Preface and Acknowledgments¹

^[]

^[] Full details of articles and books referred to in the Preface can be found in later chapters by cross-referencing with author names in the index.

I first became interested in models and modelling when I studied physics as an undergraduate at Bristol University. Or perhaps my interest was really awakened much earlier when, as a boy of nine or 10, my friend Alan Green introduced me to the board game of Monopoly. I soon became fascinated with board games of all sorts and accumulated a collection that included Cluedo (a detective murder mystery game, also known as Clue), Railroader (a game to build and operate your own wild-west railroad in competition with rival railway companies), and Buccaneer (a game of pirate ships and treasure collecting). I was intrigued by the colourful tokens, the chance cards, the rules and the evocative boards that showed city sights, a murder mansion, a treasure island or whatever was needed to fire the imagination. In Buccaneer, the game's clever distinction between the 'sailing power' and 'fighting power' of a treasure-seeking frigate is something I still appreciate today. And as a modeller I admire the game designer's artful representation of a pirate's world, set out on a blue-and-white chequered board that serves as an ocean.

Later, after graduating from Bristol, I joined Ford of Europe's operational research department, where computational decision models replaced the abstract and elegant models of physics. There I worked on investment appraisal (justifying the decision to build a new Fiesta car factory in Spain) and dealer location (whereabouts within Bromsgrove, Bury St Edmunds, or other English towns and cities, to site new car dealerships). During the second of my three years with Ford, the company sponsored me on an MSc degree in operational research at London University's Imperial College. It was at Imperial that I first encountered system dynamics, albeit briefly in an elective course on quantative methods, and this chance encounter eventually led me to apply to the doctoral programme at MIT's Sloan School of Management for a PhD in system dynamics. Hence began the journey that I have pursued ever since.

When I look back over my 30-plus years in the field I see five different phases of work, all of which have contributed to the content of this book and led me to the friends and colleagues who have shaped my thinking. My names for these phases are: (1) manufacturing dynamics and information networks; (2) bounded rationality and behavioural decision making; (3) modelling for learning; (4) the dynamics of strategy; and (5) soft systems and complementary modelling methods.

Manufacturing Dynamics and Information Networks

The first phase coincided with my doctoral dissertation at MIT when I worked on manufacturing and supply chain dynamics in Cummins Engine Company and Harley-Davidson. I was fortunate, back then, to have Jay Forrester as my PhD thesis supervisor, Jim Lyneis as a collaborator/faculty adviser on the Cummins project, and Nathaniel Mass as a faculty instructor. I learned many valuable modelling skills from them and from MIT's intensive academic apprenticeship with its special educational blend of theory and real-world practice. I still remember the sense of excitement as a first-year doctoral student, arriving by plane in Columbus Indiana, headquarters of Cummins Engine Company. There, I worked on the Cummins manufacturing dynamics project and found myself applying the inventory control, forecasting and production planning formulations I had learned at MIT. The simple factory model in Chapter 5 contains echoes of these same formulations.

My doctoral thesis topic arose from an on-the-job discovery that circumstance presented. I was working simultaneously on manufacturing models of Cummins and Harley-Davidson. When I set out the 10–15 page diagrams of these two models side-by-side on my apartment floor in Cambridge (Massachusetts), I noticed that the information flows which coordinated multi-stage production in the two factories were arranged in different patterns. Every stage of production in Harley from final assembly of motorcycles to sub-assemblies and raw materials, was coordinated from a master schedule - a kind of top-down control. There was no such master schedule in Cummins's factory at the time. Stages of production followed local order-point rules. It turned out that Harley-Davidson was operating a computer-driven top-down material requirements planning (MRP) system, which was entirely new to manufacturing firms at the time (and, back then, had scarcely featured in the academic literature on operations management). My thesis compared the long-term dynamic performance of these alternative approaches to production planning and control. A striking result was that traditional order-point rules outperformed MRP (in terms of operating cost, production stability, inventory availability and lead-time predictability). Only under special and hard-to-achieve factory conditions was MRP superior, despite the cost-savings touted by advocates of MRP. And so my curiosity in information networks began.

