the core of the problem It was useful to start the modeling with a preliminary model to kick-start the process Table C-3: Questionnaire results

n

mean

sd

8

6.00

0.76

11

5.82

0.87

8

5.00

1.41

8

4.63

1.85

11

6.18

0.87

11

5.73

0.90

11

5.27

1.01

11

5.09

1.14

11

4.27

1.19

11

4.91

1.22

8

5.13

1.13

8

5.13

1.13

8

5.38

0.92

8

5.00

0.93

8

5.38

1.19

8

6.25

0.71

D. The Usage and Utility of Participative Modeling in Change Management While system dynamics modeling is traditionally used in exploratory organizational interventions, the core of this dissertation is to investigate the usefulness of system dynamics modeling, when the purpose is to implement an already outlined strategic initiative. Regardless of whether a strategic initiative is derived from a specific strategic problem-solving process or from the ongoing strategy-forming process of the company, the implementation faces the challenges of change management.329 In this chapter, the discussion of participative modeling efforts used in supporting change management is structured into three sections: discussion of intervention context, discussion of the intervention process, and discussion of intervention outcomes.330 The intervention context is understood primarily as problem and organizational characteristics. The intervention process is described in terms of business objectives and targets, structured development of change leaders, the design of the change process, and facilitation of modeling and simulation sessions. Finally, the intervention outcomes are discussed in terms of transfer of insights and ownership from decision-makers to implementers, refining and aligning strategies through scenario testing, and organizational learning. Examples from the case study will be used as examples in the discussions. These examples are not included as a means of verifications of any point made, but merely serve as illustrations.

329

330

In Warren, Kim: “Improving strategic management with the fundamental principles of system dynamics”, System Dynamics Review, Vol. 21, No. 4, Winter 2005, p. 329, differentiation of strategic management is made between one-off challenges and the continuous direction of enterprise strategy. The ‘context – process – outcome’ structure is inspired by the ‘context – mechanism – outcome’ configurator as described Rouwette, Etiënne: Group model building as mutual persuasion, Nijmegen, 2003, pp. 87—92; and also by the ‘intervention – organization – effect’ model (a variant of a classical S-O-R model) as described in Borum, Finn: Strategier for organisationsændringer, Copenhagen, 1995, p. 56.

120

D. The Usage and Utility of Participative Modeling in Change Management

I.

Context Factors Relevant for Deciding on Usage of Modeling in Change Management

Implementing strategic initiatives calls for decisions on change strategies and resulting change management approaches.331 Palmer and Dunford discuss change management approaches in terms of controlling vs. shaping change.332 It seems reasonable to assume that modeling is ill suited for change approaches primarily focused on controlling, as a modeling process is a learning process allowing for involvement and influence. For change approaches aiming at shaping intended change outcomes, the usage of system dynamics modeling offers a way to involve key implementers with the dual purpose of transferring insights from decision makers to implementers, and having the implementers refine the implementation process. Anderson and Anderson discuss the need for the creation of change leaders when starting organizational changes, addressing the issue of change process to be something that managers have to lead rather than manage.333 Leavitt discusses the dual role of middle management as managers (respecting hierarchical decisions) and leaders (leading people).334 The creation of change leaders requires (on top of a number of personal requirements such as communication skills, interpersonal skills, etc.) establishment of internal commitment amongst the people responsible for implementing the change. Creation of such commitment is often centered on the creation of awareness, consensus, and confidence regarding the goals and the change process.335 Palmer and Dunford refer to approaches aiming at shaping intended change outcomes for ‘change management as coaching’, which is the change 331 332

333

334

335

For a discussion on change strategy taxonomies, see chapter A.II. Palmer, Ian and Richard Dunford: “Who says change can be managed? Positions, perspectives and problematics”, Strategic Change, Volume 11, August 2002, p. 244. Anderson, Linda A. and Dean Anderson: “Awake at the Wheel: Moving beyond Change Management to Conscious Change Leadership”, OD Practitioner, Vol. 33, No. 3, 2001, p. 45. Leavitt, Harold J.: Top Down – Why Hierarchies Are Here to Stay and How to Manage Them More Effectively, Boston, 2005, pp. 164—168. According to Akkermans, Henk: Modelling With Managers, Breda, p. 20, the efforts of establishing internal commitment in organizational interventions are often centered on the creation of awareness, consensus, and confidence regarding the goals and the change process

D. The Usage and Utility of Participative Modeling in Change Management

121

approach that overlaps most with organizational development and action research, and the strategy typically recommended in the prescriptive, normative management literature on change management.336 They also introduce a change approach called ‘change management as interpreting’ dealing with shaping partially intended outcome, placing managers in the role of sense-makers in organizational environments where different meanings and related change intentions are competing to be implemented. Winch and Derrick discuss the interpreting type of intervention dealing with knotty problems, where not only the problem or system to be studied is characterized by structural and dynamic complexity, but the complexity also characterizes the intervention itself.337 For this change approach, the usage of system dynamics modeling has a stronger traditional exploratory focus, compared to usage in coaching approaches aimed at transforming the managers responsible for the implementation into change leaders pursuing already outlined targets. Besides the change approach, the characteristics of the problem at hand also determine the appropriateness of using system dynamics modeling. The model in the case study is a simple system dynamics model primarily focused on the problematics of delays in aging chains, displaying no significant feedback loops. The insights gained in the project were only to a small extent directly related to the behavior of the model. The main part of the learning and consensus building took place as a result of the process around the model building, e.g. the quantification and the discussions of the importance of the different parameters leading to discussions of best practices. It is difficult to know, however, if problems requiring much more complex models can also benefit from using

336

337

Although action research has it origin from micro-organizational issues with much of the earlier literature primarily discussing changes of group behavior, the field of organizational development has included a whole range of techniques to adjust the theories of action research to be suitable for large-scale changes in organizations. Practically all of the literature from Lewin, Argyris and Schein used in this dissertation discuss mostly interventions on group level. For the evolution of action research in the organizational development literature, see Palmer, Ian and Richard Dunford: “Who says change can be managed? Positions, perspectives and problematics”, Strategic Change, Volume 11, August 2002, p. 247. Winch, Graham and Sonja Derrick: “Flexible Study Processes in ‘Knotty’ System Dynamics Projects”, Journal of Systems Research and Behavioral Science, Vol. 23, No. 4, in print, 2006, pp. 1—2.

122

D. The Usage and Utility of Participative Modeling in Change Management

system dynamics modeling in a change management perspective, although it seems fair to assume that such complexity will result in a number of challenges regarding barriers for participants to understand the model as well as the greater time investment needed. Through their practical consulting experiences in group model building Eskinasi and Fokkema have observed what they call ‘a striking correlation’ between a high level of details in the model and client disappointment and lack of success.338 Eskinasi and Fokkema’s experiences deal with models in exploratory settings rather than change management settings, which might even be a more acute argument for using simple models in change management, as participants would be expected to accept more involvement in projects dedicated to a problem to be explored. Although indications exist for preferring relatively simple models in change management purposes, one should not forget the basic argument for using system dynamics in the first place: the usage of modeling to compensate for the human brain’s difficulties in coping with feed-back structures and delays.339 For more complex problems, it could be interesting to investigate whether a combination of an exploratory modeling cycle involving top executives, and a change management modeling cycle involving the key implementers, would allow for improved sustainability in problem-solving for more complex problems. It is also important to assess the organizational situation to determine the appropriateness of using modeling. Flood and Jackson discuss the industrial relations literature on the political relationships in business organizations between individuals and groups using the three metaphors: unitary, pluralist and coercive relationships.340 These three metaphors are useful when assessing the

338

339

340

Eskinasi, Martijn and Eppie Fokkema: “Bursting the myth of making easy modeling money”, unpublished working paper (permission granted from the authors), Proceedings, Second European System Dynamics Workshop, Nijmegen, 2005, p. 63. Example: implementation of a new business process might benefit from other modeling mechanisms such as the ARIS toolkit as described in Scheer, AugustWilhem: Business Process Reengineering – Reference Models for Industrial Enterprises, 2nd edition, Berlin, 1994. Flood, Robert L. and Michael C. Jackson: Creative Problem Solving – Total Systems Intervention, Chichester, 1991, p. 12. For a discussion on power, conflict and organizational differences, see also Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Massachusetts, 1970, p. 81.

D. The Usage and Utility of Participative Modeling in Change Management

123

issues of interests, conflict and power (see table D-1). From a change management viewpoint, modeling is primarily relevant in a pre-dominantly pluralist situation. In a dominant unitary situation, change can be managed without investing time on modeling. In a dominant coercive situation, it is illusionary to expect the gain the necessary consensus and trust in the modeling sessions.

Unitary (modeling might not be worth the effort)

Pluralist

Coercive (modeling ill suited)

Interest

Common objectives - a well integrated team

Diverging group interests with the organization as a mutual focal point – loose coalition

Oppositional and contradictory interests – rival forces

Conflict

Rare and transient

Inherent, but may well have positive aspects

Inevitable and likely to lead to radical change of whole structure

Power

Replaced by conceptions such as leadership and control

Medium through which conflict of interest may be resolved

Unequally distributed thus allowing domination, subjugation and so on

Table D-1: Political characteristics of situations in terms of the issues of interest, conflict, and power341

In the case study, the Board of Directors exercised some degree of coercive power in the outlining of the overall objective. It was not a question of whether a new location strategy should take place, but only of how to implement the decision. This was clearly against the initial personal interest of a number of

341

The table is based on the table in Flood, Robert L. and Michael C. Jackson: Creative Problem Solving – Total Systems Intervention, Chichester, 1991, p. 13. The indication of modeling relevance is a modification of the table.

124

D. The Usage and Utility of Participative Modeling in Change Management

the people responsible for implementing the new structure and processes. In this situation, the overall business strategy and objectives of the companies was used to seek interest alignment, and the modeling process was used as the main mechanism ensuring involvement and focusing the discussions on both the usefulness and the practical implications of the decided change.342 The level of conflict was relatively high, resulting in strong engagement in the discussions, although not of a magnitude preventing a constructive atmosphere after the initiation of a structured and framed process. Within the frame given by the outlined targets, the project participants had a high degree of empowerment. Even radical new ideas on how to conduct business were discussed, recognizing the need for creativity to come from all levels in the organization, and not as a top-down process.343 A modeling process will inevitably be part of the social constructivism in the setting.344 In the social constructivism typology by Wenneberg, ‘the critical perspective’ and ‘the ethical reflection’ provide particularly interesting perspectives on modeling.345 In the case study, at a point in time the core team discussed whether to include the parameter ‘motivation’: having capacity buildup in low-cost locations resulting in reduced motivation among high-cost employees and consequently reducing productivity in high-cost locations. From a critical perspective, this might be an actually existing underlying social construct, but the project owners did not want to legitimize such behavior, which can be 342

343

344

345

See Snabe, Birgitte and Andreas Größler: “System Dynamics Modelling for Strategy Implementation: Case Study and Issues”, Journal of Systems Research and Behavioral Science, Vol. 23, No. 4, in print, 2006, pp. 16—17, for a discussion on ‘manipulation’ when refining and implementing a given decision. See Hamel, Gary: Leading the Revolution, Boston, 2000, p. 280, opposing the idea that new strategies, innovation and change should always start from the top. Throughout the book it is discussed how innovation and radical ideas on how to change the way a company does business should also come from ‘activists’ from inside the company; with activists being people dedicated to rule-bursting and daring unconventional business. Kieser, Alfred: “Kontruktivistische Ansätze”, in Alfred Kieser (ed.): Organisationstheorien, 3rd edition, Stuttgart, 1999, p. 288, discusses social constructivism as how human communication and interaction produce a social reality that appears as the objective reality. Wenneberg, Søren B.: “Socialkonstruktivisme som videnskabsteori – Sisyfos’ videnskab”, Online Paper, Institut for Ledelse, Politik og Filosofi, Copenhagen, 2002, pp. 8—9.

D. The Usage and Utility of Participative Modeling in Change Management

125

seen as an ethical reflection, taking responsibility for influencing social constructs. A pragmatic solution from the project owners was to agree on the importance of motivation and to establish the guideline that scenarios should only be simulated with parameter settings within a range where motivation was not reduced; and in this way to avoid including ‘motivation’ explicit in the model.346 When discussing social constructivism in a modeling perspective, Van der Smagt criticizes the field of system dynamics and group model building for having a one-sided causal model construction view with focus on creating consensus about the representation (truth seeking), and typically ignoring a constitutive model construction view dealing with definition legitimacy (negotiation).347 For system dynamics modeling applied in change management purposes, the constitutive view is particularly important, as executives often have strong commitment to shaping change. The use of a preliminary model is a central instrument in framing and targeting the intervention, allowing project owners to ‘set the scene’ aiming at aligning expected insights and results from the modeling process with the purpose of the intervention. Furthermore, the use of a preliminary model not only frames the intervention, but also gives the project owners an opportunity to advocate a certain worldview. Example: in the case study company, there was a tendency among high-cost location employees to question the efficiency and the quality of colleagues in the low-cost locations. The project owners clearly upfront explained, that they disagreed in this causal relation, and during the modeling process it was discussed and agreed, that lower productivity should only be modeled as a consequence of less years of experience. This can be seen as a commitment to constitutional change, where the project owners seek to influence the mental models of modeling participants. A modeling project might be a well-suited mechanism when executives want to

346

347

Negative consequences on organizational trust among employees has been seen to be avoided even in cases of both redeployment and relocation of employees, which was interpreted as being due to factors such as acceptance of the arguments for change, perceived organizational support, and fairness; see Ferres, Natalie, Julia Connell, and Anthony Travaglione: “The effect of future redeployment on organizational trust”, Strategic Change, Vol. 14, March-April 2005, pp. 87—88. Van der Smagt, Ton: “Causation and Constitution in System Dynamics; Modelling a Socially Constituted World”, Journal of Systems Research and Behavioral Science, Vol. 23, No. 4, in print, 2006, pp. 13—14. Van de Smagt, p. 1, furthermore argues: “causal models blind us for constitutional change.”

126

D. The Usage and Utility of Participative Modeling in Change Management

impose changes in participants’ worldview (or Weltanchauung).348 Constitutional issues can be a conscious and integrated part of establishing preliminary models, both in the argued causal relations and with regards to underlying values or beliefs. Furthermore, the project owners’ involvement as a steering committee throughout the participative modeling process allows for constitutional negotiations in the model construction. 349 On the other hand, if an organization doesn’t also strongly encourage the truth-seeking perspective, it is doubtful that modeling efforts can create learning and consensus. In a change management setting, models are used to transfer and refine insight rather than develop new insight. The target group is not made up of top executives with extensive years of experience, but employees with less experience in the overall system’s causal relations and behavior. Employees from different business units often lack the overall picture, and in the view of Lyneis, companies far too often experience underperformance due to policies not being aligned between functional areas of the company and policies not being aligned with corporate goal-settings.350 System dynamics modeling can help to make some of these misalignments explicit, which is especially relevant for people with only partial responsibilities in the company.351 Often top executives themselves find the insights yielded by system dynamics models to be intuitively correct, but

348

349

350

351

See Checkland, Peter: “Systems Thinking, Systems Practice”, Chichester, 1993, p. 219, for a discussion on the importance of the concept of Weltanschauung in both ‘hard’ and ‘soft’ system methodologies. Van der Smagt, Ton: “Causation and Constitution in System Dynamics: Modelling a Socially Constituted World”, 2006, p. 14, argues that causal models do not even ‘invite’ to think about constitutional issues. This might very well be true for the modeling workshops, but in the preparation of the modeling processes (especially in drafting the preliminary model), model owners have a clear opportunity to include structured commitment to constitution. See Lyneis, James M.: Corporate Planning and Policy Design: A System Dynamics Approach, Boston, 1980, pp. 6—9. Lyneis furthermore emphasize corporate policies not being sufficiently robust to handle changing conditions. In Thygesen, Henriette H.: System Dynamics in Action, Copenhagen, 2004, p. 189, it is argued that the system dynamics modeling process assist in creating “an atmosphere of shared reality” between project participants.

D. The Usage and Utility of Participative Modeling in Change Management

127

with little newness, due to the executives own extensive, intuitive understanding of the overall system behavior.352 A final comment regarding context factors relevant to the appropriateness of using system dynamics modeling concerns skills, attitudes and traditions among project participants. It is relevant to consider whether the employees have skills and attitudes suitable for the usage of a highly conceptual and mathematical approach. Also, the process must correspond with how the organization normally handles change, solves problems, and involves employees. In the case study, most participants were mathematical skilled people; many were engineers. The question is whether modeling is only relevant for this type of company. An argument against this viewpoint is that in most companies, major change projects make use of extensive Excel-models to plan, track and optimize. Thus, formal models are part of strategic change initiatives anyway. Implementation of change projects as a part of business planning must be able to give input to the budgets for the coming quarters and years. Therefore, managers responsible for leading change are generally used to models, used to mathematical approaches, and often appreciate the advantages that system dynamics offers compared to black-box models. Furthermore, people who want to hide weak arguments in huge datasheets will have a problem. Through the system dynamics modeling process, the numbers from the business cases in each business unit came out in the open, and “the approach made it difficult for people to play politics”. The structure of the Excel-based business cases often made them into ‘black box’ models, where it was difficult for people other than the model owners to see the critical parameter setting through the ‘jungle’ of detailed information and hidden formulas. The system dynamics modeling process served as a ‘meta-model’, aiming at having a model on a high aggregation level only including the most significant parameters—creating transparency even beyond the people doing the actual modeling.

352

A former Forrester student, who has worked five years with system dynamics modeling in McKinsey, has stated this viewpoint. The CEO of the case study company has also stated that viewpoint.

128

D. The Usage and Utility of Participative Modeling in Change Management

II.

Process Considerations

The design of the modeling process is a sub-set and closely interrelated to the overall intervention design, with its stakeholder analysis, communication strategy and plan, adjustment of business processes and measurement systems etc. In the case study, although recognizing the interdependencies as well as the iterative nature both internally between the steps and with regards to the overall intervention design, eight conceptual modeling steps crystallize:

Step 1 Frame and Design Modeling Process Step 2 Present Preliminary Model as Kick-off Step 3 Adjust and Refine Model Step 4 Discuss and Stipulate Parameters Step 5 Analyze Sensitivity and Identify Critical Parameters Step 6 Simulate Scenarios Step 7 Plan for Improvements of Critical Parameters Step 8 Provide Feedback to Initial Decision-Makers Figure D-1: Conceptual modeling steps in the case study

D. The Usage and Utility of Participative Modeling in Change Management

129

The first step, Frame and Design Modeling Process, explicitly formulated the given decisions (‘what is not for discussion’), respectively the degrees of decision freedom in the implementation and modeling process (‘what is open for discussion’). The preliminary model had a central role in step 1, both in the investigation of the appropriateness of using a model and in the process of having project owners feel comfortable with expected key learning from the model. The business objectives and the targets in combination with the preliminary model clearly framed the modeling process, and because of transparency of process intentions, the project owners expected no perceptions of manipulation. In the second step, the preliminary model was used to initiate the modeling workshops with core project participants, who represented some of the key managers responsible for implementation. Also, a number of other stakeholders were introduced to the model in early versions, in order to allow for a sense of involvement and for getting input, and to strive for both effectiveness and efficiency in the process. The third step, Adjust and Refine Model, was intended to serve the purpose of participants gaining trust in the model, as well as improving the model. Allowing for model adjustments, together with later steps, was seen as important for giving modeling participants true influence, which is an important factor in getting commitment in the change process. In step four, Discuss and Stipulate Parameters, the parameters were discussed in the group. Some of the parameters were highly subjective, e.g. how long does it take for a rookie employee to be fully productive, and it was perceived as important for participant trust, that these estimations were made without interference from top executives. The quantification process served as a means for aligning parameters between business units, to be used in the different business units in the detailed business cases. Also, the parameter estimations initiated discussion on best practices across the business units. The fifth step, Analyze Sensitivity and Identify Critical Parameters, together with the sixth step, Simulate Scenarios, enabled the change leaders to identify the crucial parameters, where improvement initiatives would have significant impact.353 The main purpose was to focus the discussion on the critical parameters in a way leading to focused implementation decisions. The facilitator played an important role in turning barriers and issues into improvement 353

Example: the group realized the importance of reducing the elapsed time of handing over tasks. If this had been a top-down request, it could very well have caused significant change resistance.

