Systems Thinking, Fast and Slow
In 2011, the Nobel Prize-winning economist Daniel Kahneman published his best-selling book “Thinking, Fast and Slow”, which extended the work on cognitive biases he had started with his co-author Amos Tversky many years earlier. Kahneman defines two systems of the mind, “System 1”, which jumps quickly to answers through simple heuristics, and “System 2”, which analyzes decisions more deliberately and looks for more evidence.
A person (or organization) with a lazy or undeveloped System 2 can get along fine in routine decision-making but will tend to make mistakes—jumping to incorrect conclusions and bad decisions—when the situation is more complicated and nuanced. The biases causing these mistakes go by such names as confirmation bias, availability bias, anchoring, and the fallacy of “what you see is all there is” (WYSIATI).
Systems thinking and modeling are supposed to help individuals and organizations make decisions more holistically and with all significant factors in mind—that is, to strengthen the role of System 2 in the world.
However, even a systems approach can fall prey to the usual cognitive biases of System 1 if it becomes dominated by certain loud voices in the room, gives too little consideration to all the evidence, applies a rigid template to a situation with potentially important idiosyncrasies, or otherwise rushes to conclusions.
Fast versus Slow may seem to pose a dilemma in the business world, where potential clients have limited time and budgets at their disposal. Although everyone knows that a thorough and careful job is ideal, some systems thinking consultants may be tempted to offer the promise of powerful insights using a “quickie” approach (e.g., using only causal-loop diagrams and no data collection) that takes weeks rather than months. Naive clients may attribute supernatural abilities to a smooth consultant who makes such promises and forget the fact that the problem at hand is difficult and has stymied all previous attempts at a solution.
Although such a consultant may get lucky a handful of times, with clients satisfied in the short term and glad to give a thumbs-up testimonial, that consultant’s luck will sooner or later run out. Short-term improvement will lead to later disaster, and the inadequacies of the consultant will be revealed for all to see. The rise and fall of a single consultant may be no cause for alarm, but if that consultant has convinced an army of others to take the same approach, the result could be catastrophic for the field as a whole. That field will have allowed its sober System 2 side to be smothered by the false attractions of go-with-the-gut System 1.
Fortunately, there is a way around the apparent dilemma, as many experienced system dynamics consultants already know. The solution is to propose a small initial simulation model (supported by data collection but not at a detailed level) for the first several weeks of work. A satisfied client may then choose to go on to a definitive second stage involving more complete data collection and expanded modeling.
As System Dynamics practitioners, we should reaffirm the scientific method and avoid being seduced by “fast” System 1 with its inevitable trajectory of overshoot-and-collapse. We must have confidence that a reliable System 2 approach will serve our clients—and our field—best in the end.
by Jack Homer (VP of Professional Practice)
Recent Posts
Call for Presenters: Seminar Series
Call for Presenters: Seminar Series We at the System Dynamics Society are continually seeking vibrant and knowledgeable presenters for our ongoing Seminar Series. As we unfold the calendar, there’s always a place for more insights, experiences, and expertise to enrich...
Honoring Excellence: A Glimpse into the Awards of the International System Dynamics Conference
Honoring Excellence: A Glimpse into the Awards of the International System Dynamics Conference The International System Dynamics Conference brings together experts, practitioners, and students to exchange ideas, showcase real-world applications, and celebrate...
From Bergen to Global: UiB’s System Dynamics Group
From Bergen to Global: UiB’s System Dynamics Group The System Dynamics Group, an autonomous research group at the University of Bergen (UiB) was established in 1971 by professor emeritus Svein Nordbotten. Inspired by the work of Jay W. Forrester, Nordbotten...
Upcoming Events
India Chapter Meet Up w/Bob Eberlein
We are very excited to invite you to another year-end meet up and chat, this time featuring Bob Eberlein – who has been a vital member of the SDS since its early days and also a pivotal developer of Vensim, and now Stella – some of the most widely used...
Recent Business cases
The World Bank Uses System Dynamics to Identify Root Causes of Poverty
The World Bank Uses System Dynamics to Identify Root Causes of Poverty EXECUTIVE Summary Madagascar has one of the highest poverty rates in the world. In 2022, an astonishingly three out of every four people in Madagascar lived below the poverty line. Poverty has...