As an aside, I should mention that the basis for the manufacturing models in my thesis was the production sector of the MIT group's National Economic Model. The production sector was essentially a generic model of the firm, residing within a system dynamics model of the US economy. The premise of the group's research at the time was that the US economy could be conceived as a micro-economic collection of interacting firms and households. Macro-economic behaviour arises from micro-structure. Jay Forrester was leading the National Model project, so he knew the production sector intimately. As my thesis supervisor he was able to swiftly critique and guide my efforts to adapt this generic model of the firm to fit what I had discovered from the company-specific models of Cummins and Harley. I learned a great deal about model formulation and behaviour analysis from those encounters. I also learned from other doctoral students in system dynamics who, at the time, included David Andersen, Alan Graham, Matts Lindquist, Ali Mashayeki, George Richardson, Barry Richmond, Khalid Saeed and Peter Senge; and then later Nathan Forrester, John Sterman, Jack Homer, Jim Hines and Bob Eberlein.

It was while working with the production sector, which was a visually complex model, that I took to drawing boundaries around sets of model symbols that belonged with a given policy function, such as capacity utilisation or scheduling and ordering. This visual simplification procedure later led to policy structure diagrams as a high-level way of representing the coordinating network in system dynamics models. I use both policy boundaries and policy structure diagrams throughout the book.

Bounded Rationality and Behavioural Decision Making

My thesis showed that sparse and 'simple' information networks in firms can often deliver business performance that is superior to more complex and sophisticated information networks. This observation led me, as a newly-appointed junior faculty member at MIT Sloan, into the literature of the Carnegie School and Herbert Simon's work on bounded rationality. The idea that the 'structure' of a firm's information feedback network determines the firm's performance and dynamic behaviour is central to system dynamics. The Carnegie literature helps to bring the information network into clear focus and to explain why human decision makers, faced with complexity and information overload, prefer sparse information networks. People and organisations are boundedly rational. They cannot gather or process all the information needed to make 'best' (objectively rational) decisions. Whenever people take decisions that lead to action, they selectively filter information sources, disregarding or overlooking many signals while paying attention to only a few. Well-designed policies recognise this human trait, while functional 'stovepipes' are an unfortunate corollary that stem from poor design (or no design at all). In practice, bounded rationality leads to departmentalised organisations in which the left hand quite literally doesn't know (and shouldn't need to know) what the right hand is doing. Loose coordination among functions, departments or sectors is normal.

Bounded rationality helped me to identify, interpret and better understand information feedback loops in business and social systems. Puzzling dynamics nearly always arise from 'hidden' coordination problems and this idea is woven throughout the book, beginning with the simple fisheries model in Chapter 1, continuing in Chapter 4's world of showers and in Chapter 5's factory model, and culminating in Chapter 7's market growth model. The information/coordination theme continues in Chapter 8 (the oil industry), in part of Chapter 9 (a return to fisheries) and in Chapter 10 (product growth dynamics in fast-moving consumer goods).

I was not alone at MIT in working on bounded rationality and system dynamics. John Sterman too was studying the topic, and using it to make sense of long-term economic cycles generated by the National Economic Model. Through conversations, seminars and papers I gained a better appreciation of the information processing assumptions of system dynamics that distinguish the subject from traditional micro-economics on the one hand and optimisation methods in management science on the other.