130

D. The Usage and Utility of Participative Modeling in Change Management

opportunities requiring innovative solutions. It should be noted, however, that other modeling processes could be building on models, where the implementation decisions and improvement opportunities are not directly linked to parameters, but could be linked to possible variations in model structure or formulas. If this would be the case, both critical parameters and underlying decisions from structure and formulas could be relevant to prioritize. The seventh step, Plan for Improvements of Critical Parameters, was closely interlinked with the development of detailed implementation plans in each business unit. Focus was on creating a constructive and action-oriented discussion on high-impact improvement opportunity. Although the entire modeling process aimed at aligning mental models, this step seemed especially important in the overcoming of critical barriers to change and the development of a shared cross-organizational implementation plan. Step 8, Provide Feedback to Initial Decision-Makers, partly took the form as a series of check-points with the project owners, and partly took the form of a formal feedback formulated as a set of critical success factors indicating policy and governing adjustments important for sustainable and successful implementation. The feedback cycle has the twin aim of providing the top executives with important information on policy changes as well as of encouraging motivation among modeling participants through influence and involvement. Figure D-1, with the eight conceptual modeling steps, relates to the case study and is not stated as a generic modeling process for modeling in change management context. The figure merely gives some indication of potential relevant steps for change-management-oriented modeling, and interventions dealing with more complex problems, different magnitude of the number of key stakeholders, different selected change strategy, or other differences, would be expected to call for adjustments in the modeling process. The next four subchapters discuss in more general terms significant process considerations regarding the usage of system dynamics modeling in change programs. The discussions will mainly be founded on the theory discussions from the previous chapters and on additional normative change management literature. Furthermore, the discussions will include fairly detailed case study insights, when these can serve the aim of illustration. The four sub-chapters are selected trying to pinpoint some of the main differentiators between modeling in exploratory

D. The Usage and Utility of Participative Modeling in Change Management

131

settings vs. modeling in a change management context.354 The four sub-chapters are: (1) Business objectives and targets directing and framing the intervention, (2) Structured development of change leaders, (3) Designing of the change process, and (4) Facilitation of modeling and simulation sessions.

1.

Business Objectives and Targets Directing and Framing the Intervention

Compared to the traditionally exploratory system dynamics modeling processes, system dynamics modeling for change management limits the problem-solving process: the question of ‘what to do’ has already been answered, ‘how to do it’ is the problem that is addressed by the modeling process. In corporate settings, the outline for change management projects will often be business objectives and targets stated by the top executives. Seeing the modeling process as a journey of learning, the business objectives and targets give the direction and the frame for the journey and the modeling and simulation activities assist in understanding implementation challenges and the creation of commitment and alignment between the individual participants. For process improvements rather than reengineering, expectations of improvements also depend on the maturity of the process. For a process that has already undertaken many improvements programs, fine-tuning is expected rather than improvement leaps.355 The usage of a system dynamics model, however, will often be applied to deal with the deeper understanding of system behavior, and will consequently typically have broader impact than merely continuous process improvement.

354

355

The selection of four significant areas is made to focus the discussion and to avoid the impossible mission of writing a ‘manual’ on how to apply system dynamics modeling in change management context. In Schneiderman, Arthur M.: “Setting Quality Goals: Use observed rates of continuous improvement to position targets”, Quality Progress, April 1988, p. 56, rational goal setting is discussed in terms of the individual company’s situation. The term “half-life“ is introduced, indicating that a constant time factor exist for reaching half of the theoretical possible process improvement at any given time (comparable with the behavior of radioactive decay).

132

D. The Usage and Utility of Participative Modeling in Change Management

With a modeling process being a journey of learning, a central challenge in a change management context is to ensure that the learning is very likely to correspond with the intended change. In this respect, the use of a preliminary model can draft the process. Akkermans discusses the usage of a preliminary model as a means to create a certain focus in a modeling process.356 In the case study, the preliminary model was constructed to illustrate the expected behavior of the new location strategy. During the development of the preliminary model in the case study, the very first results already took form as the project owners gained some interesting insights. Something first considered as a potential mistake in the model turned out to be an important insight: it became clear that one decision, which had recently been made, had a stronger negative impact in year 1 than anticipated. A decision was therefore made to modify the course of action, and make the transition over a longer time-horizon. In addition, the preliminary model appropriately illustrated the high-level idea behind the new strategy, which was a prerequisite for the project owners to initiate the modeling activities. The project owners had no intentions to start a group model building process from clean sheets of paper with the risk of losing control, but wanted to see a preliminary model that in broad terms supported their relatively firm viewpoint on the future direction of the company. It is likely be a general trend in corporate environments that executives have a clear view of the direction they want to drive a given change. Executives could therefore be expected only to initiate a modeling process, if they are shown a model whose behavior corresponds with the insights that they want to transfer to the implementers, and if the model at the same time seems to be an effective and appropriate “tool” to demonstrate and investigate the problem. Furthermore, lack of experience and comfort with participative modeling processes could result in the fear of a project resulting in a model with hidden errors or having the problem being addressed or conceptualized erroneously. In the case study, top executive initial trust in the modeling process was gained through the preliminary model. From the viewpoint of modeling participants, the modeling process in a change management context is restricted to some degree, compared to the exploratory freedom in traditional modeling studies. Akkermans discusses the possible reduced participant ownership of the model, when using a preliminary model, compared to an ‘empty whiteboard’ approach.357 Regarding the location

356 357

Akkermans, Henk: Modelling With Managers, Breda, 1995, pp. 116—117. Akkermans: Modelling With Managers, 1995, p. 116.

D. The Usage and Utility of Participative Modeling in Change Management

133

strategy case study, one extreme viewpoint could be that no free exploration of all possible scenarios took place, no free discussions over sound and just business objectives were held, and only limited hierarchy-free conversations resulted. Sterman discusses situations, where clients use models to support already reached conclusions, recommending modelers to “speak truth to power” and get “a better client,” and Borum discusses the importance of free, informed choices in establishing commitment in change processes. 358 A different viewpoint on the same matter could be that everyone in the process was informed about the expected goals of the project, no-one was given the impression that he or she had more influence on the direction of the strategy than was actually the case, and noone did generally criticize the task of implementing a strategy that was decided by someone else. On the contrary, everyone agreed that by modeling, structure was given to a usually chaotic and emotionally charged process and with the system dynamics model a basis for discussion existed. Furthermore, free, informed choices took place both with regards to the model refinement, parameter settings, and the decisions on how to implement the strategy. This was a deliberate part of enhancing participant ownership of the model. It can be argued that the context in which the case study took place reflects a rather typical and frequently occurring set-up in corporate environments with which organizational members do not necessarily disagree. In the evaluation interviews, there were some remarks concerning the fact that participants in the process were quite informed about their options to change things. Most agreed that their task was to refine and implement a given strategy, not to re-formulate the strategy. However, for others the modeling process served as a disciplinary tool, as the following quote from a participant shows: “A few participants did not agree with the business objectives and for that reason they also did not agree with the process, but nevertheless the process forced them in the decided direction, and through the modeling they gained some of the insights motivating the intervention in the first place”. However, this can be true for any intervention approach and for many situations in organizations: people in business firms are well aware of the fact that they sometimes have to implement things they would

358

Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, 2000, p. 85; Borum, Finn: Strategier for organisationsændringer, Copenhagen, 1995, p. 58. It is important to note, that Sterman’s recommendation is made in a modeling context different from change management.

134

D. The Usage and Utility of Participative Modeling in Change Management

rather not. In interviews with the project owners, the issue of possible ‘manipulation’ was addressed explicitly. The project owners did not agree with this viewpoint, as it is normal business that the top executives outline a strategy and the level below is responsible for implementing it. In their view, top executives as a part of their role and responsibility exercise legitimate use of power, which has noting to do with manipulation because it is openly stated upfront. Thus, there are frequently decisions in larger organizations that are not open for discussion. It is difficult to say whether the problematic of a framed intervention had a negative impact on the participants’ ownership and trust in the model. The questionnaires do not explicitly include questions regarding this possible impact of a preliminary model, due to the problem of a measure that influences the system (in this case, creating negative attitudes by explicitly raising the issue). As a final remark on the role of business objectives and targets, it is relevant to consider the common trends in the literature discussing change management. In general, clear objectives and targets are seen as important and necessary elements in creating planned change. Therefore, when modeling efforts are used within the context of change management it could be seen as quite natural that the process is framed by business objectives and targets.359 Furthermore, the literature also emphasizes the iterative nature of problem solving and strategic processes, implying the need for implementation issues and new insights gained in an implementation process feeding back to the strategy forming or decision-making phases.360 In the case study, feedback of modeling insights to the original decision-makers took place as formally stated critical success factors, implying needed adjustments in the overall company governance model.361

359

360

361

A parallel can be drawn to the importance of clear objectives in learning and experimentation situations supported by simulators, see Größler, Andreas: “Don’t Let History Repeat Itself: Methodological Issues Concerning the Use of Simulators in Teaching and Experimentation”, System Dynamics Review, Volume 20, Number 3, Fall 2004, pp. 268—269. See discussion in chapter A on the diagnostics and decision-making cycle, and the change management cycle, in problem-solving processes. One example is the importance of head-count focus vs. cost focus in the governance procedures, which is discussed elsewhere in this dissertation.

D. The Usage and Utility of Participative Modeling in Change Management

2.

135

Structured Development of Change Leaders

The more difficult a change program is and the more internal commitment is necessary for effectiveness, the more the key employees need to be involved in the design, execution and monitoring of the changes in order to ensure sustainable change.362 In many consulting projects, a practical solution on how to establish such involvement is to strive for the development of what could be called ‘change leaders.’363 In this context, change leaders are the critical managers from across the organizational structure and hierarchy with the highest direct influence on implementation. The term ‘reference group’ is sometimes used in the literature for a similar type of employee involvement in projects, but often reference groups have a broader scope than the establishment of change leaders, also encompassing other groups of stakeholders.364 Anderson and Anderson discuss the creation of change leaders in what seems to be a somewhat different context, focused on the leadership skills among the top executives formally responsibility for the change process.365

362

363

364

365

Argyris, Chris: Interventions Theory and Method – A behavioural Science View, Reading, Massachusetts, 1970, p. 83. This observation is based on a number of consulting projects, which the author has carried out as consultant or engagement manager in IBM Management Consulting and Deloitte Consulting Group between 1993 and 2002. Andersen Consulting uses a similar approach (although it uses other terms), see Lochmann, Hans-Dieter und Michaela Rüsch-Kornasoff: “Organization Change Strategy – Ein wesentlicher Baustein des Reengineerings”, in Manfred Perlitz, Andreas Offinger, Michael Reinhardt and Klaus Schug (eds.): Reengineering zwischen Anspruch und Wirklichkeit, Wiesbaden, 1996, pp. 329—340. Involvement of reference groups is often discussed in change management literature with regards to the capture of important information and input, and in order to ease a later implementation by avoiding negative attitudes to changes only due to the fact that some people feel offended that they were not at all involved in the project. In the approach “The Reference Group” by Jørgen Randers, referees are involved primarily through interviews and workshops, see Rouwette, Etiënne: Group model building as mutual persuasion, Nijmegen, 2003, p. 44. In group model building approaches emphasis is rather on direct participation in modeling and simulations efforts. Anderson, Linda A. and Dean Anderson: “Awake at the Wheel: Moving beyond Change Management to Conscious Change Leadership”, OD Practitioner, Vol. 33, No. 3, 2001, p. 45.

136

D. The Usage and Utility of Participative Modeling in Change Management

The purpose of development of change leaders is to transfer project ownership from the top executives who have decided and launched a change project to the managers being responsible of implementing it. This could be done with the dual objective of: (1) using the change leaders’ input to establish effective action and communication plans, and maybe to refine or alter the change program,366 and (2) get the change leaders committed to the change, taking into account the full range of cognitive, affective, and conative elements, social norms elements, and perceived control elements.367

Discussing change leaders, it is interesting to draw a parallel with the usage core team members, as seen in many Business Reengineering projects, where individuals are taken out of their previous environment, to use their skills and experiences in the creation of reengineered business processes.368 In reengineering projects, the role of the core team members has a strong diagnostic focus of the problem solving process, whereas the use of change leaders has a strong change management focus building upon the individual change leaders’

366

367

368

The structured involvement of change leaders is a way to exercise what Klein calls ‘sympathetic understanding’ regarding the change resistance, seeking valuable input about the nature of the system that is going to change, potentially motivating a modification of the change itself or the change implementation process, see Klein, Donald: “Some Notes on the Dynamics of Resistance to Change: The Defender Role”, in Warren G. Bennis, Kenneth D. Benne and Robert Chin: The Planning of Change, 4 th edition (first published in 1966), New York, 1985, p. 103. The five elements are taken from Ajzen’s framework for theories of planned behavior; see discussion in chapter B.II.1. See also Schein, Edgar H.: Process Consultation, Boston, 2000, part I, p. 68, emphasizing the importance that the handover from the decision-makers to the implementers should be carefully planned to avoid communication breakdown. As seen in Business Reengineering projects, where individuals are taken out of their previous environment, to use their skills and experiences in the creation of reengineered business processes Bungard, Walter: “Zur Implementierungsproblematik bei Business-Reengineering Projekten”, in Manfred Perlitz, Andreas Offinger, Michael Reinhardt and Klaus Schug (eds.): Reengineering zwischen Anspruch und Wirklichkeit, Wiesbaden, 1996, pp. 264—265.

D. The Usage and Utility of Participative Modeling in Change Management

137

existing organizational platform.369 Compared to projects where project members are selected primarily based on knowledge, experience, and formal responsibilities, change leaders are selected using additional criteria. Change leaders are also selected based on peer network and informal power-bases, which means that not only managers formally responsible for the implementation, but also managers with strong informal influence are relevant for consideration as change leaders.370 The change leaders will (if it works out successfully) function as early adaptors of the change, and this fits well with the thoughts of Schein, who recommends that the personal power-base of individuals, their connections with peer employees, as well as their change readiness should influence with whom to start the change process.371 These individuals have the potential to positively influence the change process in informal ways, and alternatively (if not being involved) they might contribute to the stiffness in the organizational change resistance. For the later reason, it might be relevant to involve powerful employees even with low change readiness, primarily to avoid or reduce their counter-productiveness in the implementation phase. The learning and establishment of commitment strived for through a modeling process might be especially relevant for such individuals. Involving change leaders in change planning activities often encompasses key employees from different hierarchical levels. In this regard, the planning of activities and forums should take into concern the challenges regarding hierarchical-free discourse, which is an area that still troubles the field of social science despite the fact that significant research has already been made in this

369

370

371

Although it should be noted, that even though the change leaders are going to implement an already outlined strategic initiative, the implementation in itself also involves creative problem-solving. Projects involving system dynamics will be likely to regard discontinuous change, requiring the change leaders to navigate in a new landscape, where new and innovative solutions are required. It should be noted that having change leaders also being informal change leaders implies a broader definition of the term change leader compared to the use in Anderson, Linda A. and Dean Anderson: “Awake at the Wheel: Moving beyond Change Management to Conscious Change Leadership”, OD Practitioner, Vol. 33, No. 3, 2001, pp. 40—48, where focus is mostly on top executives. Schein, Edgar H.: Organisationspsykology, Danish translation, Herning, 1990, p. 257.

138

D. The Usage and Utility of Participative Modeling in Change Management

area.372 A number of arguments exist for ‘educating’ a group or a number of groups of change leaders, as opposed to educate the change leaders individually. Decisions made by individuals in group settings, even regarding the individual’s own goals, seems to have a significantly more endurable behavioral effect compared to those made in a 1-on-1 lecturing setting.373 Also, there are strong indications that it is easier to change the ideology and social practice of a small group handled together than of single individuals.374 Furthermore, most individuals do not want to divagate too far from the standards of the group they belong to or wish to belong to.375 Therefore, the feeling of belonging to a ‘prestigious’ group of change leaders might be important. To avoid the change leaders “falling back” into old norms or behavior, the new group ties to the other change leaders could be expected to be of importance. Also, the selection process of change leaders might benefit from considerations regarding the individual’s personality characteristics, such as whether he or she easily feels uncomfortable representing new norms or behavior. In the efforts to strive for creating sustainable change, the change leaders could also be expected to influence the social norms in broader organizational contexts, and the discussions within the group of change leaders could be expected to positively influence the individual change leaders perceived behavioral control.376 Gladwell presents an interesting analogy for the structured usage of change leaders in change management, proposing that the spread of messages and behavior in social systems can be compared to epidemics spreading a virus.377 An

372

373

374

375

376

377

Jöns, Ingela: Managementstrategien und Organisationswandel, Mannheim University, 1995, p. 157. Lewin, Kurt: “Group Decision and Social Change (first published in Newcomb and Hartley’s Readings in social psychology, 1948, pp. 330—341), in Martin Gold: The Complete Social Scientist – A Kurt Lewin Reader, Washington, American Psychological Association, 1999, pp. 276—279. Lewin, Kurt: “Group Decision and Social Change”, in Martin Gold: The Complete Social Scientist – A Kurt Lewin Reader, Washington, 1999, p. 273. Lewin, Kurt: “Group Decision and Social Change”, in Gold, Martin: The Complete Social Scientist – A Kurt Lewin Reader, 1999, p. 281. For a discussion on the importance of social norms and perceived behavioral control in the changing of behavior, see Ajzen, Icek: Attitudes, Personality and Behavior, Chicago, 1988, pp. 121—133. Gladwell, Malcolm: The Tipping Point, paperback edition, New York, 2002, p. 9.