Fast-Track Cities Uses System Dynamics to Enhance HIV Care
Fast-Track Cities Uses System Dynamics to Enhance HIV Care EXECUTIVE Summary Low levels of viral suppression at 69% for people with HIV make it hard to believe the 95% target level will be achieved by 2030 in St. Louis, USA. As a solution, Fast-Track Cities-STL opted...
System Dynamics Unravels ICU Tensions at the Portuguese Oncology Institute
System Dynamics Unravels ICU Tensions at the Portuguese Oncology Institute EXECUTIVE Summary The Portuguese Oncology Institute (IPO) faced a critical challenge in its intensive care unit (ICU) where doctors and nurses experienced high turnover due to tensions with...
Join us
Hi
I am again this dualistic or Manichean approach. Instead of opposing system thinking and system dynamics for methodological reasons it would be better to take a client’s point of view, and compare their respective results, making it possible to find how to move smoothly from a ST approach to a SD one. This move would permit to find the right mix of SD and SD starting proposal depending on the nature of the client and his needs and the type of problem, leaving after this first experience, the possibility to move into a more quantitative approach. JJ.
Correction: I am against this dualistic approach and right mix of ST and SD. Sorry for the mistake.
Hi Jack and JJ,
I think a wholistic approach at the outset with the client may be the most time efficient approach. I have read Kahneman’s book and perceive there are valid reasons to integrate fast with slow thinking. I think present day society’s education system appears to so highly value fast thinking and getting an answer, too often at the expense of teaching the value of the slow thinking which how to generate better quality questions during the learning process. I think we need to promote more ‘Inquiry based learning’ and System Dynamics is so good at doing this! For SD practitioners, perhaps what could occur as an initial investigation into the client’s ‘problem’, would be an integrated ST and SD outline with the focus of this investigation being to outline a few possible avenues for SD simulation analysis which could bring insight to the problem. Outlining several possible SD avenues to the client would educate the client about what is possible using SD simulation analysis and also provide an opportunity for helpful feedback and also a sense of participating (agency) in the problem solving process by the client…a win-win step to take for both SD practitioner and client! Martha Toy
Hi folks, I didn’t mean to be doctrinaire or dualistic in my blog piece. I was rather just trying to suggest that careful, properly-done SD can rarely reach a definitive answer in just a few weeks’ time. If a consultant is offering “powerful insights” in just weeks, then generally some shortcut is being taken, and the client will fall victim to cognitive bias and “loudest voice in the room” syndrome. My suggestion was to start with exploratory modeling, which can be a mix of diagramming and simulation–but to be very clear that the insights coming out of that first exploratory phase will be only tentative. It may take some strength to tell a client that the first insights will be only tentative, but that, I believe, is what we should be doing. Otherwise, we’re cheapening the field. Jack
Without being an SD practitioner myself, yet having worked in another client financed professional setting for many years, I couldn’t agree more that initial education of the client about what to expect from professional service and how long it will take to achieve a truly valued and appreciated result feels essential to protect both the quality of the professional service and also public respect and trust in a benchmark standard of professional ethic. Martha.
Hi everybody
The too fast problem is not specific to SD. Being first a client problem owner before a modeler, I will illustrate this by 3 examples.
Forty years ago, I was associated with 3 other vehicle renters in a franchising system having roughly 150 locations in France, half of them being owned by ourselves. We met twice a year one day in Paris in a hotel near the airport to discuss and talk about our problems. This was unfortunately far from enough, and the result is that this franchise that should be today leader in Europe, is no more existing. And we did not need any special sophisticated method but just time and a minimum of ability to manage this kind of meeting.
In 2001, I was faced with a sudden decrease of the turnover and of the market price of the competition and thought wrongly that SD could help me. Unfortunately, I had not the experience nor the time to chose this strategy. What I needed was to spend much more time thinking about my problems with vey conventional methods and if I had done that I would not have been obliged to sell progressively all my locations in France later on.
In 2008, having studied SD since 2002, I tried to work with a renowned SD consultant on a specific problem. He proposed me a model built in 8 weeks. The model was finished 8 months later (bugged and poor) when I did not need it anymore.
Just to say that SD modeler should first learn to help clients with conventional methods before eventually trying if necessary other more sophisticated methods and there is a huge market for that. If this was done, I would not have attended a consultant meeting in the Washington SD conference in 2011, where the SD strategy consultants complained about their fees being half less than ordinary strategy consultants. And buying a french fry stand and trying to generate profits with it, will be much more helpful than any PHD to become credible when talking with a business client and understanding his problems.
JJ