Modelling for Learning

After more than 10 years at MIT, I returned to England in 1986 to join London Business School. John Stopford made possible this return and I joined him in the School's Strategy department. From this new academic base I entered a productive and enjoyable phase of 'modelling for learning'. I was invited by Arie de Geus to collaborate with his Group Planning department in Royal Dutch/Shell, based at the headquarters of Shell International in London. There, over a period of six years, a series of modelling projects (some conducted by me, and others conducted by David Kreutzer and David Lane) unfolded within the framework of Arie's 'planning as learning' initiative. The idea was to take a fresh view of planning and decision making in organisations and see them as collective learning processes. A vital empirical finding, from educational psychologists' studies of child learning, was that learning and doing often go hand-in-hand; children learn as they play. Arie de Geus made the logical step from child's play to decision making by play. It was a big step. But it was insightful if you took the idea seriously, as he and others in Group Planning did. Modelling and simulation fit naturally with this new approach to planning since models are in essence representations of reality (toys) and simulators allow role-playing with a modelled (and much simplified) reality.

An important consequence of my collaboration with Arie and Shell was the launch, at London Business School, of a week-long residential executive education programme called Systems Thinking and Strategic Modelling (STSM). The programme used learning-by-doing to engage executives with the core principles of feedback systems thinking and system dynamics modelling. Chapter 2 (Introduction to Feedback Systems Thinking) and Chapter 3 (Modelling Dynamic Systems) are derived from STSM. Moreover, the programme brought together, for a period of 10 years, a faculty team at London Business School that helped to develop system dynamics in many important ways and materially contributed to the content of this book. The team members were Arie de Geus, Erik Larsen, Ann van Ackere and Kim Warren and then later Shayne Gary. I enjoyed working with this special group of people and know that together we accomplished a great deal. Thanks to you all.

The shower models in Chapter 4 were sparked by Erik Larsen who felt, in the spirit of modelling for learning, that we shouldn't simply lecture STSM participants about the tricky balancing loop in a shower 'system'. Instead, we should build a simulator that would allow participants to see (or even experience) the resulting dynamics. So together we developed prototype simulators that became the basis for the World of Showers A and B models in Chapter 4. Alessandro Lomi and Ari Ginsberg later joined us to write a journal article based on these models, entitled 'The dynamics of resource sharing – a metaphorical model'. Two MBA students at London Business School, Thomas Furst and Derrick D'Souza, helped me to develop an early version of the gaming interface, and my wife Linda Morecroft worked on the user guide and interface enhancements for World of Showers.

There is an anecdote to accompany the shower project. After Erik Larsen and I had formulated the model's equations, we needed to supply parameters. Erik suggested that the 'desired temperature' of the shower should be set at 25 °C. I asked him if that number was high enough. He said it didn't matter as the choice would make no difference to the resulting dynamics, which was what we wanted the model to demonstrate. He was right in principle, but in practice (as I discovered by taking a thermometer into my home shower) water at 25°C feels distinctly cool. Erik was not easily moved by this piece of empirical evidence and so, as an amusing compromise, we decided to locate the model's imaginary shower taker in a hot and humid climate where a cool shower would be both desirable and plausible.

Perhaps the most memorable project from the modelling for learning era was a study of the structure and long-term dynamics of global oil markets. This study, conducted with the help of Kees van der Heijden, led to the Oil Producers' model described in Chapter 8. At the time, Kees was head of Group Planning's renowned scenario development team. He brought together 10 Shell managers who contributed to the model's conceptualisation. The project was a good opportunity to engage these managers with the model building process and to build a model that captured a collective view of their oil world as the basis for subsequent scenario development. The original Oil Producers' model was developed in the iThink modelling language. But several years later, prompted by a suggestion from Erik Larsen, the model's equations were transported into Visual Basic and a dramatic new interface was overlaid as the basis for experimental work on decision making in dynamically complex environments (the global oil industry is certainly dynamically complex). This work was carried out by Paul Langley as part of his doctoral thesis at London Business School ('An experimental study of the impact of online cognitive feedback on performance and learning in an oil producer's microworld', November 1995) and led to the Oil Producers' Microworld, which is in the CD folder for Chapter 8. My wife Linda worked on the user guide and documentation accompanying the microworld. Meanwhile, Paul went on to create other gaming-simulator interfaces for system dynamics models developed by the STSM team at London Business School, such as the popular Beefeater Restaurants Microworld and the ingenious BBC World Service Microworld. He subsequently took his modelling and gaming design skills to McKinsey's Business Dynamics practice in London.