D. The Usage and Utility of Participative Modeling in Change Management

139

epidemic will spread widely as a result of the behavior of the people transmitting the virus, the characteristics of the virus, and environmental factors.378 Using this analogy in a change management context gives rise to conscious reflections regarding: who to involve, how to make the message and behavior ‘stick’ and how to create settings and conditions that support the change process. In terms of the use and selection of change leaders, Gladwell in particular has some interesting thoughts regarding the first aspect: who to involve. He has formulated the ‘Law of the Few,’ arguing that the effective spread of messages and behavior often depends on relatively few individuals in a social system, namely individuals with special skills regarding relations with others, regarding information accumulation or regarding persuasion.379 The people with special skills regarding relations with other people are called ‘Connectors.’ Compared to other individuals, they know and interact with significantly more people. They are the social glue of an organization, and they spread messages. The people who accumulate information, the ‘Mavens’ are the one’s whose opinion most people will take very seriously. The last group, the ‘Salesmen,’ consist of those who are able to persuade unconvinced colleagues. Gladwell argues, that all three types of personalities are critical to involve if a change process is to take advantages of the ‘word-of-mouth’ phenomena. Gladwell’s thoughts on creating change as initiating epidemics has some similarities with Hamel’s 8-step process on how to “revolt” the business by creating a movement within the organization, although Hamel’s process is a bottom-up approach to be used by innovative employees rather than a top-down change approach. The 8-step process includes “infecting” others with the idea, creating coalitions, and working closely together with representatives from across the organization.380 The structured development of change leaders should also be viewed in terms of organizational learning. An organizational intervention cannot be seen independently of the ongoing organizational development. The single intervention is highly dependent on the flexibility and change readiness of the individuals involved and the existing system structures. The intervention, on the 378

379 380

When an epidemic starts to spread widely, it has passed, what Gladwell calls ‘the tipping point, see Gladwell, Malcolm: The Tipping Point, paperback edition, New York, 2002, p. 18. Gladwell, Malcolm: The Tipping Point, New York, 2002, p. 19 and pp. 30—88. Hamel, Gary: Leading the Revolution, Boston, 2000, pp. 187—206.

140

D. The Usage and Utility of Participative Modeling in Change Management

other hand, is also a part of the continuous learning and shaping of the organization, its change readiness, and its change capabilities, why it seems relevant to have the intervention planning including organizational development considerations as well as considerations of each change leader’s individual personal transition.381 The usage of modeling can help in the establishment of understanding and commitment among change leaders regarding a change process. But other elements are important for the effectiveness of change leaders, mainly regarding general leadership skills such as interpersonal relations and communication skills, e.g. presenting ideas in a way that activates feelings and makes the message memorable.382

3.

Designing the Change Process

Looking for models and frameworks for the design of change processes, most approaches still build upon Lewin’s change model and the action research theories, suggesting a general framework for planned change in organizations with four basic activities:383 1. Entering and contracting (initial data gathering and committing resources) 2. Diagnosing 3. Planning and implementing change 4. Evaluating and institutionalizing change

381

382

383

See Schein, Edgar H.: Organisations Psychology, Herning, Forlaget systime, 1990, p. 40; Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Massachusetts, 1970, chapter 1 and 2. McKee, Robert: “Storytelling That Moves People”, Harvard Business Review, June 2003, p. 52. Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, p. 28. This framework fits also well with generic consulting methodologies; e.g. a phase model of reengineering projects described by Perlitz: (1) initializing, incl. project initialization and project understanding, (2) problem analysis and redesign, (3) implementation, and (4) anchoring and continuous development, see Perlitz, Manfred, Jürgen Bufka, Andreas Offinger, Michael Reinhardt, und Klaus Schug: “ReengineeringProjekte erfolgreich umsetzen – Ergebnisse einer Erfolgsfaktorenstudie”, in Perlitz, Manfred, Andreas Offinger, Michael Reinhardt and Klaus Schug (eds.): Reengineering zwischen Anspruch und Wirklichkeit, Wiesbaden, 1996, p. 186.

D. The Usage and Utility of Participative Modeling in Change Management

141

The dedicated change management literature focuses especially on the third step in Lewin’s model: planning and implementing change. The theory of planned change, with the “unfreezing-movement-freezing” process, takes as point of departure the need for weakening existing behavior or attitudes, establishment of feelings of dislike for the present situation, and also establishment of psychological feeling of safeness in the change process.384 The establishment of dislike for the present situation is often discussed using terms such as ‘change imperative’, ‘burning platform’, and ‘establishing sense of urgency’, and De Geus and Dörner discuss the usage of both positive and negative goals for motivating change.385 The change imperative functions as a mechanism to open minds to change and to lower change resistance. Change resistance, or rather overcoming change resistance, is a central element in the discussions of planned change and implementation of organizational changes.386. Doppler and Lauterburg give three causes of change resistance: (1) the goals, background or motives for change are not understood, (2) the people concerned understand what is said but do not believe it, and (3) the message is understood and believed, but the people concerned will not or cannot comply with the change.387 Furthermore, they summarize symptoms for change resistance in two dimensions: verbal vs. nonverbal and active vs. passive, see table D-2. In the case study regarding the implementation of a location strategy, a large number of the symptoms in table D-2 could be observed directly or indirectly. The descriptions of goals, background and motives for the change (described in chapter C.II.1 and C.II.2) are intuitively easy to dismiss as “wishful

384

385

386

387

For further discussion on the unfreezing-movement-freezing process, see chapter A.II as well as Schein, Edgar H.: Organisationspsykologi, Herning, 1990, pp. 254—255. See Kotter, John P: Leading Change, Boston, 1996, pp. 35—49; de Geus, Arie P.: “Planning as Learning”, Harvard Business Review, Vol. 66, No. 2, MarchApril 1988, pp. 70—74; and Dörner, Dietrich: The Logic of Failure, New York, 1996, pp. 49—54. Chin, Robert and Kenneth D. Benne: “General Strategies for Effecting Changes in Human Systems”, in Bennis, Warren G., Kenneth D. Benne and Robert Chin: The Planning of Change, 4 th edition, New York, 1985, p. 22. Doppler, Klaus and Christoph Lauterburg: Change Management – Den Unternehmenswandel gestalten, 10 th edition, New York, 2002, p. 324.

142

D. The Usage and Utility of Participative Modeling in Change Management

thinking from the Board.”388 For the people expected to lead the change as well as for the employees as such, the project owners found that all the three main causes for change resistance listed by Doppler and Lauterburg were relevant when designing the change process: addressing whether the reasons for change were understood, addressing whether the reasons were believed, and addressing whether people were willing and able to conform with the change.

Verbal (Talk)

Nonverbal (Behavior)

Active OPPOSITION (aggressive) Augmenting against Accusations Threats Polemic Stubborn formalism

REBELLION Disturbances Controversy Intrigue Rumors Building of cliques

Passive (escape)

DEMOTIVATION Lack of attention Lack of energy Absence Withdrawal Illness

EVASIVENESS Silence Treat it as unimportant Make it look stupid Make fun of it Discuss less important issues

Table D-2: Generic symptoms of change resistance389

In the early years of the 21st century, normative management literature has increased attention devoted to the affective side of planning and implementing

388

- The change imperative was stated as: “right now is the right time to hire people in the low-cost locations, because due to company growth, it can be done right now without staff reduction at high-cost locations, and the expected results are improved competitiveness and further company growth, also securing jobs at high-cost locations in the future.” 389 Own translation and modification from Doppler, Klaus and Christoph Lauterburg: Change Management – Den Unternehmenswandel gestalten, 10th edition, New York, 2002, p. 326.

D. The Usage and Utility of Participative Modeling in Change Management

143

change. Kotter and Cohen have introduced the term “see-feel-change” arguing that, “people less change what they do because they are given analysis that shifts their thinking than because they are shown a truth that influences their feelings.”390 Roberto and Levesque use similar expressions, arguing for the “show me” rather than the “tell me” approach.391 To make people “see” includes communicating the visions and the need for change using emotionally engaging approaches, such as storytelling, symbolic actions, a video of an angry customer or an off-site event.392 The usage of participative system dynamics modeling fits well with the affective oriented change approaches. The theories of system dynamics are based upon not only a cognitive, but also an affective and ‘seeing is believing’ learning approach, supporting the learning process by stimulating experimentation and simulating experience. The importance of the latter is expressed by Brown as follows: “It’s never enough to just tell people about some new insight. Rather, you have to get them experience it in a way that evokes its power and possibility”.393 A modeling process fosters involvement and participation, which is widely recognized as some of the most effective strategies for overcoming change resistance.394 Although dealing with change in large organizations, there is the problem of scalability. For large-scale organizational change it seems rather unrealistic to have a significant number of the implicated employees participate 390

391

392

393

394

See Kotter, John P. and Dan S. Cohen: The Heart of Change, Boston, 2002, p. 1. As the book is practical rather than research oriented, there is no discussion of the “see-feel-change” in terms of cognitive, affective and conative elements. Roberto, Michael and Lynne Levesque: “The Art of Making Change Stick”, MIT Sloan Management Review, Summer 2005, Vol. 46, No. 4, Summer 2005, p. 56. See Kotter and Cohen: The Heart of Change, 2002, p. 181; Roberto, Michael and Lynne Levesque: “The Art of Making Change Stick”, MIT Sloan Management Review, Summer 2005, Vol. 46, No. 4, Summer 2005, pp. 56—57. Feelings facilitating change include faith, trust, optimism, urgency, reality based pride, passion, excitement, hope, and enthusiasm, whereas feelings like anger, false pride, pessimism, arrogance, cynicism, panic, exhaustion, insecurity, and anxiety undermine change, see Kotter and Cohen: The Heart of Change, 2002, p. 180. Brown, John Seely: Research That Reinvents the Corporation, Harvard Business Review, August 2002, p. 108. Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, p. 158.

D. The Usage and Utility of Participative Modeling in Change Management

144

The process by which the organization defines the initiative’s purpose, its scope and the way people will work with one another on the program. • Boundary Setting – Definition of scope of initiative

LEARNING

How managers develop, test and refine ideas through experimentation before full-scale rollout..

MOBILIZING

• Team Design – Definition of roles, responsibilities, norms and ground rules for teamwork

The use of symbolism, metaphors and compelling stories to engage hearts as well as minds in order to build commitment to the project.

REALIGNING

CHARTERING

in modeling efforts.395 In the case study, participative modeling did not support the general learning process of the organization, but targeted the key managers responsible for the implementation, by aiming to developing these individuals into change leaders. The modeling efforts addressed the initial parts of the change management process – or what could be called bridging the launch of the strategic initiative and the implementation efforts. Roberto and Levesque have proposed a change framework arguing that certain parts of the change processes must be planned well before actually starting the main intervention; hence the name Four Antecedent Processes (figure D-2).

A series of activities aimed at reshaping the organizational context, including a redefinition of roles and reporting relationships as well as new approaches to monitoring, measurement and compensation.

• Discovery – Data and information gathering to define goals of initiative and means of achieving objectives • Experimentation – Testing and refinement of initiative prior to full-scale rollout.

• Storytelling – Use of stories and metaphors to create compelling accounts about need for initiative and explain specific changes • Symbolic actions – Use of symbols to reinforce creditability and legitimacy of core team and its message

• Job Redesign – Alteration of underlying structures and processes that support jobs. • Performance Management – Invention of new metrics to measure effectiveness of initiative and incorporation of the metrics into employee appraisal process

Figure D-2: Four antecedent processes in organizational interventions 396

395

396

Although some of the same underlying mechanisms can be applied in larger scale in the form of gaming environments, simulators, etc. The figure is taken from Michael and Lynne Levesque: “The Art of Making Change Stick”, MIT Sloan Management Review, Summer 2005, Vol. 46, No. 4,

D. The Usage and Utility of Participative Modeling in Change Management

145

Participative modeling is especially relevant in the second of the four antecedent processes, the learning process, as both individual and group learning is the essence of modeling. Planning of the modeling process is also closely interlinked with the chartering process, with focus on traditional change management activities such as total intervention design and planning, traditional stakeholder analysis, strategies for communication and involvement of key stakeholders.397 Regarding the mobilizing process, a modeling process has the potential of providing a deeper understanding of the problem as well as simulating experience, which will contribute to change leaders’ ability to perform effective leadership. However, the individual leaders’ ability to perform effective communication and their interpersonal skills must be expected to be the main factors for the effectiveness of the mobilization phase.398 When it comes to the realigning process, the modeling process has offered the possibility of create a deeper understanding of the problem, and cross-organizational discussion in the search for best practice solutions and an aligned implementation plan. Furthermore, actual model insights will be likely to influence the realigning process through input to both job redesign and performance measurements. All in all, modeling seems to have the potential to contribute to the establishment of the foundation for change roll-out, and due to its contribution to the development of a cross-organizational shared mental model, it also seems to have the potential to support the refreezing or, in more modern terms, the establishment of sustainable change. With the modeling process being an integral part of a planned change process, it is necessary to ensure that the process to ‘stays on track’, with an ongoing alignment between modeling learning and the objectives of the change project. The Workbook methodology as described in the Group Model Building literature offers a way to ensure communication along the process internally among modeling participants as well as with stakeholders who do not take part in the modeling sessions. When modeling for change management purposes, it also

397

398

Summer 2005, p. 56 and p. 60. The four processes are stated to be critical to successful change; although the need for clear objectives, sound project management, accountability and control systems are also stressed (p. 55). An example of this kind of planning is described in the case study description of the intervention design, see description chapter C.II.2. The importance of leaders presenting ideas in a way that activate feelings and make the message memorable is also discussed in McKee, Robert: “Storytelling That Moves People”, Harvard Business Review, June 2003, p. 52.

146

D. The Usage and Utility of Participative Modeling in Change Management

seems fair to assume that at least one of the facilitators in the modeling projects must be able to represent the viewpoints from the top executives. Furthermore, traditional project steering mechanisms such as frequent check-points with a steering committee would seem to be a likely approach in securing alignment between modeling learning and change objectives. Although such aligning is a balancing act: on the one hand to get unbiased input from and discussions within the group, on the other hand to control the scope and orientation of the modeling process. For change projects dealing with more complex or cross-organizational problems involving highly specialized managers, the cognitive capabilities of the individuals must be expected to yield special challenges. Kieser and Koch discuss the problem of shared knowledge when the individuals have difficulties overlooking the possible combinations of knowledge, and they present case study findings suggesting the relevance of thinking in terms of re-combining knowledge rather than sharing knowledge.399 For organizational learning in organizations with high specialization, Kieser and Koch also call for a knowledge integration mechanism that does not rely on cross-learning, and discuss the benefits of simulated prototyping, simulated experience, joint thought experiments et cetera, and in their conclusions they emphasize the importance of teams and communities in creative knowledge creation.400 It could be an area of further research to investigate the possible role of system dynamics modeling in such a context; maybe in a combination of a top executive “meta-modeling” process in iterations with change leaders participating in modeling processes involving one or more sub-models. For complex problems, the involvement of change leaders is important in the search for sustainable change, but it must also be expected to be crucial to involve the employees representing the critical specialized knowledge.401

399

400

401

Kieser, Alfred and Ulrich Koch: Organizational Learning through Rule Adaptation: From the Behavioral Theory to Transactive Organizational Learning, Mannheim, 2000, p. 12 and pp. 15—16. Kieser and Koch: Organizational Learning through Rule Adaptation, 2000, pp. 18—27. See Forrester, Jay W.: “Policies, decisions and information sources for modeling”, European Journal of Operational Research, Vol. 59, No. 1, 1992, pp. 42—63. See also Kieser and Koch: Organizational Learning through Rule Adaptation, Mannheim, 2000, p. 19, discussing the transactive memory concept

D. The Usage and Utility of Participative Modeling in Change Management

147

As reducing change resistance is a core challenge in the actual roll-out of change initiatives, a counterproductive element is the ‘worse-before-better’ effect. Implementing change typically means changing routines, which requires the employees to perform their job in new ways, and consequently can be used as an argument against the usefulness of the change in the early phases of the changes.402 A number of system dynamics models have been published exploring and describing the effect of an initial lowering in performance due to improvement programs, including a model by Repenning and Sterman simulating a ‘working smarter strategy’ in creating and sustaining process improvement. Jacobsen and Samuel have modeled planned organizational change, considering the change process in terms of gap between current performance and target (assuming a certain gap will make a planned change process necessary), the simplicity of the change program, the cost of the change program, the pacing of the change program, the organizational change resistance, and lastly also two factors making employees accept a change program, namely employee involvement in the change and inducements (e.g. bonuses and other compensatory mechanisms).403 A major insight indicated by the use of this system dynamics model is that any planned organizational change will temporarily impair the organization’s performance. In the case study, the model clearly showed this ‘worse-before-better’ effect and thereby had a major role in expectation settings for the roll-out phase.

402

403

in terms of meta-knowledge or directory knowledge: knowledge about where to find knowledge. Repenning, Nelson P. and John D. Sterman.: Nobody Ever Gets Credit for Fixing Problems that Never Happened: Creating and Sustaining Process Improvement”, California Management Review, Vol.43, No.4, Summer 2001, p. 74. Jacobsen, Chanoch and Yitzhak Samuel: “Planned Organizational Change: Theory, Model, Data and Simulation”, in Milling, Peter M. and Erich O.K. Zahn (eds.): Computer-Based Management of Complex Systems, Proceedings of the 1989 International Conference of the System Dynamics Society, Heidelberg, 1989, pp. 104—118.

148

4.

D. The Usage and Utility of Participative Modeling in Change Management

Facilitation of modeling and simulation sessions

Kotter’s 8-stage process of change and Roberto and Levesque’s Four Antecedent Processes both address the questions that, according to Doppler and Lauterburg, all employees ask themselves about a planned change initiative:404 1. Why and for what reasons do we need to change? (What are the goals? Are all the motivations made explicit or are there some hidden ones? Is the change really important) 2. Am I able? (Is it too difficult? Is it possible for me to do? What are my chances for success?) 3. Am I willing? (What’s in it for me? Is it interesting? Will I loose benefits or other factors that I value like colleagues, career opportunities etc.?) In the modeling discussions and the change process around them, these three questions are important when explicitly or implicitly addressing the change resistance for each participant. The more open and honest the workshop atmosphere is, the more directly the questions can be addressed. In general, the atmosphere of the modeling sessions is important. Gladwell describes a number of scientific research projects conveying that a positive atmosphere (e.g. a smile from a charismatic person or physical head-nodding among the receivers) influences the receiver’s likelihood to agreeing with a message.405 The organizational situation and interpersonal factors influence the magnitude of the challenge for the facilitator to establish a constructive atmosphere. If the change is due to present and serious problems, for example, managers are likely to exhibit behavior including hiding information, secrecy and denial, blame, avoidance and/or passivity and helplessness.406 In such an atmosphere, the

404

405

406

Own translation from Doppler, Klaus and Christoph Lauterburg: Change Management – Den Unternehmenswandelgestalten, 10th edition, New York, 2002, pp. 326—327. Gladwell, Malcolm: The Tipping Point, paperback edition, New York, Back Bay Books, 2002, pp. 74—87. Kanter, Rosabeth Moss: “Leadership and the Psychology of Turnarounds”, Harvard Business Review, June 2003, p. 61.

D. The Usage and Utility of Participative Modeling in Change Management

149

facilitator’s job of establishing trust in the process is challenging, but also crucial for team members to be willing to contribute and cooperate.407 Although a good atmosphere improves the learning potential, an ‘extremely’ good atmosphere also constitutes facilitation challenges. In groups characterized by group cohesiveness and consensus, groupthink can occur with lack of critical thinking, resulting in limiting discussions to few scenarios without due consideration of alternative scenarios or possible alternative gains.408 Furthermore, groupthink often leads to not attaining sufficient information (even from experts within the organization), ignoring facts not supporting the favorite scenarios, dismissing feedback or new information that should lead to changing earlier group decisions, as well as underestimating implementation challenges.409 In a change management context, the preliminary model and intervention objectives and targets frame the modeling project, but still, within the framework, such a project also involves innovative problem-solving and decision-making regarding the implementation strategy and plan. O’Connor and McDermott stress the importance of widening perspectives by stimulating creativity, having a variety of different viewpoints represented, and getting different sorts of feedback.410 The facilitator furthermore has the option to use different curiosity stimulating mechanisms, as curiosity makes people question mental models and generate new perspectives411 With regards to creativity and curiosity, creative problem-solving literature often emphasizes the importance of “asking stupid 407

408

409

410

411

Henttonen, Kaisa and Kirsimarja Blomqvist: “Managing distance in a global virtual team: the evolution of trust through technology-mediated relational communication”, Strategic Change, Vol. 14, March-April 2005, p. 108. Here, trust is discussed in respect to ‘normal’ teams in ‘normal’ projects, but Henttonen and Kirsimarja later also discus evolution of trust in virtual teams (which was actually one of the issues discussed among project participant the case study outlining how to implement the location strategy). Poor decision-making as result of groupthink is described in Janis, Irving L.: “Groupthink: The Problems of Conformity” (original printed in Psychology Today, Nov. 1971, pp. 271—279), in Morgan, Gareth: Creative Organization Theory, Newbury Park, California, 1989, pp. 225—227. Janis: “Groupthink: The Problems of Conformity”, in Morgan, Gareth: Creative Organization Theory, Newbury Park, California, 1989, p. 227. O’Connor, Joseph and Ian McDermott: The Art of System Thinking – Essential skills for creativity and problem solving, London, 1997, p. 141. O’Connor and McDermott: The Art of System Thinking – Essential skills for creativity and problem solving, London, 1997, p. 141.

150

D. The Usage and Utility of Participative Modeling in Change Management

questions” in an open an honest atmosphere. Leonard discusses the value of making constructive use of tensions between organizational units in the search for innovations, and calls for mechanisms to “translate across different languages and finding ways of encouraging the depersonalization of conflicting perspectives.”412 This is also related to the importance of aligning language and codes in change projects.413 Conflicting perspectives can follow from conflicting policies and objectives, structural misalignments, and from individual conflicting cognitive and affective perceptions. Warren describes the example of sales-driven policies vs. earnings-driven policies, where a sales-driven policy (with corresponding investment in marketing) can be better in the long term, whereas the earningsdriven policy (focusing on cost cutting) would seem better in the short term.414 Structural misalignments include conflicting measurement systems between both departments and individuals, resulting in focus on own performance measures rather that ‘the big picture.’ Differing cognitive and affective perceptions influence the alignment of goals and objectives on the individual level. In the case study an interesting workshop observation is, that even the core-group person showing the strongest personal disagreement with the location strategy objectives became involved in committed and vital discussions after a very short time, showing a positive attitude in the sessions. One of the explanations could be that he could not ‘resist the fun of modeling’ when he took part in sessions, as he and the other core members were very mathematically skilled and interested individuals. Another explanation could be that modeling efforts offer a cognitive framework for reducing personal barriers for involvement and honesty in discussions.415 For this reason, modeling might have particular benefits in diverse 412

413 414 415

Leonard, Dorothy: Wellsprings of Knowledge – Building and Sustaining the Sources of Innovation, Boston, 1998, pp. 74—75. See discussion in chapter B.III.2.d. Warren, Kim: Competitive Strategy Dynamics, Chichester, 2002, p. 264. This has some similarities with how astrologers use horoscopes as a virtual reality, where humans re-arrange their perception of their own life, see Munk, Kirstine: “In the Airy Spaces of Our Minds…: Cosmology and ritual design in modern, Western astrology”, in York, Michael (ed.): Nature, Religion, and Culture, London, in print, 2006, p. 27. Munk refers to the Danish astrologer and psychotherapist Pia Balk-Møller, who suggests that the use of a chart (a horoscope) makes a discussion ‘safer’, as the secrets are already on the table, and Munk furthermore stresses the importance of imaginative involvement in discussions of charts, where insight “not only has to be understood intellectually, but also has to be imagined and felt” (pp. 29—30).

D. The Usage and Utility of Participative Modeling in Change Management

151

teams. Distefano and Maznevski present research, showing that diverse teams often perform either significant worse (more often) or significant better (less often) compared to homogeneous teams.416 Managed well, diverse teams have potential to outperform homogeneous teams, and modeling efforts might be a relevant approach in such a setting due to possibilities of lowering personal barriers as well as due to providing structure to otherwise diverse discussions. The creation of an open and honest modeling atmosphere, which at the same time has some degree of tension and investigative interest, has much to do with creating a platform for learning and directing discussions, not with regards to efficient model completion, but with regards to allowing exchange of viewpoints and aligning mental models. In scenario simulations, focus should be on understanding why the model produces the given behavior, rather than using the model as “an evaluator” of different scenarios. The insights from discussion of the model behavior can lead to new insights. For real life evaluations, Farson and Keyes discuss the importance of moving beyond both success and failure, analyzing the underlying reasons with less focus on evaluation and more focus on interpretation.417 Being rewarded or complimented for success can actually be as de-motivating as criticism, whereas most humans are motivated by getting a deeper insight into their problems. In model understanding, not taking model results at face value, but using them to generate more discussion on model behavior encourages the long process of testing and gaining trust in the model, and avoiding the risk of false trust in the model. It is important that there are enough group activities to enable the necessary discussions and that each individual has been sufficient involved in the model building to gain both a sense of “ownership” and to gain trust in the model. Participant’s trust in the model is a prerequisite for establishing learning, commitment, and alignment of mental models. This is important when striving for sustainable change built upon a new shared mental model.418 Effective facilitation depends not only on fostering discussions, but also on making sure that the process relies on available facts. In the case study, the

416

417

418

Distefano, Joseph J. and Martha L. Maznevski: “Creating Value with Diverse Teams in Global Management”, Organizational Dynamics, Vol. 29, No. 1, p. 46. Quoted in Farson, Richard and Ralph Keyes: “The Failure-Tolerant Leader”, Harvard Business Review, August 2002, pp. 66—67. For a discussion of the importance of shared mental models in organizational learning, see chapter B.II.3.

152

D. The Usage and Utility of Participative Modeling in Change Management

initial setting of each of the parameters involved issues that could have been discussed for hours in both workshops and other related meetings. Opinions on parameters were often very different within the core project team. As an example, it was a widely accepted ‘fact’ among many of the project participants that employees in low-cost countries very often stayed only 1–2 years, because as soon as they attained experience in R&D, they could get a better-paid job in a high-cost country. But with the ‘forced quantification process’ with parameter stipulations, facts came on the table, as the data were actually available, documenting a very low employee turnover in the low-cost countries.419 The parameter stipulations furthermore served to set benchmarks between business units, and this approach enabled cross-business-unit knowledge and experience exchange. For example, one business unit already had high-scale experience with building up resources in low-cost countries, reflecting very low hiring costs due to procedures that the other business units decided to adopt. The modeling approach this way served as a forum for transfer of best practices, which was one of the objectives of the modeling project in the first place. In the case study, the quantification of the model and the simulation of scenarios helped to provide an understanding of which parameters most strongly influenced the effectiveness of the strategy. Based on this understanding, the facilitator directed the discussion on how the strategy could and should be executed, as the parameter setting reflected implementation decisions. For example, the discussion was focused on reducing the time spent on hand-over of tasks, resulting in discussions on structuring knowledge transfer both during and after the initial transfer period. The discussions also focused on how to reduce both time and costs of classroom training. Without modeling, the intuitive choice could have been to focus primarily on the costs, including the travel costs, but due to the model-guided new understanding of the influence of the different parameters, increased focus was placed on reducing the time spent on classroom training. For system dynamics practitioners, the model seems very simple, but it is interesting to notice that the project team first tried to handle the problem with the use of a normal Excel-spreadsheet, which became a complicated “black box”,

419

Christensen, Søren and Jan Molin: Organisationskulturer, Copenhagen, 1987, p. 27, discuss myths as inward, empty explanations used to legitimize certain behavior. See also Ackoff’s morale: “There is nothing so deceptive as an apparent truth“: Ackoff, Russel L.: The Art of Problem Solving – Accompanied by Ackoff’s Fables, New York, 1978, p. 84.

D. The Usage and Utility of Participative Modeling in Change Management

153

where it was difficult to see and understand significance of the different parameters in their influence of model behavior. Not just identifying possible parameter values, but using the modeling workshop to discuss how to ‘optimize’ the most influential parameters was seen as a cornerstone in the development of the implementation plan of the new location strategy. Briggs, de Vreede, and Nunamaker Jr. discuss in general terms the need for facilitation as a barrier to the diffusion of the usage of group support systems; arguing that despite the documented advantages of such systems, economic and political factors result in facilitators not remaining long term at facilities supporting non-routine, ad hoc projects.420 To attempt to overcome this barrier, they propose the concept of ‘thinkLet’, packaging facilitation skills in a preconfigured process aimed at supporting company-specific decision processes.421 The packaging of facilitation skills is supposed to allow practitioners without broad facilitation competencies to guide a “repeatable, predictable pattern of collaboration among people working towards a goal.”422 Whether this would also be suitable for system dynamics modeling in change management settings is an area for further research. In the case study, the questionnaire results scored relatively low on expectations to use modeling in other projects, despite the higher score on the usefulness of using modeling.423 One could expect this to be connected with the problem of facilitator accessibility. As a special comment on the facilitation challenges, it is worth noting the rare, but interesting, problematic of ‘flirting-with-disaster.’ Salge and Milling discuss the “flirting-with disaster” behavior in the Chernobyl on-line operations, where bypassing of rules resulted only in positive experiences, and therefore

420

421

422

423

Briggs, Robert, Gert-Jan de Vreede, and Jay F. Nunamaker Jr.: “Collaboration Engineering with ThinkLets to Pursue Sustained Success with Group Systems”, Journal of Management Information Systems, Vol. 19, No. 4, Spring 2003, p. 32. Briggs, de Vreede, and Nunamaker Jr.: “Collaboration Engineering with ThinkLets to Pursue Sustained Success with Group Systems”, Journal of Management Information Systems, Vol. 19, No. 4, Spring 2003, p. 45. Briggs, de Vreede, and Nunamaker Jr.: “Collaboration Engineering with ThinkLets to Pursue Sustained Success with Group Systems”, Journal of Management Information Systems, Vol. 19, No. 4, Spring 2003, p. 46. See table C3 in chapter C.III.2.

154

D. The Usage and Utility of Participative Modeling in Change Management

slowly became de facto routine.424 Although the effect is discussed in terms of learning in real systems, one could anticipate the same type of unwanted learning taking part in simulation environments. In simulation environments, this ‘flirtingwith-disaster’ might even be seen relatively more frequently, as the consequences would be only of a virtual nature. If the probability of the disaster occurring is small, requiring very specific parameter combinations, then even extensive simulations could take place without resulting in the experience of negative consequences. Although not qualifying for such a strong word as ‘disaster’, the case study yielded a glimpse of such behavior. As discussed before, the core project team should only simulate scenarios within a range of parameter settings, where no significant negative motivational responses would be anticipated. Nevertheless, it is very likely that some of the simulations had parameter settings that flirted with the risk of a company-wide decrease in motivation. As the final implementation decisions were based on qualitative, subjective evaluations of organizational acceptance, the ‘flirting-with-disaster’ scenarios had no consequences for the present case study, but in other modeling projects the context could be imagined to be of such character, that the facilitator should pay attention to possible unwanted learning effects. In general, the literature stresses the importance of facilitation for the success of participative modeling processes.425 In the case study, it was perceived that one of the facilitator’s most crucial tasks was, while balancing the allowing for learning and the pursuing of efficiency, to keep the discussion focused on solving the critical implementation issues. Having the group of change leaders taking responsibility for identifying as well as constructively solving the critical implementation issues, commitment and ownership for the implementation plan seemed to be building up.

424

425

Salge, Markus and Peter Milling: “Who is to blame, the operator or the designer? Two stages of human failure in the Chernobyl accident”, System Dynamics Review, Vol. 22, in print, 2006, the figures 8 and 9. See discussion in chapter B.III.2.d.

D. The Usage and Utility of Participative Modeling in Change Management

III.

155

Outcomes of Participative Modeling Efforts in the Implementation of Change Programs

Discussing how big, hierarchical organizations manage change, Leavitt states, that many “do a good job of bringing new ideas into their organizational tents, but then they do a poor job of weaving those ideas into their cultures. Instead, they dump that grab bag of changes onto the labs of their middle managers, expecting them to integrate and implement the whole potpourri.” 426 A modeling project provides the opportunity to avoid such ‘dumping’ of change initiatives, and the following sub-chapters discuss possible outcomes of participative modeling projects used from the outset of larger change projects in hierarchical corporations.

1.

Modeling and simulation as a tool for transfering insights and ownership from decision-makers to implementers

Taking theoretical point of departure in the two cycles of the problem-solving process as described in chapter A, modeling efforts in a change management context focus on bridging the diagnostics and decision-making cycle with the change management cycle.427 The diagnostics and decision-making cycle is in itself a learning process among decision-makers. The decision-makers identify and discuss problems in a groping process, where, especially for highperformance management teams, they develop a shared mental model of the problems at hand.428 Following diagnostics and decision-making, formulating and launching a change project inevitably means involving more people, who have not taken part in the learning journey of the diagnostics and decision-making

426

427

428

Leavitt, Harold J.: Top Down – Why Hierarchies Are Here to Stay and How to Manage Them More Effectively, Boston, 2005, p. 98. The discussion in chapter A.I on the two cycles of problem-solving, Diagnostics and Decision-Making (cycle I) and Change Management (cycle II), is based on the problem solving process described in Schein, Edgar H.: Process Consultation, Boston, 2000, part I, p. 61. Eisenhardt, Kathleen M.: “Strategy as Strategic Decision Making”, Sloan Management Review, Spring 1999, pp. 66—67. Eisenhardt calls the mental model alignment “building collective intuition.”

156

D. The Usage and Utility of Participative Modeling in Change Management

cycle. A modeling process can utilize a system dynamics model in bridging the two cycles of the problem-solving process, transferring the main insights motivating the change and illustrating the intended outcome of the change project. Furthermore, by allowing for model refinement, parameter stipulation, and scenario testing with a high degree of empowerment among the modeling participants, the modeling process serves as an instrument ensuring the implementers true involvement and influence in the change process. Involvement and influence are often discussed as essential mechanisms in lowering change resistance and ensuring the commitment among key employees necessary for creating sustainable change. A modeling process will impact the transfer of insights and ownership at individual, group, and organizational level. On the individual level, learning of causal relationships will influence cognitive structures, and if designed well, the process in a group setting could be expected to also influence affective and conative elements as well as social norms and perceived behavioral control for the project participants.429 Consequently, based on the framework of Ajzen, one could expect behavioral changes.430 In the case study, the participants’ questionnaire answers scored relatively high on the usefulness of including the model in the project, its usefulness in facilitating discussions, the usefulness of starting the project with a preliminary model, individual learning in general, and the building of a shared view.431 From the project owners’ perspective one of the main objectives of the modeling process was establishing consensus about the change imperative, as significant change obstacles existed of both cognitive and affective characteristics.432 A year and a half after the modeling process, it was

429

430

431

432

This conclusion is based on the discussions throughout the dissertation, supported by both the organizational psychology literature and the normative, prescriptive change management literature. Ajzen, Icek: Attitudes, Personality and Behavior, Chicago, 1988, p. 133. See also discussion in chapter B.II.1. At the same time, the lowest score was given to expectations of actual implementation of the recommendations. However, in follow-up discussions this gave the impression of being due to pessimistic expectations to the required shift from headcount-orientation to cost-orientation, see further description in chapter C. The short version of the change imperative being that in times of company growth it is possible to build up capacity in low-cost locations without reducing high-cost locations - and this is a necessary strategy for the continued

D. The Usage and Utility of Participative Modeling in Change Management

157

observed, that the recommendations derived from the modeling process were implemented at an even higher speed than outlined in the implementation plan. Although many influential factors exist, the project owners give significant credit to the modeling project’s influence on the change leaders commitment to the change. At group level, the modeling process allowed for discussion and exchange of experiences, and the high-level aggregated model secured alignment of mental models and codes among key implementers across the organization. At organizational level, the change affected significantly more people than were involved in the modeling project. However, indirect influence takes place through the development of true change leaders.

2.

Refining and Aligning Implementation Plans Through Scenario Simulation

Apart from the transfer of insights and ownership, a more tangible outcome of the modeling usage is the development and cross-organizational alignment of implementation plans. In the case study, the conceptual modeling step ‘Discuss and Stipulate Parameter’ contributed to the alignment of parameters across the different business cases in the different business units.433 The step ‘Analyze Sensitivity and Identify Critical Parameters’ focused the discussion on the critical parameters, which was seen as contributing to the effectiveness and efficiency of the project. For instance, the time horizon was deliberately set to five years to focus the discussion on the present implementation challenges. It was well known to the executives that employees feared the long-term consequences for the number of jobs in high-cost locations. Deliberately limiting the modeling time horizon to five years focused the discussion on the implementation efforts, rather than on potential long-term policies and effects. Although one could argue, that this was a way of ‘hiding’ a relevant discussion from the official meetings (and leaving this issue to the unofficial meetings and coffee-breaks), it was also a way to set up a barrier against a very emotional discussion that popped up in most meetings and mentally blocked participants from addressing the implementation

433

competitiveness of the company without determining the distant future of the company. For an overview of the conceptual modeling steps in the case study, see chapter D.II.

158

D. The Usage and Utility of Participative Modeling in Change Management

challenge at hand. Furthermore, the long-term effects were indirectly addressed, as the modeling efforts and the focused discussions actually established a sense of urgency: The need to build up capacity in low-cost locations in times of growth, seeking competitive advantages and securing jobs in high-cost locations in the long run. The step ‘Simulate Scenarios’ helped to focus attention on the critical improvement opportunities of the implementation strategy and helped in the discussion of the appropriateness of the different alternative scenarios.434 The step ‘Plan for Improvements of Critical Parameters’ stimulated discussions on practical solutions to improvement opportunities and contributed to a crossorganizational alignment of implementation plans.435 In the case study, the modeling process served as a framework for discussion and exchange of experiences important for decisions on the actual detailed implementation plans. However, even though the model was perceived as addressing the core of the problem, a number of additional issues and aspects were discussed without any direct link to the model or the modeling process, but they were nevertheless important for reaching the objectives and targets as well as for developing true change leaders. It seams fair to assume, that it is a general trend that a model can only address a certain part of a given problem at hand, and that an intervention should always include sessions and discussions not directly linked to the modeling and simulation efforts.