Systems Thinking and Strategic Modelling ran bi-annually for 15 years and brought system dynamics to more than 700 managers and senior staff from countries around the world.

The Dynamics of Strategy

Around 1995, I began working with Kim Warren on the dynamics of strategy. This development was motivated by our shared interest in strategy (we were both in the Strategy department at the time) and also by our familiarity with a widely cited paper in the academic management literature entitled 'Asset stock accumulation and sustainability of competitive advantage'. The paper was written by INSEAD's Ingemar Dierickx and Karel Cool and appeared in Management Science in 1989. Their argument was that the sustainability of firms' competitive advantage could be better understood by thinking about the way firms accumulate the asset stocks that underpin their business. A firm might achieve competitive advantage by building-up a distinctive set of asset stocks that rivals would find difficult to imitate. Sustainability of competitive advantage would stem in part from the time it takes to accumulate or reconfigure such assets. We realised that here was a dynamic view of firm performance that could be further developed by formally linking system dynamics (SD) with the resource-based view (RBV) of the firm (an important branch of contemporary strategy theory and practice).

Our way of carrying out this SD-RBV synthesis was to jointly design and launch a new MBA elective course at London Business School, which we called the Dynamics of Strategy, abbreviated to Dynos. It ran for the first time in the spring term of 1996 and still runs today alongside strategic modelling and business dynamics SMBD. In addition, with Aime Heene and Ron Sanchez, I edited a book about systems perspectives on resources and capabilities that grew out of a Strategic Management Society mini-conference held in Oslo in 1998. Also, several PhD theses were written on SD-RBV at London Business School by Edoardo Mollona, Shayne Gary, Abhijit Mandal and Martin Kunc. The theses used the RBV literature as a backdrop for dynamic resource-based studies and models of important strategy topics such as the dynamics of diversification, resource heterogenity from feedback and the dynamics of competitive industries. Incidentally, all four dissertations were part-funded by commercial research partners, my London version of the business–academic partnerships I had known at MIT. Two of the research partners were past participants of STSM: Francois Delauzun from BBC World Service and Bill Howieson from Scottish Power. Another partnership was with the London Office of McKinsey & Co., during 1996–2000, when the Business Dynamics practice was in full-swing. The company assembled a strong team of consultants with expertise in modelling, and they provided a sounding board for many fledgling ideas about system dynamics and strategy. My thanks to Andrew Doman who led the Business Dynamics initiative in London and to Maurice Glucksman, Paul Langley, Norman Marshall, Panos Ninios and Hendrick Sabert who collaborated on various Dynos-related projects and publications.

There are fragments of this strategy dynamics work in Chapter 6 of the book, particularly the materials on the rise and fall of People Express Airlines in the CD folder for Chapter 6 (inspired by John Sterman's well-known management flight simulator). Also, Chapter 10 includes edited extracts from Martin Kunc's dissertation about product growth dynamics and industry competition in fast moving consumer goods. Kim Warren went on to further develop the SD-RBV theme in his Forrester Award-winning book, Competitive Strategy Dynamics.

Meanwhile, I did some hands-on resource building for London Business School when I became Associate Dean of the Executive MBA Programme (a 24-month modular MBA for participants who also hold a full-time job). During my three-year term we expanded the intake of students and also designed and launched EMBA-Global, an innovative 18-month, dual-degree Executive MBA offered in collaboration with Columbia Business School in New York City.

It was about this time that I encountered rather strong academic turbulence at London Business School. A number of friends and academic colleagues helped me, in various different ways, to weather the storm and to keep afloat the ship of system dynamics. They include Dennis Sherwood, David Andersen, Derek Bunn, Zeger DeGraeve, Rob Goffee, Costas Markides, Peter Mitchell, Alastair Nicholson, John Quelch and Ken Simmonds. Without their help then, I doubt this book would ever have been written. My wife Linda, my family and our retreat in North Cornwall also provided stability and perspective at this moment, as they have done throughout my career.