3.

Organizational Learning in Change Management Oriented Modeling

In the system dynamics tradition, modeling projects should not only aim at creating insights of relevance in a given intervention, but also strive for the general improvement of system thinking skills among the modeling project 434

435

Snabe Birgitte: Brugtvognskoncept for Skandinavisk Motor Co., Thesis, Det erhvervsøkonomiske diplomstudium, Copenhagen Business School, Copenhagen, 1994, p. 9 and p. 89, discusses scenario evaluation and selection among alternative scenarios in terms of strategic fit, profitability, risk, competitor reactions and reversibility. A discussion of the importance and the challenges of cross-organizational alignment of robust policies can be found in Lyneis, James M.: Corporate Planning and Policy Design: A System Dynamics Approach, Massachusetts, 1980, pp. 6—9.

D. The Usage and Utility of Participative Modeling in Change Management

159

participants. With the development of change leaders being an essential part of modeling efforts in a change management context, the mere experience of a system-oriented approach is likely to contribute to the development of system thinking skills. However, change-management-oriented modeling is more focused on developing practical plans compared to traditional exploratory modeling, for which reason it seems fair to expect less significant improvement of system thinking skills. Taking a broader view of organizational learning, a modeling process in a change management context contributes to the participants’ general understanding of cross-organizational systems due to the process of refining and discussing a shared model with colleagues from other parts of the organization. This could often be expected to be a learning experience contributing to a more holistic understanding of the company. With a modeling process involving a larger number of participants developing a shared mental model, this constitutes clear elements of organizational learning.

E. Targeted Participative Modeling in Change Management System dynamics modeling in organizations has been applied to identify strategic objectives, to explore potential strategic behavior modes and, ultimately, to formulate improved organizational policies to achieve strategic advantage. Although implementation of improved policies has always been on the agenda in the field of system dynamics, the literature rarely discusses how system dynamics modeling can also be used in the change process of implementing already outlined strategic initiatives.436 The purpose of this dissertation has been to investigate the usefulness of system dynamics modeling used in a change management context, as opposed to the often seen usage of system dynamics modeling in an exploratory context. The change management context is defined here as the implementation process of a planned change initiative with already identified objectives and targets directing and framing the modeling intervention. Chapter A discussed the need for and the challenges in organizational interventions, and illustrates a problem-solving process of iterative stages with two conceptual cycles: the diagnostic and decision-making cycle, and the change management cycle.437 The practical implications of the two conceptual cycles include issues of transfer of insights and ownerships, as it will often be the case, that different people are responsible for the diagnostic and decision-making process (often top executives) and the change management process (primarily the

436

437

For processes, where modeling has been used designing alternative policies and structures, Forrester discusses the need for educating and debating with people who will be involved the implementation. The challenge often includes changing deeply embedded policies and emotional beliefs, where “it is not that people disagree with the goals, but rather how to achieve them.“ See Forrester, Jay W.: “System dynamics, system thinking, and soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, p. 247. The two cycles of the problem-solving process are inspired from the problemsolving process in Schein, Edgar H.: Process Consultation, Boston, 2000, part I, p. 61.

162

E. Targeted Participative Modeling in Change Management

operational managers). Chapter A also discussed taxonomies for change strategies, and furthermore places the usage of system dynamics modeling in an organizational intervention perspective. Chapter B mainly discussed the theories underlying the usage of system dynamics, as well as methods and procedures for the use of system dynamics modeling in organizational interventions. The system dynamics theories, methods, and procedures described in chapter B draw from the natural science literature as well as cognitive and social psychology, economics, and other social sciences literature, and constitute the main foundation for the research presented in this dissertation. Chapter C and D concentrated on the usage of system dynamics modeling in a change management context, from both a case study and a literature based perspective. The case study is a single-site case study in the action research tradition, where modeling efforts were used for the purpose of implementing a strategic initiative. Objectives and targets for the intervention were established in the decision-making process leading to the launching of the new location strategy, and a preliminary model was established prior to the initiation of the modeling project to focus and frame the modeling efforts. The model served as a cognitive framework for discussing the change imperative as well as implementation issues. The planned change intervention had a strong focus on developing the managers responsible for the implementation into change leaders through true involvement and influence in the change process, including an iteration with the top executives based on the insight gained in the modeling process. Also, the practical development and alignment of cross-organizational implementation plans was central in the process, where scenario simulations supported discussions and exchange of experiences important for decisions on the actual detailed implementation plans. Based on the evaluation feedback received, the case study yielded indications of valuable outcomes at individual level (e.g., understanding the importance and the motivation for the change), at group level (e.g., facilitation of communication, exchange of ideas and experiences, and alignment of mental models), and at organizational level (e.g., insights incorporated in actual budgets and business plans across the organization). From the case company perspective, it seems fair to conclude that the modeling process was an effective and efficient way of refining a strategy that fulfils board objectives as well as preparing the grounds for sustainable implementation. Especially due to the sensitive nature of the topic, it was interesting to observe how the modeling and simulation efforts helped to direct and focus the discussions and facilitate individual learning. However, it is not

E. Targeted Participative Modeling in Change Management

163

possible to know whether a different approach would have been more effective or efficient, as the case study was a single-site study without any test group. The usage of system dynamics modeling for change management purposes could be called a ‘targeted participative modeling’ process, addressing how to implement a strategic initiative with already established objectives and targets. The managers responsible for an implementation process often played only a small part, if any, in the decision-making process from which the strategic initiative originated. Targeted participative modeling looks especially promising with regards to: (1) The usage of modeling and simulation as a tool to transfer insights and ownership from decision-makers to implementers.438 In the pursuit of sustainable change, iterations with top executives might be relevant, with the dual objective of adjusting the strategic initiative according to implementation issues and giving the managers responsible for implementation true influence on the entire change process. 439 (2) The refining and aligning of cross-organizational implementation plans through scenario simulation. Although simulations should not serve as ‘an evaluator’ of scenarios, they can aid in the investigation of expected system behavior for alternative scenarios and, as such, support decisionmaking at the operational level critical for successful implementation. The positive indications of both the effectiveness and efficiency yielding from the case study could give rise to expectations to potential significant usage of targeted participative modeling, as larger organization often launch strategic initiatives involving the need to bridge the decision-making process with the implementation process. A question arising from the case study is whether a modeling approach as used in the case study would also work if the problem at hand had been significantly more complex. For some complex problems, a business simulator might prove more suitable compared to a modeling process in

438

439

With regards to the transfer of insights, a targeted participative modeling process has certain similarities with the usage of management simulators. Communication skills, interpersonal skills and other personal leadership skills influence effectiveness in the change process, together with company factors including change readiness, change capabilities and culture, see Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, 2001, p. 144.

164

E. Targeted Participative Modeling in Change Management

a change management context. Alternatively, a modeling process supporting the total problem-solving process might yield additional benefits; i.e. modeling used in an iterative set-up supporting both the decision-making and the change management cycle. Further research in the design and evaluation of targeted participative modeling processes could address this question, as well as barriers for practical application (e.g., the challenge of acceptance and availability of modeling facilitators). The evaluation of targeted participative modeling processes constitutes a significant challenge due to the complexity of social systems and the difficulties in measuring learning processes.440 However, the indications of effectiveness and efficiency of the usage of targeted participative modeling discussed in this dissertation improve the prospects for further research into balanced constitutional and causal modeling processes, as opposed to the one-sided commitment to causal modeling typically seen in the field of system dynamics.441

440

441

Some of the same challenges are seen in the design and evaluation of business simulators, see Größler, Andreas: Entwicklungsprozess und Evaluation von Unternehmenssimulation für lernende Unternehmen, Frankfurt am Main, 2000, p. 178. See Van der Smagt, Ton: “Causation and Constitution in System Dynamics: Modelling a Socially Constituted World”, Journal of Systems Research and Behavioral Science, Vol. 23, No. 4, in print, 2006, pp. 13—14, as well as the discussion in chapter D.I.

Appendices Appendix A: Model Parameter Descriptions........……………………. 167 Appendix B:

Model Equations…………………………………………… 171

Appendix C:

Equations for Stock Initializations ………………………… 183

Appendix D

Preliminary Model ………………………………………… 185

Appendix E:

Model without ‘Rate-on-Rate’ Modeling………………….. 187

Appendix F:

Facilitator Observations and Key Quotes From Interviews………………………….. 189

Appendix A: Model Parameter Descriptions The following three tables give a short description of the model input parameters. The first table describes parameters particularly relevant to high-cost locations. The next table describes parameters particularly relevant to low-cost locations, and the third table describes parameters mainly relevant for transfer of tasks and build-up of employees in low-cost locations.

Parameter (and Units)

Description

INI TOTAL FTE HC (Person)

Number of R&D employees in High Cost locations (new + rookies + experienced by Year end 2004)

HC QUIT FRACTION (1/Month)

The fraction of people quitting HC locations (internal rotation in the company not included)

HC TRAINING TIME (Month)

Training time (basic new-hire training)

HC TRAINING COST (EUR/(Person*Month))

Basic new-hire training costs per person per month

HC ROOKIE TIME (Month)

On-the-job training

HC ROOKIE PRODUCTIVITY (Dimensionless)

Productivity factor for Rookies in HC (Rookies are inexperienced and therefore have reduced productivity)

INI HC AVERAGE PERSON COSTS (EUR/(Person*Month))

Total cost (including travel, rent, licenses etc.) allocated per HC FTE by Year-end 2004. Salaries are only XX% of this

HC PERSON COST INCREASE RATE (1/Month)

Increase rate in HC person costs

Appendix A, Table 1: Parameters mainly relevant to high-cost locations

168

Appendix A

Parameter (and Units)

Description

INI TOTAL FTE LC (Person)

Number of R&D employees in Low Cost locations (new + rookies + experienced by Year-end 2004)

LC QUIT FRACTION (1/Month)

Fraction of people leaving LC locations

LC TRAINING TIME (Month)

Training time for the basic training of newly hired employees (classroom training)

LC TRAINING COST (EUR/(Person*Month))

Basic training cost for newly hired employees (per person per month)

LC TOTAL ROOKIE TIME (Month)

Time spent on on-the-job training after finished classroom training. This includes both the time as normal rookie and the time as hand-over rookie

LC ROOKIE PRODUCTIVITY (Dimensionless)

Productivity factor for Rookies (although only in the part of the time NOT used to hand over tasks)

INI LC AVERAGE PERSON COSTS (EUR/(Person*Month))

Total cost (excluding travel) allocated to LC FTE by Year-end 2004. Salaries are only XX% of this. The cost will grow with the rate below

LC PERSON COST INCREASE RATE (1/Month)

Increase rate in LC person cost

LC ONGOING TRAVEL COST (EUR/(Person*Month))

Average Travel + Hotel + Rented cars etc. for all LC employees

LC PRODUCTIVITY REDUCTION (Dimensionless)

Productivity reduction due to low average experience (1-2 years in LC vs. 5-10 years in HC)

Appendix A, Table 2: Parameters mainly relevant to low-cost locations

169

Appendix A

Parameter (and Units)

Description

REPLACEMENT IN HC VS. LC (Dimensionless)

Factor for how many of HC quit who will be replaced in HC. When a person leaves a high-cost location, the job is either given to a new-hired employee, or (when replacement rate lower than 1) the job is given to an employee who has transferred his or her job to a low-cost location. The replacement rate is only below 1 for a limited period of time (to finance the initial build up in lowcost locations).

ADDITIONAL GROWTH (Dimensionless)

Factor for the ramp-up of productive LC employees (fixed for the first period, then gradually decreasing to zero after the 36 th month)

HAND-OVER FRACTION The fraction of the time a newly hired LC Rookie spends on hand-over tasks (as opposed to normal OF LC ROOKIE TIME rookie on-the-job training). Example: if the (Dimensionless) parameter is 1/3, and the total LC Rookie time is 6 months, an LC New Hire will after finished classroom training spend 2 months as Hand-OverRookie (with zero productivity), and 4 months as “normal rookie” (with the normal LC ROOKIE PRODUCTIVITY). LC HANDOVER TRAVEL Travel costs related to hand-over, i.e. only effect hand-over fraction of Rookies. COST (EUR/(Person*Month)) HC CAPACITY USE ON LC TRAINING (Dimensionless)

1-on-1 hand-over will result in XX% reduction of the HC employee’s productive time (LC employee will have 0 productivity)

Appendix A, Table 3: Parameters mainly relevant for transfer of tasks and build-up of employees in low-cost locations

Appendix B: Model Equations The two following tables give an overview of the significant equations in the model. The first table describes the main equations for the rates influencing stock levels, and the next table describes the main equations for calculation of the monthly production, and well as monthly costs. After the two tables, a complete list of model equations is to be found.

Main equations for the rates influencing stock levels HC hire = HC quits * REPLACEMENT IN HC VS. LC LC replacement hire = LC quit Comment: This hiring only compensates for the employee turn-over LC new hire = + Productive FTE LC*ADDITIONAL GROWTH + LC replacing HC quit Comment: LC replacing HC quit = HC quit – HC hire The rates between stages are calculated as delay-functions of the inflow-rates. For the training period, a high-order delay was used to imitate a pipeline delay, as this is a fixed period of time for each employee. For the period as a Rookie, a lower order delay was used, to reflect the variability in the learning curve for individuals Appendix B, Table 1: Main equations influencing stock levels

172

Appendix B

Main equations for production per month and cost per month Handover capacity reduction = H-O-R FTE LC * HC capacity use on hand-over Comment: Each H-O-R (Hand-Over-Rookie) in a low-cost location will need physically to sit together with the person from whom he or she is taking over tasks. For this reason, the experienced person in the high-cost location will have reduced capacity (an average stipulated in the parameter: HC capacity use on hand-over) Productivity per month = + Productive FTE HC + HC ROOKIE PRODUCTIVITY*Rookie FTE HC - handover capacity reduction + Productive FTE LC*(1-LC PRODUCTIVITY REDUCTION) + LC ROOKIE PRODUCTIVITY*Rookie FTE LC Comments: 1) Unit of productivity is person/month (the case company often uses manyear, man-month or man-days as unit for projects or production) 2) Number of productive days in LC higher than in HC. This is not included in the model, but “equals out” with coordination overhead Cost per month = +HC person cost*total FTE HC +New hired FTE HC*HC TRAINING COST +(LC person cost+ LC ONGOING TRAVEL COST)*total FTE LC +New hired FTE LC*LC TRAINING COST +handover travel costs Comment: HC person costs and LC person costs develop over time with the person cost increase rate. In Vensim this is can be managed by treating person costs as stocks: input rate = PERSON COST INCREASE RATE*LC person cost (an initialized with the initial values of the person costs) Appendix B, Table 2: Main equations influencing production per month

Appendix B

173

Starting below is a complete list of the model equations from the case study model. According to the agreement with the case study company, the most confidential numbers (such as average employee costs etc.) have been made unrecognizable, through having ‘NN’ replacing the first digits. A number of the parameters do not relate directly to the model, but serve the purpose of producing nice output graphs.

List of equations (generated in Vensim):

Formulas ADDITIONAL GROWTH= (0.03*PULSE(0, 61) +RAMP(-0.03/36, 12, 48)) LC new hire= +(Productive FTE LC*ADDITIONAL GROWTH) +LC replacing HC quit +"HC micro-site optim." +"LC micro- site optim."

Dimension

Dmnl

Person/Month

INI LC AVERAGE PERSON COST= NN000/12 EUR/(Person*Month) LC PERSON COST INCREASE RATE= +(LN(1.15)/12)*PULSE(0, 24) +(LN(1.1)/12)*PULSE(24, 36)

1/Month

LC person cost= INTEG ( +LC PERSON COST INCREASE RATE*LC person cost, INI LC AVERAGE PERSON COST) EUR/(Month*Person)

174 "HC job-trained"= DELAY N(HC trained, HC ROOKIE TIME, Rookie FTE HC/HC ROOKIE TIME, 3)

Appendix B

Person/Month

Productive FTE HC= INTEG ( +"HC job-trained"-HC quit-"HC micro-site optim.", INI TOTAL FTE HC/ (1+(HC ROOKIE TIME+HC TRAINING TIME)*HC QUIT FRACTION)) Person "LC repl. hire"= LC quit

Person/Month

LC 2a= DELAY N( ("LC repl. hire"), LC TRAINING TIME, (New hired FTE LC/LC TRAINING TIME) , 12)

Person/Month

LC 1a= DELAY N( LC new hire, LC TRAINING TIME, 0, 12)

Person/Month

New hired FTE LC= INTEG ( +"LC repl. hire"+LC new hire -LC 2a-LC 1a, LC TRAINING TIME/ (LC TRAINING TIME+LC TOTAL ROOKIE TIME) *INI TOTAL FTE LC/ (1+(1/((LC TRAINING TIME+LC TOTAL ROOKIE TIME) *LC QUIT FRACTION)))) cost per month= +HC person cost*total FTE HC +New hired FTE HC*HC TRAINING COST

Person

Appendix B

175

+(LC person cost+LC ONGOING TRAVEL COST)*total FTE LC +New hired FTE LC*LC TRAINING COST +handover travel costs EUR/Month Rookie FTE LC= INTEG ( +LC 2a +LC 1b -LC 2b-LC 1c, LC TOTAL ROOKIE TIME/ (LC TRAINING TIME+LC TOTAL ROOKIE TIME) *INI TOTAL FTE LC/ (1+(1/((LC TRAINING TIME+LC TOTAL ROOKIE TIME) *LC QUIT FRACTION)))) LC 2b= DELAY N(LC 2a, LC TOTAL ROOKIE TIME, Rookie FTE LC/LC TOTAL ROOKIE TIME, 3)

Person

Person/Month

Productive FTE LC= INTEG ( +LC 2b+LC 1c -"LC micro- site optim."-LC quit, INI TOTAL FTE LC/ (1+(LC TOTAL ROOKIE TIME+LC TRAINING TIME)*LC QUIT FRACTION))

Person

"H-O-R FTE LC"= INTEG ( LC 1a-LC 1b, 0)

Person

LC 1b= DELAY N(LC 1a, LC TOTAL ROOKIE TIME* "HAND-OVER FRAKTION OF LC ROOKIE TIME", 0, 12)

Person/Month

LC 1c= DELAY N(LC 1b, LC TOTAL ROOKIE TIME* (1-"HAND-OVER FRAKTION OF LC ROOKIE TIME"), 0, 3)

Person/Month

LC QUIT FRACTION= 0.07/12

1/Month

176 handover capacity reduction= "H-O-R FTE LC"*"HC CAPACITY USE ON HAND-OVER" handover travel costs= "H-O-R FTE LC"*"LC HAND-OVER TRAVEL COST"

Appendix B

Person

EUR/Month

total FTE LC= New hired FTE LC+Productive FTE LC +Rookie FTE LC+"H-O-R FTE LC"

Person

LC TOTAL ROOKIE TIME= 6

Month

production per month= (+Productive FTE HC +HC ROOKIE PRODUCTIVITY*Rookie FTE HC - handover capacity reduction +((1-LC PRODUCTIVITY REDUCTION)*Productive FTE LC) +LC ROOKIE PRODUCTIVITY*Rookie FTE LC) *make unit per month

"HAND-OVER FRAKTION OF LC ROOKIE TIME"= 0.333

Person/Month

Dmnl

output old year cost= +IF THEN ELSE(Time=4, Nice Graph Year Cost/TIME STEP, 0) +IF THEN ELSE(Time=16, Nice Graph Year Cost/TIME STEP, 0) +IF THEN ELSE(Time=28, Nice Graph Year Cost/TIME STEP, 0) +IF THEN ELSE(Time=40, Nice Graph Year Cost/TIME STEP, 0) +IF THEN ELSE(Time=52, Nice Graph Year Cost/TIME STEP, 0) +IF THEN ELSE(Time=64, Nice Graph Year Cost/TIME STEP, 0) EUR/Month

Appendix B

177

input year cost= +IF THEN ELSE(Time=0, cost per month*12*STY month/TIME STEP, 0) +IF THEN ELSE(Time=12, Year cost/TIME STEP, 0) +IF THEN ELSE(Time=24, Year cost/TIME STEP, 0) +IF THEN ELSE(Time=36, Year cost/TIME STEP, 0) +IF THEN ELSE(Time=48, Year cost/TIME STEP, 0) EUR/Month Nice Graph Year Cost= INTEG ( input year cost-output old year cost, 0) adding= production per month*1/12 summing= cost per month LC FTE fraction of total FTE= total FTE LC/total FTE

EUR

Person/Month

EUR/Month

Dmnl

LC replacing HC quit= HC quit-HC hire

Person/Month

LC quit= Productive FTE LC*LC QUIT FRACTION

Person/Month

"HC CAPACITY USE ON HAND-OVER"= 0.5 LC TRAINING COST= NN00

Dmnl

EUR/(Person*Month)

"LC micro- site optim."= (1-HC FRACTION OF MICROSITES)*"MICRO-SITE OPTIMIZATION" Person/Month LC PRODUCTIVITY REDUCTION= 0.2 +0*0.2*PULSE(0, 48)

Dmnl

178

Appendix B

LC ONGOING TRAVEL COST= NN0

EUR/(Person*Month)

"LC HAND-OVER TRAVEL COST"= NN00

EUR/(Person*Month)

cost per production= cost per month/production per month

EUR/Person

total FTE= total FTE HC+total FTE LC

Person

HC quit= Productive FTE HC*HC QUIT FRACTION

Person/Month

HC hire= HC quit*"REPLACEMENT IN HC VS. LC"

Person/Month

"MICRO-SITE OPTIMIZATION"= 0*(NN*PULSE(1, 1)+NN*PULSE(12, 1) +NN*PULSE(24, 1))

Person/Month

HC FRACTION OF MICROSITES= 0.5 "HC micro-site optim."= "MICRO-SITE OPTIMIZATION" *HC FRACTION OF MICROSITES

Dmnl

Person/Month

STY month= 1

Month

HC person cost= INTEG ( +HC PERSON COST INCREASE RATE*HC person cost, INI HC AVERAGE PERSON COSTS) EUR/(Month*Person) INDEX production per month= (production per month*STY month)/startindex production per month

Dmnl

Appendix B

179

INI HC AVERAGE PERSON COSTS= NN0000/12

EUR/(Month*Person)

HC PERSON COST INCREASE RATE= LN(1.025)/12

1/Month

INDEX cost per month= (cost per month*STY month)/startindex cost per month

Dmnl

startindex production per month= INTEG ( production per month-production per month, production per month*STY month)

Person

INDEX cost per production= INDEX cost per month/INDEX production per month

Dmnl

startindex cost per month= INTEG ( cost per month-cost per month, cost per month*STY month)

EUR

INI TOTAL FTE HC= NN20

Person

Year production= INTEG ( adding-reset production every year, 0)

Person

make unit per month= 1 reset production every year= +IF THEN ELSE(Time=12, +IF THEN ELSE(Time=24, +IF THEN ELSE(Time=36, +IF THEN ELSE(Time=48,

1/Month

Year Year Year Year

production/TIME production/TIME production/TIME production/TIME

Year cost= INTEG ( +summing-reset cost every year, 0)

STEP, STEP, STEP, STEP,

0) 0) 0) 0)

Person/Month

EUR

180 reset cost every year= +IF THEN ELSE(Time=12, +IF THEN ELSE(Time=24, +IF THEN ELSE(Time=36, +IF THEN ELSE(Time=48,

Appendix B

Year cost/TIME STEP, Year cost/TIME STEP, Year cost/TIME STEP, Year cost/TIME STEP,

HC trained= DELAY N(HC hire, HC TRAINING TIME, New hired FTE HC/HC TRAINING TIME, 12)

0) 0) 0) 0)

EUR/Month

Person/Month

LC ROOKIE PRODUCTIVITY= 0.5 HC TRAINING COS NN00

Dmnl

EUR/(Person*Month)

LC TRAINING TIME= 2

Month

New hired FTE HC= INTEG ( HC hire-HC trained, HC TRAINING TIME/(HC TRAINING TIME+HC ROOKIE TIME) *INI TOTAL FTE HC /(1+(1/((HC TRAINING TIME+HC ROOKIE TIME) *HC QUIT FRACTION))))

Person

HC ROOKIE PRODUCTIVITY= 0.5 HC TRAINING TIME= 3 "REPLACEMENT IN HC VS. LC"= 0.4*(PULSE(0, 12) +PULSE(12, 12) +PULSE(24, 12)) +1*(PULSE(36, 12) +PULSE(48, 12))

Dmnl

Month

Dmnl

Appendix B

181

total FTE HC= Rookie FTE HC+Productive FTE HC+New hired FTE HC

Person

HC ROOKIE TIME= 6

Month

Rookie FTE HC= INTEG ( HC trained-"HC job-trained", HC ROOKIE TIME/(HC TRAINING TIME+HC ROOKIE TIME) *INI TOTAL FTE HC /(1+(1/((HC TRAINING TIME+HC ROOKIE TIME) *HC QUIT FRACTION))))

Person

HC QUIT FRACTION= 0.0NN/12 INI TOTAL FTE LC= NN0

1/Month

Person

FINAL TIME = 60 ~ The final time for the simulation.

Month

INITIAL TIME = 0 ~ The initial time for the simulation.

Month

SAVEPER = TIME STEP ~ The frequency with which output is stored.

Month

TIME STEP = 0.25 ~ The time step for the simulation.

Month

Appendix C: Equations for Stock Initializations The stocks in the model are initialized based on the condition for equilibrium, where the total number of hirings in both low cost locations and high cost locations equals the number of employees leaving. For both aging chains, the three stocks to be initialized are: Productive FTE, New hire FTE, and Rookie FTE. The model input parameters used to calculate the initial values are: INI total FTE, QUIT FRACTION, TRAINING TIME, and ROOKIE TIME. Three equilibrium equations are: (1): INI total FTE = Productive FTE + New hire FTE + Rookie FTE (2): New hire FTE+ Rookie FTE = Productive FTE * QUIT FRACTION * (TRAINING TIME+ROOKIE TIME) (3): New Hire FTE = (TRAINING TIME/(TRAINING TIME+ROOKIE TIME) ) *(New hire FTE+ Rookie FTE) resulting in the following stock initializations (see workings on next page):

• •

Productive FTE (ini) = INI TOTAL FTE /(1+(ROOKIE TIME+TRAINING TIME) *QUIT FRACTION) New hire FTE (ini) = (TRAINING TIME / (TRAINING TIME+TOTAL ROOKIE TIME)) * INI TOTAL FTE /(1+(1/((TRAINING TIME+TOTAL ROOKIE TIME)*QUIT FRACTION)))

•

Rookie FTE (ini) = (ROOKIE TIME/(TRAINING TIME+ROOKIE TIME)) *INI TOTAL FTE /(1+(1/((TRAINING TIME+ROOKIE TIME)*QUIT FRACTION)))

(and furthermore is the stock of LC H-O-R Rookies initialized with 0

TIME+LC ROOKIE TIME)*LC QUIT FRACTION)))

New hire FTE = LC TRAINING TIME/(LC TRAINING TIME+LC ROOKIE TIME)**INI TOTAL FTE LC/(1+(1/((LC TRAINING

And the initial values of each of New hire FTE and Rookie FTE are then found by the equation (3):

<=> (New hire FTE + Rookie FTE) = INI total FTE / (1+(1/ QUIT FRACTION * (TRAINING TIME+ROOKIE TIME) )

FTE <=> INI total FTE = (New hire FTE + Rookie FTE) * (1+(1/ QUIT FRACTION * (TRAINING TIME+ROOKIE TIME) )

INI total FTE = ((New hire FTE+ Rookie FTE) / QUIT FRACTION * (TRAINING TIME+ROOKIE TIME))+New hire FTE + Rookie

and also, when the result of Productive FTE from equation (2) is entered to equation (1):

<=> Productive FTE = INI total FTE / (1+(TRAINING TIME+ROOKIE TIME)* QUIT FRACTION

<=> INI total FTE = Productive FTE + Productive FTE * QUIT FRACTION *(TRAINING TIME+ROOKIE TIME)

gives when New hire FTE + Rookie FTE from equation (2) is entered into equation (1):

(3): New Hire FTE = (TRAINING TIME/(TRAINING TIME+ROOKIE TIME) )*(New hire FTE+ Rookie FTE)

(2): Productive FTE * QUIT FRACTION * (TRAINING TIME+ROOKIE TIME) = New hire FTE+ Rookie FTE

(1): INI total FTE = Productive FTE + New hire FTE + Rookie FTE

The three equilibrium equations:

Initializing of the stocks in equilibrium (both LC and HC) :

184 Appendix C

INI TOTAL FTE LC

LC hire

LC TRAINING TIME

ADDITIONAL GROWTH

LC ROOKIE TIME

LC ROOKIE PRODUCTION FRACTION

LC QUIT FRACTION

LC PERSON COST

cost per month

total FTE HC

HC TRAINING COST HC HIRE COST

HC PERSON COST DEVELOPMENT RATE

HC person cost

MICRO -SITE OPTIMIZATION LC FRACTION OF MICROSITES

Appendix D, Figure 1: The preliminary model in the case study

TRANSFER FACTOR LC TRAINING COST LC HIRE COST

total FTE LC

HC ROOKIE TIME

HC ROOKIE PRODUCTION FRACTION

REPLACEMENT FACTOR HC HC TRAINING TIME

Rookie New hired Productive FTE HC HC FTE HC HC FTE HC HC trained hire job-trained HC microsite optim.

HC quit

HC QUIT FRACTION

production per month

Rookie Productive New hired FTE LC LC FTE LC LC FTE LC LC quit trained job-trained & siteopt.







INI HC AVERAGE PERSON COSTS

INI TOTAL FTE HC

Appendix D: Preliminary Model

Appendix E: Model without ‘Rate-on-Rate’ Modeling The next page (Appendix E, Figure 2) is an adjusted version of the model, where the out-flow rates from the stocks are modeled as simple fractions of the level-values. This way, rate-on-rate modeling is avoided, which is often recommended in the literature. Some of the stocks had to be split up in two parts, in order to use this approach. Figure 1 shows a simulation run with the same parameter setting as used in the simulation runs in Chapter C. It should be noted that the main trends – and thereby the main model insights – are the same as in the original model, also for year 1, even though the adjusted model results in the new hiring policy influencing the stock of experienced employees from nearly the very beginning.

Index Cost per Production 1.2 1.15

1 1

1.1

1 1 1

1.05 32 1 0

41 3 6

1 4

4

2

3 2 4

3 2 4

3

3 4

2

4

2

3 2 4

9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 Months 1 1 1 1 1 INI Dmnl Dmnl Base run 2 2 2 2 2 Dmnl 40% HC replacement 3 3 3 3 Faster training and hand-over Dmnl 4 4 4 4

Appendix E, Figure 1: Simulation run of adjusted model (avoiding rate-on-rate)



LC ROOKIE PRODUCTIVITY

LC QUIT FRACTION

HAND-OVER FRAKTION OF LC ROOKIE TIME

LC TOTAL ROOKIE TIME

LC TRAINING TIME



LC flow input

cost per month handover travel costs



REPLACEMENT IN HC VS. LC

INI LC AVERAGE PERSON COST

INI HC AVERAGE PERSON COSTS

HC TRAINING TIME

total FTE HC

HC TRAINING COST

HC PERSON COST INCREASE RATE

Appendix E, Figure 2: Model without rate-on-rate modeling



LC HAND-OVER TRAVEL COST

INI TOTAL FTE LC INI TOTAL FTE HC

Rookie New hired Productive FTE HC HC FTE HC HC FTE HC HC job-trained trained hire

HC ROOKIE TIME

HC ROOKIE PRODUCTIVITY

HC CAPACITY USE ON HAND-OVER

LC person cost HC person cost

LC PERSON COST INCREASE RATE

LC ONGOING TRAVEL COST

LC TRAINING COST

total FTE LC (arrows hidden)

HC quit

HC QUIT FRACTION

production per month

LC PRODUCTIVITY REDUCTION

LC add. hire LC A2 LC A1 Rookie LC A3 New hired H-O-R LC quit FTE LC FTE LC A FTE LC A Productive A FTE LC Rookie New hired FTE LC BLC B2 FTE LC B LC B1 LC repl. hire

ADDITIONAL GROWTH

non-replaced HC quit



handover capacity reduction

188 Appendix E

Appendix F: Facilitator Observations and Key Quotes From Interviews The observations and interview quotes are structured according to the evaluation framework described in chapter C.III. When nothing is indicated, observations and interview quotes were made during the project, i.e. fourth quarter 2004.

Personal reactions to the modeling process: Very early in the modeling process, an engaged and vital discussion started, showing a positive attitude in the sessions. A few persons were resistant to the process, due to disagreement with the intervention objectives. It is an interesting point, however, that not even the core-group person disagreeing with the objectives of the process could “resist the fun of modeling” when he took part in modeling sessions, as he and the other core members were very mathematically skilled and interested individuals. “A few participants did not agree with the business objectives, and did therefore never really buy in” Gain of learning, and changes in goal structures and mental models: In the development of the preliminary model, the very first results already took form as the project owners gained some interesting insights. Something first considered as a potential mistake in the model turned out to be an important insight, and it became clear that one decision, that had just been made, had a stronger negative impact in year 1 than anticipated, and it was therefore decided to modify the decision, and make the transition over a longer time-span. Through the discussions and model simulations in the modeling workshops, the core project team gained

Observation

Interview

Observation

Observation

190

Appendix F

insights and exchanged experience relating to the location strategy. Also, the model was a framework for the setting of parameters to be used in the business case in each of the business areas. Through investigations of effects of the changes in the different parameters of the model, the core team identified effective optimization opportunities, as well as sensitivity risks. Some of the most important insights gained were the understanding of the ‘reinforcing growth loop’ motivating the intervention, of how relatively few non-replacements in high-cost countries could compensate for the costs of building up the required volume of R&D employees in low-cost countries, as well as distinct benefits of reducing training and hand-overtime compared to reducing costs of training and traveling.

Observation

Often, simulations served as an “eye-opener”. In a few cases a parameter was perceived as “not possible to reduce”, but through simulations with increased value, a strong impact was seen, and the individuals then opened up for discussion on what it would take to optimize a certain parameter, e.g. reduce hand-over time.

Observation

The initial setting of each of the most important parameters could be discussed for hours in both workshops and other related meetings. For example, it was a widely accepted “fact” among many of the project participants, that employees in low-cost countries often stayed for only 1-2 years, because as soon as they attained experience in R&D, they could get a higher paying job in a high-cost country. Through the parameter stipulation, facts came on the table, documenting a very low employee turnover in the low-cost countries. (This could be viewed in the context of Ackoff’s morale: “There is nothing so deceptive as an apparent truth”).

Observation

191

Appendix F

In the modeling sessions, especially in the parameter stipulation, knowledge and experience exchange took place across the business units. Especially one business unit had already high-scale experience with build-up of resources in low-cost countries. Using this modeling approach served as a forum for the transfer of best practices.

Observation

“The preliminary model was important to get confirmation on the feasibility of the objectives, and the preliminary model also gave a better understanding of the dynamics of the problem”

Interview

“Parameter discussions were effective in challenging assumptions”

Interview

“Simulations were strong in showing the importance of the different parameters”

Interview

“The exchange of Best Practices was one of the objectives for starting a cross-business unit process in the first place”

Interview

Commitment to the outcome of the modeling sessions: Modeling participants often argued supporting the insights gained in the modeling, when presenting the results in other meetings – but there were also a few examples, when this was not the case, primarily in situations with divergence between insights and personal interests.

Changes in behavior: “In general the team members developed business cases in compliance with the modeling insights and results, with only one exception”

Observation

Interview

192

Group communication: The discussion seemed to be both very structured and very open and frank. The result-oriented process, however, did not leave time to go into depth in all of the relevant discussions, but due to the structure, most of the time invested by the participants in discussions was used very effectively.

Appendix F

Observation

A couple of times, the modeling helped to take focus away from discussions, when simulations proved the low importance of a parameter. Therefore, there was little relevance of the continuous discussion about the exact stipulation of the value.

Observation

“The discussion improved radically compared to the rather unstructured communication we had in the project, before we decided to use system dynamics. The model directed the discussions back to the core of the problem”

Interview

Communication and consensus: Opinions on parameters were often very different within the core project team, and the model proved to function as a structure for fact-finding and alignment of perceptions. “The approach makes it difficult for people to play politics” Consensus: Opinions on the importance of different causalrelationships differed initially, but through the modelbuilding process a more shared understanding of the problem and its dynamics was created. The discussion of the parameters often initiated longer discussions on how the strategy could and should be

Observation

Interview

Observation

Observation

193

Appendix F

executed, as the parameter setting reflected implementation decisions; e.g. the logistic and cost model of traveling, how to structure knowledge transfer, etc. Common language: Within the core project team, there was a tendency to increased alignment, but this was difficult to transfer to non-core members in the relatively short meetings with these people. Parts of the language did spread to some extent, such as “one employee is one employee” regardless of the type of location. The factor for reduced productivity in low-cost countries only reflects a lower average experience-level. A stronger outcome on this dimension would have required a less result-oriented process, with more time to in-depth discussion. “Even more of an effect – especially outside the core team – would have been better”

Transfer of insights: The model clearly confirmed some viewpoints that the project team wanted to communicate to the board and the corporate controlling. Whereas these insights did not have much “newness” value, it was very valuable to have a model that distinctly and clearly “proved” the matter. These type of insight included the worse-before-better effect, implying that the division even receiving a relatively large number of additional head-counts in year 1, would have no additional productivity, but rather a slightly reduced productivity. Also, the model showed very clearly that even the relatively large growth in the fraction of low-cost employees compared to the total number of employees, does not result in a decreased cost per produced development hour, as the inflation has stronger influence than the benefits to be realized through a location strategy of the discussed scope.