Soft Systems and Complementary Modelling Methods

In November 2001, I was invited by Mike Pidd of Lancaster University Management School to join the INCISM network, and it was here, in a series of meetings that spanned two years, that I learned much more about soft systems than I had previously known. INCISM is an abbreviation for Interdisciplinary Network on Complementarity in Systems Modelling and its meetings were funded by the UK's Engineering and Physical Sciences Research Council (EPSRC). The network brought together a mix of academics and practitioners to explore the combined use of what have become known as 'hard' and 'soft' approaches to systems modelling. One result was a book entitled Systems Modelling – Theory and Practice. Through the network, the book and subsequent conversations with both Peter Checkland and Mike Pidd, I have come to better understand where system dynamics fits on the hard–soft model spectrum. It seems to me that the juxtaposition of system dynamics and soft systems methodology (SSM) reveals, in tangible terms, quite a lot about the abstract philosophy of modelling – by which I mean the different ways in which modellers interpret situations in business and society. I touch on this topic in Chapter 2 (under 'event-oriented thinking') in Chapter 5 (under 'modelling for learning and soft systems') and again in Chapter 10 (under 'mental models, transitional objects and formal models'). INCISM also inspired a plenary session on soft systems and modelling at the 2004 International Conference of the System Dynamics Society in Oxford. Presentations by Mike Pidd and Peter Checkland described the territory covered by hard and soft modelling approaches and opened up discussion about the role of both qualitative and quantitative system dynamics. In the UK there is a long tradition of qualitative system dynamics which was started by Eric Wolstenholme and Geoff Coyle. The Oxford conference built on this tradition with its theme of collegiality as a social and scientific process to mediate between competing or complementary world views.

My interest in complementary modelling methods has been further reinforced through collaboration with the Operational Research and Information Systems (ORIS) group at Warwick Business School. It all began in October 2003 when I took a much-needed sabbatical year. I decided to split my time between home, London Business School and Warwick Business School. My home town of Beaconsfield is conveniently located on the Chiltern train line between the two Schools, about 25 miles from London and 80 miles from Warwick. Within ORIS, I found colleagues working at the interface of operational research and strategy and others who were interested in simulation methods. We had much in common. The use of complementary models and frameworks for strategic development became the focus of activity for an informal research group that included Robert Dyson, Maureen Meadows, Frances O'Brien, Abhijit Mandal and Alberto Franco (from Warwick), Jim Bryant (from Sheffield Hallam) and me. In monthly meetings, joint teaching, seminars and workshops we examined a variety of strategy support methods such as scenario planning, drama theory, system dynamics, SWOT (strengths, weaknesses, opportunities and threats) and the resource-based view. Our intention was to understand how they can help firms create and rehearse strategic initiatives as an integral part of a 'managed' strategic development process. This collaborative work is reported in Supporting Strategy, a book edited by Frances O'Brien and Robert Dyson (2007) with contributions from the core research group and others.

At Warwick, I also found experts in discrete-event simulation (DES). We soon discovered a shared interest in simulation methods that transcended our differences. With Stewart Robinson, I conducted a mini-project that compared system dynamics and discrete-event models of fishery dynamics. We each built a small model of a fishery following the normal modelling conventions of our respective fields. Then we compared notes. The project led to many interesting conversations about modelling and simulation. Some of our thoughts and conclusions are reported in the appendix of Chapter 9 on alternative simulation approaches. Although both system dynamics and discrete-event simulation are commonly viewed as hard system modelling approaches, their comparison illustrates an interplay and clash of world-views worthy of a soft systems study. In a sense, this comparison was our mini-project as we built separate fishery models and then reflected how our professional backgrounds led us to interpret and represent the problem situation in fisheries. For me, the fishery models also opened the door to the discrete-event simulation community (or at least showed where the door was), making possible further collaborative research with Ruth Davies, Sally Brailsford and others at the boundary of system dynamics and DES.

Book Writing and Production

My sabbatical at Warwick provided a period of tranquillity in which to plan this book. The first step was to gather materials from my London Business School lectures. An MBA course in Strategic Modelling has run at the School for 20 years and introduced almost 1000 graduate students to system dynamics. Any long-lived course evolves. The content of Strategic Modelling and Business Dynamics has benefited significantly from applied research projects I conducted at Royal Dutch/Shell, BBC World Service and Mars Inc. There are even distant echoes of projects at Harley-Davidson and Eaton-Kenway dating from my MIT days. Faculty friends and colleagues at London Business School have also contributed to the evolution of the course. Over the years Strategic Modelling has been taught by Ann van Ackere, Shayne Gary and Scott Rockart, who each brought their own interpretations to the core materials. My thanks to them for the innovations and refinements they introduced.

The main period of book writing spanned the years 2004–2006. For part of that time, Martin Kunc was completing his doctoral thesis at London Business School and we had many enjoyable conversations about modelling and problem structuring that provided a stimulating backdrop to my writing. Martin's thesis also contributed directly to the book. The soap industry model in Chapter 10 is derived from an applied research project and field study reported in his thesis. The model's special blend of conceptual and empirical richness made it ideal as a practical illustration of tests for building confidence in system dynamics models. Martin's work also brought into focus the interpretive stance of system dynamics during model conceptualisation, as neatly illustrated in the contrast between Chapter 10's preliminary and refined models of the soap industry. Supervising his thesis enabled me to see more clearly the links between phases of my own work in modelling for learning, dynamics of strategy and soft systems/complementary modelling methods. My thanks also to Gail Hohner for setting up the soap industry project that led to so many useful insights about the modelling process and its role in strategic development.

As draft chapters took shape many people helped to transform the raw manuscript into a completed book. Martin Kunc, Linda Morecroft and Mark Ratnarajah each read and corrected individual chapters. Special thanks to Shayne Gary who read almost half the manuscript and made numerous perceptive editorial comments. The models available on CD form an important part of the book. My thanks to Joanne Egner, Jesse Richmond, Anna Richmond and Karim Chichakly for providing the iThink software to run the models, and for adapting it to fit the particular purpose of the book (while taking an active interest in the entire project). Graduate student volunteers Joanna Bartlett, Simon Bennett, Stephen Boyle, Krish Rohatgi and Antuela Tako read the text, tested the models and checked the instructions for running them.

I am also indebted to the late Barry Richmond for his vision and energy in creating the iThink modelling environment. Although there are several excellent software tools available for system dynamics modelling, iThink's emphasis on visual modelling is particularly suited to my book. This compatibility is no accident since Barry was passionately committed to the communication and dissemination of system dynamics. Hierarchy, sector frames and enforced consistency of model diagrams and equations are three attributes of the modelling environment that reflect his bias as a visual modeller. Also, he worried and cared about the distinction in models between information flows and hard-wired causal links. As a result the original dotted-line convention for information flows from Industrial Dynamics is retained. I use this small but important visualisation feature repeatedly in the book's models.

My thanks to the book production team at Wiley for their friendly and professional support throughout the project. Sarah Booth initiated the book, won Wiley's support, and diligently followed its progress. Emma Cooper managed all stages of production from page design to copy editing and typesetting. She also handled the CD model collection and software provision. Anneli Anderson worked tirelessly on permissions for both the book chapters and the CD content and set up the book website. Wendy Mould carefully copy-edited the text and Emma Redfern helped with the proofreading. Meanwhile, my secretary Suzanne Shapiro kept my other School activities going smoothly while the manuscript was in preparation. As always, she remained cheerful and dedicated, even in the most stressful of times. Finally, I thank my wife Linda and my three daughters Katie, Jennifer and Rachel who have provided the motivation, love, stability and perspective of family life that I needed to complete this book.