Observation

Interview

Observation

194

Appendix F

Running a few simulations appeared to be very convincing in the discussions with non-core stakeholders.

Observation

“the model made the strategy very transparent, with clear definitions – and was better than words for communication”

Interview

System changes: The board accepted the business cases developed in the project, and the hirings for 2005 went approx. as planned. Also, the business case was implemented in the 3 years business plan, and the execution should follow the plans.

Observation

The business unit that disagreed most with the project objectives has actually been “overperforming” year-end 2005 with regards to hirings in low-cost locations.

Observation

“The business cases are approved by the board, and incorporated in the budgets and business plans”

Interview

Results: 1½ years after the approval of the detailed implementation plans, the FTE ratio between low-cost and high-cost locations has increased beyond the plan. This must also be seen in relation to the fact, that due to the success of the company, more hiring were needed and approved than originally planned. “The modeling clearly helped in getting the managers committed to the strategy to use the current success to build up capacity in low-cost locations, seeking to avoid having to lay off employees in less fortunate periods.”

Observation (Q2 2006)

Interview (Q2 2006)

195

Appendix F

Efficiency: The project kept established deadlines. Some disturbance and discussion took place due to the fact, that the intervention also encompassed many elements not included in the modeling.

Observation

Especially in the beginning of the project, there was a tendency among core team members to think of the modeling as an additional task, increasing the workload in an already stressed period of time. But on the other hand, the model helped both to structure and to facilitate the discussions, which are likely to have reduced the overall time spent by the core team. To obtain this efficiency, however, a lot of efforts were made in workshop and meeting preparation by the facilitator and project owners were made.

Observation

Overall, it seemed very efficient to use the chosen software to make a shared model on a high abstraction level, with easy simulation options. Some improvement in the software, however, would be welcomed, as quite a lot of “behind the scenes” work was needed in order to create nice and effective output-graphs in separate views, avoiding waisting time and thereby creating irritation towards the modeling efforts. Also, even relatively small changes in structures could be very time-consuming to implement in the chosen software (Vensim).

Observation

“It was a very structured and effective process”

Interview

“The project progressed even better than planned due to discipline and focus in the process”

Interview

“Maybe even too efficient: more difficult to act politically in the budget-negotiation” (said with a smile)

Interview

196

Quality in results: For SD practitioners, the model seems very simple, but it should be recognized that the project team first tried to handle the problem using a normal Excel-model, which became a complicated “black box”, where it was difficult to see and understand how the different parameters influenced the model.

Further use of SD: One of the project owners has been partly involved in a later project using system dynamics. This later modeling project was initiated in another part of the organization, and due to the positive experiences in the location strategy project, the project owner positively supported the idea in his new role as project sponsor.

Intervention driven by business objectives and targets: The intervention was initiated with clear objectives and targets (directions from the board). Only a modeling process supporting this type of intervention was considered by the project owners. No participants questioned this circumstance. “Most of our strategic projects are initiated with very clear business objectives and targets”

Project framing: The project owners had no intentions of starting a group model building process from clean sheets of paper with the risk of losing control. This might be a general trend in the corporate environment; that executives have a clear view of the direction they want to drive a given change, and that they will not take the risk that a model could show contradicting results, which in their view could be

Appendix F

Observation

Observation (Q2 2006)

Observation

Interview

Observation

197

Appendix F

due to hidden errors in the model or the problem being addressed or conceptualized erroneously. Trust in the modeling process was gained through the preliminary model. Compared to exploratory modeling, the targeted participative modeling approach restricts the problemsolving process (with regards to “what to do,” not in “how to do”). It is difficult to say whether this had negative impacts on the participants’ ownership and trust in the model. The questionnaires do not explicitly include questions regarding this possible impact of a preliminary model, due to the problem of measurements influencing the system (in this case creating negative attitudes). The preliminary model confirmed some intuitive expectations of the project owners, and showed to be an effective mean of communicating these cause-effect relationships, which was a cornerstone in continuing the modeling efforts.

“A few participants did not agree with the business objectives and for that reason also not with the process, but nevertheless the process forced them in the decided direction, and through the modeling they gained some of the insights motivating the intervention in the first place”

Observation

Observation

Interview

“We were open about the premises for the process, and participants should therefore not feel in the slightest way manipulated” (This was the answer to a question, if the use of a preliminary model and fixed business objectives could have caused the participants to feel somewhat manipulated)

Interview

“Initially I was a bit skeptical, but along the process I started to trust the model”

Interview

198

Appendix F

“We got were we wanted to”

Interview

“The modeling helped changing the focus from ‘seeing only problems’ to discussing sustainable and fair execution”

Interview

Context comparative conditions: The problem was more politically sensitive than truly messy. There were clearly defined objectives and targets.

Observation

The case company has a strong tradition for employee empowerment and is a relatively un-hierarchical organization.

Observation

Attitude to intervention: it was a top-down decision to initiate the intervention, initially against the “true wish” of many of the participants, although most of them could agree with the rationale behind the intervention.

Observation

There was a technical environment with young and highly educated people with a tradition of mathematical and “rational” problem solving. All participants were perceived as high-performers and have been with the company for years.

Observation

Mechanism comparative conditions: A preliminary quantitative model was used to investigate whether a model was appropriate to illustrate the change objectives. The preliminary model showed what main learning to anticipate.

Observation

The modeling process was focused on developing a relatively simple model that could illustrate an idea of the overall behavior of the problem-system without including

Observation

199

Appendix F

too many details, partly because overview was considered more important than detailed correctness (avoiding black-box effect), partly due to the fact that the system dynamics Vensim model was complemented by a more detailed excel-model with the format needed in the budgeting and business planning. The result was a model that was relatively easy to explain in even 1-2 hour meetings. The observer (and facilitator) primarily had a theoretical foundation for SD modeling, with only little SD modeling experience, but has more than 10 years of planned organizational intervention experience, including other types of modeling.

2nd order observation

The observer (and facilitator) had personal relationships with company executives, which could be expected to influence both ‘positively’ with regards to access to information and dialogues with the decision-makers, and ‘negatively’ with regards to creating biases. Although being a highly subjective observation, the observer had the impression that most of the participants were rather indifferent to the existence of personal relationships, which might be due to the fact that most of the participants were high-placed managers themselves with a high degree of self-confidence.

2nd order observation

Bibliography Ackoff, Russel L. : The Art of Problem Solving – Accompanied by Ackoff’s Fables, New York, (John Wiley & Sons) 1978. Ajzen, Icek: Attitudes, Personality and Behavior, Chicago, (The Dorsey Press) 1988. Ajzen, Icek and Martin Fishbein: “The Prediction of Behavior from Attitudinal and Normative Variables”, Journal of Experimental Social Psychology, No. 6, 1970, pp. 466—487. Akkermans, Henk: Modelling With Managers, Breda, The Netherlands, 1995. Andersen, David F., George P. Richardson and Jac A. M. Vennix: “Group model building: adding more science to the craft”, System Dynamics Review, Vol. 13, No. 2, 1997, pp. 187—201. Andersen, David F and George P. Richardson: “Scripts for group model building”, System Dynamics Review, Vol. 13, No. 2, 1997, pp. 107—129. Anderson, Linda A. and Dean Anderson : “Awake at the Wheel: Moving beyond Change Management to Conscious Change Leadership”, OD Practitioner, Vol. 33, No. 3, 2001, pp. 40—48. Argyris, Chris: Interventions Theory and Method – A Behavioural Science View, Reading, Massachusetts, (Addison-Wesley Publishing Company) 1970. Argyris, Chris: Reasoning, Learning, and Action – Individual and Organizational, San Francisco, (Jossey-Bass Publishers) 1982. Argyris, Chris: On Organisational Learning, 2 nd edition, Oxford, (Blackwell Publishers) 1999.

202

Bibliography

Argyris, Chris and Donald Schön : “ Organizational Learning: A theory of Action Perspective” (first printed as part of book with the same title in 1978) , reprint in Gareth Morgan: Creative Organization Theory, Newbury Park, California, (SAGE Publications) 1989, pp. 139—142. Ashby, W. Ross: An Introduction to Cybernetics, paperback version, London, (Methuen & Co. Ltd.) 1964. Bakken, Bent E.: Learning and Transfer of Understanding in Dynamics Decision Environments, Boston, (MIT dissertation) 1993. Barlas, Yaman : “Formal aspects of model validity and validation in system dynamics”, System Dynamics Review, Vol. 12, No. 3, 1996, pp. 183—210. Baron, Jonathan : Thinking and Deciding, 3 rd edition, Cambridge, (Cambridge University Press) 2000. Beer, Stafford : Beyond Dispute – The Invention of Team Syntegrity, Chichester, (John Wiley & Sons) 1994. Bennis, Warren G., Kenneth D. Benne and Robert Chin : The Planning of Change, 4 th edition, New York, (CBS College Publishing) 1985. Berger, Ulrike and Isolde Bernhard-Mehlich : “Die Verhaltenswissenschaftliche Entscheidungstheorie”, in Alfred Kieser (ed.): Organisationstheorien, 3 rd edition, Stuttgart, (Kohlhammer) 1999, pp. 133—168. Bonabeau, Eric: “Don’t Trust Your Guts”, Harvard Business Review, May 2003, pp. 116—123. Borum, Finn : Strategier for organisationsændringer, Copenhagen, (Handelshøjskolens Forlag) 1995. Brown, John Seely: Research That Reinvents the Corporation, Harvard Business Review, August 2002, pp. 105—114. Briggs, Robert, Gert-Jan de Vreede, and Jay F. Nunamaker Jr .: “Collaboration Engineering with ThinkLets to Pursue Sustained Success with Group Systems”, Journal of Management Information Systems, Vol. 19, No. 4, Spring 2003, pp. 31—63.

Bibliography

203

Bungard, Walter and Catrin Niethammer: “Psychologische Aspekte des Change Management im interorganisationalen Kontext”, in Walter Bungard, Jürgen Fleischer, Holger Nohr, Dieter Spath and Erich Zahn (eds.), Customer Knowledge Management, Stuttgart, 2003, pp. 105—117. Bungard, Walter: “Zur Implementierungsproblematik bei Business-Reengineering Projekten”, in Manfred Perlitz, Andreas Offinger, Michael Reinhardt and Klaus Schug (eds.): Reengineering zwischen Anspruch und Wirklichkeit, Wiesbaden, (Gabler Verlag) 1996, pp. 253—273. Checkland, Peter: “Soft System Methodology”, in Jonathan Rosenhead (ed.): Rational Analysis for a Problematic World, Chichester, (John Wiley & Sons) 1989, pp. 21—42. Checkland, Peter: “Systems Thinking, Systems Practice”, Chichester, (John Wiley & Sons) 1993. Chin, Robert and Kenneth D. Benne: “General Strategies for Effecting Changes in Human Systems”, in Warren G. Bennis, Kenneth D. Benne and Robert Chin: The Planning of Change, 4 th edition, New York, (CBS College Publishing) 1985, pp. 22—45. Christensen, Søren and Jan Molin : Organisationskulturer, Copenhagen, (Akademisk Forlag) 1987. Cummings, Thomas G. and Christopher G. Worley: Organizational Development and Change, Ohio, (South-Western College Publishing) 2001. Conant, Roger C. and W. Ross Ashby: “Every Good Regulator of a System Must Be a Model of that System”, International Journal of System Sciences, Vol. 1, No. 2, 1970, pp. 89—97. Davenport, Thomas H.: Process Innovation, Boston, (Harvard Business School Press) 1993. Denscombe, Martyn : The Good Research Guide for small-scale social research projects, 2 nd edition, Philadelphia, (Open University Press) 2003. De Bono, Edward : Lateral Thinking for Management, England, (American Management Association) 1971.

204

Bibliography

de Geus, Arie P.: “Planning as Learning”, Harvard Business Review, Vol. 66, No. 2, March-April 1988, pp. 70—74. de Geus, Arie P.: The Living Company, Boston, (Longview Publishing Limited) 1997. Distefano, Joseph J. and Martha L. Maznevski : “Creating Value with Diverse Teams in Global Management”, Organizational Dynamics, Vol. 29, No. 1, pp. 45—63. Doppler, Klaus and Christoph Lauterburg : Change Management – Den Unternehmenswandel gestalten, 10 th edition, Frankfurt/Main, (Campus Verlag) 2002. Doyle, James K. and David N. Ford : “Mental models concepts for system dynamics research”, System Dynamics Review, Vol. 14, No. 1, Spring 1998, pp. 3—29. Doyle, James K. and David N. Ford : “Mental models concepts revisited: some clarifications and a reply to Lane”, System Dynamics Review, Vol. 15, No. 4, Winter 1999, pp. 411—415. Dörner, Dietrich : The Logic of Failure, New York, (Metropolitan Books) 1996. Eco, Umberto : A Theory of Semiotics, Bloomington, (Indiana University Press) 1976. Eden, Colin : “Using cognitive mapping for strategic options development and analysis”, in Jonathan Rosenhead (ed.): Rational Analysis for a Problematic World, Chichester, (John Wiley & Sons) 1989, pp. 21—42. Eden, Colin : “Cognitive mapping and problem structuring for system dynamics model building”, System Dynamics Review, Vol. 10, Nos. 2-3, Summer-Fall 1994, pp. 257—276. Eisenhardt, Kathleen M.: “Building Theories from Case Study Research”, Academy of Management Review, Vol. 14, No. 4, 1989, pp. 532—550. Eisenhardt, Kathleen M.: “Strategy as Strategic Decision Making”, Sloan Management Review, Spring 1999, pp. 65—72.

Bibliography

205

Eskinasi, Martijn and Eppie Fokkema: “Bursting the myth of making easy modeling money”, unpublished working paper (permission to quote granted from the authors), Proceedings, Second European System Dynamics Workshop, Nijmegen, 2005, pp. 62—66. Farson, Richard and Ralph Keyes: “The Failure-Tolerant Leader”, Harvard Business Review, August 2002, pp. 64—71. Ferres, Natalie, Julia Connell, and Anthony Travaglione: “The effect of future redeployment on organizational trust”, Strategic Change, Vol. 14, March-April 2005, pp. 77—91. Fey, Willard and John Trimble: “The Evaluation and Development of Knowledge Acquisition in System Dynamics Studies”, in Proceedings, System Dynamics Conference, System Dynamics Society, 1992, pp. 173—182. Fine, Charles H. : “Clockspeed-based strategies for Supply Chain Design”, Production and Operation Management, Vol. 9, No. 3, 2000, pp. 213—221. Flood, Robert L. and Michael C. Jackson : Creative Problem Solving: Total Systems Intervention, Chichester, (John Wiley & Sons) 1991. Flood, Robert L.: Solving Problem Solving, Chichester, (John Wiley & Sons) 1995. Forrester, Jay W. and Peter Senge: “Tests for Building Confidence in System Dynamics Models”, in Augusto A. Legasto, Jay W. Forrester und James M. Lyneis (eds.), TIMS Studies in the Management Sciences, Vol. 14, Amsterdam, 1980, pp. 209—228. Forrester, Jay W.: Principles of Systems, Cambridge, (MIT Press) 1968. Forrester, Jay W.: Industrial Dynamics, Cambridge, (MIT Press) 1961. Forrester, Jay W.: “Counterintuitive Behavior of Social Systems”, in Collected Papers of Jay W. Forrester – Foreword by Gordon S. Brown, Cambridge, (Wright-Allen Press, inc.) 1975, pp. 211—244. Forrester, Jay W.: “Policies, decisions and information sources for modeling”, European Journal of Operational Research, Vol. 59, No. 1, 1992, pp. 42—63.

206

Bibliography

Forrester, Jay W.: “System dynamics, system thinking, and soft OR”, System Dynamics Review, Vol. 10, No. 2, 1994, pp. 245—256. Gladwell, Malcolm: The Tipping Point, paperback edition, New York, (Back Bay Books) 2002. Gold, Martin : The Complete Social Scientist: A Kurt Lewin Reader, Washington, (American Psychological Association) 1999. Größler, Andreas: Entwicklungsprozess und Evaluation von Unternehmenssimulation für lernende Unternehmen, Frankfurt am Main, (Peter Lang) 2000. Größler, Andreas, Peter Milling and Graham Winch : “Perspectives on rationality in system dynamics: a workshop report and open research questions”, System Dynamics Review, Vol. 20, No. 1, 2004, pp. 75—87. Größler, Andreas: “Don’t Let History Repeat Itself: Methodological Issues Concerning the Use of Simulators in Teaching and Experimentation”, System Dynamics Review, Volume 20, Number 3, Fall 2004, pp. 263—274. Größler, Andreas: “A Content and Process View on Bounded Rationality in System Dynamics”, Systems Research and Behavioral Science, Vol. 21, No. 4, July/August, 2004, pp. 319—330. Hamel, Gary: Leading the Revolution, Boston, (Harvard Business School Press) 2000 Henttonen, Kaisa and Kirsimarja Blomqvist: “Managing distance in a global virtual team: the evolution of trust through technology-mediated relational communication”, Strategic Change, Vol. 14, March-April 2005, pp. 107—119. Hodgson, A. M.: “Hexagons for system thinking”, European Journal of Operational Research, Vol. 59, No. 1, 1992, pp. 123—136. Homer, Jack B. : “Why we iterate: scientific modeling in theory and practice”, System Dynamics Review, Vol. 12, No. 1, Spring 1996, pp. 1—19. Hogarth, Robin : Judgment and Choice – The Psychology of Decision, 2 nd edition, Chicago, (John Wiley & Sons) 1987.

Bibliography

207

Huff, Anne S. and Rhonda Kay Reger: “A Review of Strategic Process Research”, Journal of Management, Vol. 13, No. 2, 1987, pp. 211—236. Huz, Steven, David F. Andersen, George P. Richardson and Roger Boothroyd : “A framework for evaluating systems thinking interventions: an experimental approach to mental health system change”, System Dynamics Review, Vol. 13, No. 2, 1997, pp. 149—169. Imai, Masaaki : Kaizen: Der Schlüssel zum Erfolg der Japaner im Wettbewerb, german translation, München, (Wirtschaftsverlag Langen Müller Herbig) 1992. Jacobsen, Chanoch and Yitzhak Samuel: “ Planned Organizational Change: Theory, Model, Data and Simulation”, in Peter M. Milling and Erich O.K. Zahn (eds.): Computer-Based Management of Complex Systems, Proceedings of the 1989 International Conference of the System Dynamics Society, Heidelberg, (Springer-Verlag) 1989, pp. 104—118. Janis, Irving L. : “Groupthink: The Problems of Conformity” (original printed in Psychology Today, Nov. 1971, pp. 271—279), reprint in Gareth Morgan: Creative Organization Theory, Newbury Park, California, (SAGE Publications) 1989, pp. 224—228. Johnson-Laird, P. N.: Mental Models, Cambridge, (Cambridge University Press) 1983. Jöns, Ingela : Managementstrategien und Organisationswandel, Mannheim, (Mannheim University Dissertation) 1995. Kahnemann, Daniel and Amos Tversky: “On the Psychology of Prediction”, Psychological Review, Vol. 80, No. 4, July 1973, pp. 237—251. Kampmann, Christian P. E. : Feedback complexity and market adjustment – An experimental approach, Boston, (MIT Dissertation) 1992. Kanter, Rosabeth Moss: “Leadership and the Psychology of Turnarounds”, Harvard Business Review, June 2003, pp. 58—67. Keough, Mark and Andrew Doman : “The CEO as organization designer: An interview with Professor Jay W. Forrester, the founder of system dynamics”, The McKinsey Quarterly, No. 2, 1992, pp. 3—30.

208

Bibliography

Kieser, Alfred : “Human Relations-Bewegung und Organisationspsycholgie”, in Alfred Kieser (ed.): Organisationstheorien, 3 rd edition, Stuttgart, (Kohlhammer) 1999, pp. 101—131. Kieser, Alfred : “Kontruktivistische Ansätze”, in Organisationstheorien, 3 rd edition, Stuttgart, pp. 287—318.

Alfred Kieser (Kohlhammer)

(ed.): 1999,

Kieser, Alfred and Ulrich Koch : Organizational Learning through Rule Adaptation – From the Behavioral Theory to Transactive Organizational Learning, Mannheim, 2000. Kim, Daniel H. and Peter M. Senge: “Putting systems thinking into practice”, System Dynamics Review, Vol. 10, Nos. 2-4, Summer-Fall 1997, pp. 277—290. Klein, Donald : “Some Notes on the Dynamics of Resistance to Change: The Defender Role”, in Warren G. Bennis, Kenneth D. Benne and Robert Chin (eds.): The Planning of Change, 4 th edition, New York, (CBS College Publishing) 1985, pp. 98—105. Koningswieset, Roswita and Alexander Exner: Systemische Intervention, Stuttgart, (Klett-Cotta) 1998. Kotter, John P : Leading Change, Boston, (Harvard Business School Press) 1996. Kotter, John P. and Dan S. Cohen : The Heart of Change, Boston, (Harvard Business School Press) 2002. Kotler, Philip : Marketing Management, 7 th edition, New Jersey, (Prentice-Hall, Inc.) 1991. Kvale, Steinar: InterViews – An Introduction to Qualitative Research Interviewing, Thousand Oaks, (SAGE Publications) 1996. Lane, David C. : “Modelling as Learning: A consultancy methodology for enhancing learning in management teams”, European Journal of Operational Research, Vol. 59, No. 1, 1992, pp. 64—84. Lane, David C. : “Friendly amendment: A commentary on Doyle and Ford’s proposed re-definition of ‘mental model’”, Systems Dynamics Review, Vol. 15, No. 2, Summer 1999, pp. 185—194.

Bibliography

209

Lane, David C. : “Should System Dynamics be Described as a ‘Hard’ or ‘Deterministic’ System Approach?”, Systems Research and Behavioral Science, Vol. 17, 2000, pp. 3—22. Leavitt, Harold J.: Top Down – Why Hierarchies Are Here to Stay and How to Manage Them More Effectively, Boston, (Harvard Business School Press) 2005. Leavitt, Harold J., William R. Dill, and Henry B Eyring: The Organizational World – A systematic view of managers and management, New York, (Harcourt Brace Jovanovich, Inc.) 1973. Leonard-Barton, Dorothy: “A Dual Methodology for Case Studies: Synergistic use of a Longitudinal Single Site with Replicated Multiple Sites ”, Organizational Science, Vol. 1, No. 3, 1990, pp. 248—266. Leonard, Dorothy: Wellsprings of Knowledge – Building and Sustaining the Sources of Innovation, Boston, (Harvard Business School Press) 1998 Lewin, Kurt: “Group Decision and Social Change” (first published in Newcomb and Hartley’s Readings in social psychology, 1948, pp. 330—341), in Martin Gold: The Complete Social Scientist – A Kurt Lewin Reader, Washington, (American Psychological Association) 1999, pp. 265—291. Lochmann, Hans-Dieter und Michaela Rüsch-Kornasoff : “Organization Change Strategy: Ein wesentlicher Baustein des Reengineerings”, in Manfred Perlitz, Andreas Offinger, Michael Reinhardt and Klaus Schug (eds.): Reengineering zwischen Anspruch und Wirklichkeit, Wiesbaden, (Gabler Verlag) 1996, pp. 329—340. Luna-Reyes, Luis Felipe and Deborah Lines Andersen : “Collecting and analysing qualitative data for system dynamics: methods and models”, System Dynamics Review, Vol. 19, No. 4, 2003, pp. 271—296. Lübcke, Poul (ed.): Politikens filosofi leksikon, Copenhagen, (Politikens forlag) 2001. Lyneis, James M.: Corporate Planning and Policy Design – A System Dynamics Approach, Boston, (Pugh-Roberts Associates, Inc.) 1980.

210

Bibliography

Maani, Kambiz E. and Robert Y. Cavana : Systems Thinking and Modeling – Understanding Change and Complexity; Auckland, (Person Education New Zealand, Lt.) 2000. Maier, Frank und Andreas Größler : ”What are we talking about? A taxonomy of Computer Simulations to Support Learning”, System Dynamics Review, Vol. 16, No. 2, 2000, pp. 135—148. Maier, Frank: Die Integration wissens- und modellbasierter Konzepte zur Entscheidungsunterstützung im Innovationsmanagement, Berlin, (Duncker & Humblot), 1995. Martinez, Ignacio J and George P. Richardson: “An Expert View on the System Dynamics Modeling Process: Concurrences and Divergences Searching for Best Practices in System Dynamics Modeling”, at CD-ROM of Proceedings, System Dynamics Conference, System Dynamics Society, 2002. Martinez, Ignacio J. and Luis F Luna : “The Dynamics of Best Practices: A Structural Approach”, at CD-ROM of Proceedings, System Dynamics Conference, System Dynamics Society, 2001. McKee, Robert: “Storytelling That Moves People”, Harvard Business Review, June 2003, pp. 51—55. McPherson III, L. Fillmore: “Organizational Change: An Industrial Dynamics Approach”, in Edward B. Roberts (ed.): Managerial Applications of System Dynamics, Cambridge, Massachusetts, (Productivity Press) 1978, pp. 447—459. Miller, George A.: “The Magical Number Seven, Plus Minus Two: Some Limits on Our Capacity for Processing Information”, The Psychological Review, Vol. 63, No. 2, March 1956, pp. 81—97. Milling, Peter: Systemtheoretische Grundlagen zur Plannung der Unternehmenspolitik, Berlin, (Verlag Duncker & Humblot) 1981. Milling, Peter: “Leitmotive des System-Dynamics-Ansatzes”, Wirtshaftswissenschaftliches Studium, Vol. 10, 1984, pp. 507—513.

Bibliography

211

Milling, Peter M. and Erich O.K. Zahn (edt.): Computer-Based Management of Complex Systems, Proceedings of the 1989 International Conference of the System Dynamics Society, Heidelberg, (Springer-Verlag) 1989. Milling, Peter: “Organisationales Lernen und seine Unterstützung durch Managementsimulatoren”, in: Zeitschrift für Betriebswirtschaft, Ergänzungsheft 3/95, Lernende Unternehmen (Sonderausgabe), 1995, pp. 93—112. Milling, Peter: “Modeling Innovation Processes for Decision Support and Management Simulation”, System Dynamics Review, Vol. 12, No. 3, 1996, pp. 211—234. Milling, Peter: “Kybernetische Überlegungen beim Entscheiden in komplexen Systemen”, in Peter Milling (ed.): Entscheiden in komplexen Systemen, Wirtschaftskybernetik und Systemanalyse, Band 20, Berlin, (Duncker & Humblot) 2002. Mintzberg, Henry: The Rise and Fall of Strategic Planning, New York, (The Free Press) 1994. Morecroft, J. D. W. : “Executive knowledge, models and learning”, European Journal of Operational Research, Vol. 59, Vol. 1, 1992, pp. 9—27. Morecroft, John D. W. and John D Sterman (ed.): Modeling for Learning Organizations, Portland, Oregon, (Productivity Press) 1994. Morgan, Gareth : Creative Organization Theory, Newbury Park, California, (SAGE Publications) 1989. Munk, Kirstine: “In the Airy Spaces of Our Minds…: Cosmology and ritual design in modern, Western astrology”, in Michael York (ed.): Nature, Religion, and Culture, in print, London, (Equinox) 2006. Newell, Allen and Herbert A. Simon : Human Problem Solving, Englewood Cliffs, New Jersey, (Prentice-Hall, Inc.) 1972. O’Connor, Joseph and Ian McDermott : The Art of System Thinking – Essential skills for creativity and problem solving, London, (Thorsons) 1997.

212

Bibliography

Palmer, Ian and Richard Dunford : “Who says change can be managed? Positions, perspectives and problematics”, Strategic Change, Volume 11, August 2002, pp. 243—251. Pearson, Alan : “You Drive for the Show but you Putt for the Dough”, appendix in Stafford Beer: Beyond Dispute: The Invention of Team Syntegrity, Chichester, (John Wiley & Sons) 1994, pp. 311—322. Perlitz, Manfred, Jürgen Bufka, Andreas Offinger, Michael Reinhardt, und Klaus Schug : “Reengineering-Projekte erfolgreich umsetzen: Ergebnisse einer Erfolgsfaktorenstudie”, in Manfred Perlitz, Andreas Offinger, Michael Reinhardt and Klaus Schug (eds.), Reengineering zwischen Anspruch und Wirklichkeit, Wiesbaden, (Gabler Verlag) 1996, pp. 181—208. Perlitz, Manfred : Internationales Management, 5 th edition, Stuttgart, (Lucius & Lucius) 2004. Randers, Jørgen : “Guidelines for Model Conceptualization”, in Jørgen Randers (ed.): Elements of the System Dynamics Method, Cambridge, Connecticut, (Productivity Press) 1980, pp. 117—139. Rao, Abhijit: “Recognition of Conative and Affective Behavior in Web Learning using Digital Gestures”, North America Web-Based Learning Conference, Online Proceedings, New Brunswick, 2001. Repenning, Nelson P. and John D. Sterman : Nobody Ever Gets Credit for Fixing Problems that Never Happened: Creating and Sustaining Process Improvement”, California Management Review, Vol.43, No.4, Summer 2001, pp. 64—88. Richardson, George P. and Alexander L. Pugh : Introduction to System Dynamics Modeling with DYNAMO, Cambridge, (MIT Press) 1981. Richardson, George P. and David F. Andersen: “Teamwork in group model building”, System Dynamics Review, Vol. 11, No. 2, 1995, pp. 113—137. Richmond, Barry: “The Strategic Forum: aligning objectives, strategy and process”, System Dynamics Review, Vol. 13, No. 2, 1997, pp. 131—148.

213

Bibliography

Roberto, Michael and Lynne Levesque: “The Art of Making Change Stick”, MIT Sloan Management Review, Summer 2005, Vol. 46, No. 4, Summer 2005, pp. 53—60. Roberts, Edward B .: “Strategies for Effective Implementation of Complex Corporate Models”, in Edward B. Roberts (ed.): Managerial Applications of System Dynamics, Cambridge, Massachusetts, (Productivity Press) 1978, pp. 77—85. Rockart, John F .: “Chief executives define their own data needs”, Harvard Business Review, March-April 1979, pp. 81—93. Rosenhead, Jonathan (ed.): Rational Analysis for a Problematic World, Chichester, (John Wiley & Sons) 1989. Rouwette, Etiënne: Group model building as mutual persuasion, Nijmegen, (Wolf Legal Publishers) 2003. Salge, Markus and Peter Milling : “Who is to blame, the operator or the designer? Two stages of human failure in the Chernobyl accident”, System Dynamics Review, Vol. 22, in print, 2006 Schein, Edgar H. : Organisationspsykologi, Danish translation, Herning, (Forlaget systime) 1990. Schein, Edgar H.: Process Consultation – MAOM Capstone Course for the University of Phoenix, Boston, (Person Custom Publishing) 2000. Schneiderman, Arthur M.: “Setting Quality Goals: Use observed rates of continuous improvement to position targets”, Quality Progress, April 1988, pp. 51—57. Scheer, August-Wilhem: Business Process Reengineering – Reference Models for Industrial Enterprises, 2 nd edition, Berlin, (Springer-Verlag) 1994. Scheper, Willem J.: (Dissertation) 1991.

Group

Decision

Support

Systems,

Tilburg/Utrecht,

Scherer, Andreas G.: “Kritik der Organisation oder Organisation der Kritik? Wissenschaftstheoretische Bemerkungen zum Umgang mit Organisationstheorien”, in Alfred Kieser (ed.): Organisationstheorien, 3 rd edition, Stuttgart, (Kohlhammer) 1999, pp. 1—37.

214

Bibliography

Senge, Peter M. and John D. Sterman : “Systems thinking and organizational learning: Acting locally and thinking globally in the organization of the future”, European Journal of Operational Research, Vol. 59, No. 1, 1992, pp. 137—150. Senge, Peter M.: The Fifth Discipline, paperback edition, New York, (Currency Doubleday) 1994. Senge, Peter M. : The Dance of Change, New York, (Currency Doubleday) 1999. Simon, Herbert A.: “Rational Decision Making in Business Organizations”, The American Economic Review, Vol. 69, No. 4, 1979, pp. 493—513. Simon, Herbert A.: The Science of the Artificial, 3 rd edition, Cambridge, (The MIT Press) 1996. Snabe, Birgitte and Jakob F. Ølgaard : Informationsstrategi, Master Thesis, Institute of Mathematical Statistics and Operations Research, The Technical University of Denmark, Lyngby, 1992. Snabe, Birgitte: Brugtvognskoncept for Skandinavisk Motor Co., Thesis, Det erhvervsøkonomiske diplomstudium, The Copenhagen Business School, Copenhagen, 1994. Snabe, Birgitte and Andreas Größler : “System Dynamics Modelling for Strategy Implementation: Case Study and Issues”, Journal of Systems Research and Behavioral Science, Vol. 23, No. 4, 2006, pp. 467—481. Snabe, Birgitte, Andreas Größler, and Peter M. Milling : “Policies and Politics of Establishing R&D Capacity in Low-Cost Locations”, Tagungsband, GWSTagung 2005, Greifswald, in print, 2006. Stake, Robert E.: The Art of Case Study Research, Thousand Oaks, (SAGE Publications) 1995. Sterman, John D.: “ Misperceptions of Feedback in Dynamic Decision Making”, in Peter M. Milling and Erich O.K. Zahn (eds.): Computer-Based Management of Complex Systems, Proceedings of the 1989 International Conference of the System Dynamics Society, Heidelberg, (Springer-Verlag) 1989, pp. 21—31.

Bibliography

215

Sterman, John D.: Business Dynamics – Systems Thinking and Modeling for a Complex World, Boston, (McGraw Hill Higher Education) 2000. Sterman, John D.: “All models are wrong: reflections on becoming a systems scientist”, System Dynamics Review, Vol. 18, No. 4, Winter 2002, pp. 501—531. Taylor, Frederick W.: The Principles of Scientific Management, New York, first time published in 1911, (Norton & Company) 1967. Takahara, Yasuhiko : “A Formal Model of Organization”, in Singo Takahashi, Kyoichi Kijima and Ryo Sato (eds.): Applied General Systems Research on Organizations, Tokyo, Springer, 2004, pp. 3—35. Thun, Jörn-Henrik: Die zeitbasierte Fertigungsstrategie, Wiesbaden, (Mannheim University Dissertation) 2002. Thun, Jörn-Henrik, Peter M. Milling, and Uwe Schwellbach : ”The Impact of Total Employee Involvement on Time-based Manufacturing“, in Kate Blackmon, Steve Brown, Paul Cousins, Andrew Graves, Christine Harland, Richard Lamming, and Harvey Maylor (eds.): ”What Really Matters in Operations Management“, Bath, 2001, pp. 131—143. Tilebein, Meike: “Eine strukturwissenschaftliche Betrachtung von Diversity Management”, Tagungsband, GWS-Tagung 2005, Greifswald, in print, 2006. Tversky, Amos: “Additivity, utility and subjective probability”, in Ward Edwards and Amos Tversky (eds.): Decision Making, Middlesex, (Penguin Modern Psychology) 1967, pp. 208—238. Thygesen, Henriette H.: System Dynamics in Action, Copenhagen, (Samfundslitteratur) 2004. Van der Smagt, Ton : “Causation and Constitution in System Dynamics; Modelling a Socially Constituted World”, Journal of Systems Research and Behavioral Science, Vol. 23, No. 4, in print, 2006.

216

Bibliography

Vennix, Jac A. M, David F. Andersen, George P. Richardson and John Rohrbaugh : “Model-building for group decision support: Issues and alternatives in knowledge elicitation”, European Journal of Operational Research, Vol. 59, No. 1, 1992, pp. 28—41. Vennix, Jac A. M. : Group Model Building, Chichester, (John Wiley & Sons) 1996. Voss, Chris, Nikos Tsikriktsis, and Mark Frohlich : “Case research in operations management”, International Journal of Operations & Production Management, Vol. 22, No. 2, 2002, pp. 195—219. Warren, Kim: Competitive Strategy Dynamics, Chichester, (John Wiley & Sons) 2002. Warren, Kim: “Improving strategic management with the fundamental principles of system dynamics”, System Dynamics Review, Vol. 21, No. 4, Winter 2005, pp. 329—350. Webster’s Encyclopedic Unabridged Dictionary of the English Language, New York, (Portland House) 1989. Weil, Henry B.: “The Evolution of an Approach for Achieving Implemented Results from System Dynamics Projects”, in Jørgen Randers (ed.): Elements of the System Dynamics Method, Cambridge, Connecticut, (Productivity Press) 1980, pp. 269—291. Weil, Henry B. and Kenneth P. Veit: “Corporate Strategic Thinking: The role of System Dynamics” in Peter M. Milling and Erich O.K. Zahn (Eds.): ComputerBased Management of Complex Systems, Proceedings of the 1989 International Conference on the System Dynamics Society, Stuttgart, (Springer-Verlag) 1989, pp. 66—75. Wenneberg, Søren B .: “Socialkonstruktivisme som videnskabsteori: Sifyfos’ videnskab”, Online Paper, Institut for Ledelse, Politik og Filosofi, Copenhagen, 2002. Winch, Graham and Sonja Derrick: “Flexible Study Processes in ‘Knotty’ System Dynamics Projects”, Journal of Systems Research and Behavioral Science, Vol. 23, No. 4, in print, 2006.

Bibliography

217

Wolstenholme, Eric F .: “The definition and application of a stepwise approach to model conceptualisation and analysis”, European Journal of Operational Research, Vol. 59, 1992, pp. 123—136. Wolstenholme, Eric: “Using generic system archetypes to support thinking and modelling”, System Dynamics Review, Vol. 20, No. 4, Winter 2004, pp. 341—356. Yin, Robert K.: Case study Research, 3 rd edition, Thousand Oaks, (SAGE Publications) 2003. Zock, Alexander: “A critical review of the use of System Dynamics for organizational consulting projects”, at CD-ROM of Proceedings, System Dynamics Conference, System Dynamics Society, Oxford, 2004.