Stephen Sacca, SF ’90
Director, MIT Sloan Fellows Program
You’ve been there, I’m sure. You’re on the phone with a service provider waiting to transact a simple task, but you spend long minutes on hold between short conversations with people who seem at a loss to help. You know there has to be a better way, and no doubt there is. One way of discovering that better way is through system dynamics (SD)—the analysis and redesign of any interdependent social, managerial, economic, or ecological system. System dynamics was a mid-century invention, the brainchild of MIT Sloan Professor Jay Forrester during the School’s earliest days, but Forrester’s concept has gained legions of adherents over the decades, including many Sloan Fellows, some of whom studied with Forrester.
In these pages, you will meet MIT Sloan Fellows alumni and faculty—including Forrester himself—who are innovating a diverse array of processes around the world with the help of SD. The sheer versatility of the approach is astonishing, but more astonishing still are the outcomes.
“When it comes to system dynamics,” says Professor of Management Science and Organization Studies Nelson Repenning, “MIT Sloan is both the birthplace and current center of expertise. The toolset was invented here, and I think it’s fair to say that we continue to offer a depth, rigor, and variety of applications you can’t find anywhere else.”
The origins of the field lie in the work of MIT Sloan Professor Emeritus of Management Jay Forrester. Forrester arrived at MIT in 1939 and joined MIT Sloan in 1956 with a desire to apply his background in science and engineering to the core issues that determine the success or failure of organizations.
Although computer simulation models have become synonymous with system dynamics (SD), Forrester’s initial investigations were calculated by hand. Forrester assembled a team of graduate students, including MIT Sloan Fellows, to interview people in a company about how they made decisions at their individual operating points. “By tapping the wealth of information people possessed in their heads,” Forrester says, “we were able to build a model that revealed overall organization behaviors that resulted from the interconnected actions of many local decision-making individuals.”
In the late ’50s and early ’60s, Forrester and his students began building their simulation models with computers. By translating the important policies, information flows, and interconnections of a company into computer instructions, they were able to create a model that could act out individual roles at each decision point. The computer would then feed the results of those actions into other connected decision points, allowing the model to generate the basis for its next round of simulated decision-making.
The technology enabled the team to create a laboratory replica of a company that could be used to observe the behavioral consequences of company policies. “What we showed,” explains Forrester, “was that the policies believed to be the solution to the company’s problems were, instead, the cause of those problems.”
System dynamics modeling also provided the basis for a new generation of teaching tools known as management flight simulators. One of the most famous, The Beer Game, was developed in the late 1950s by Jay Forrester and his students at the MIT System Dynamics Group. The game puts teams of students in charge of managing four components of a distribution chain. Players within each component—retailer, wholesaler, distributor, and factory—receive shipments of beer, fill as much of their customers’ orders as possible, and place new orders for beer with suppliers.
The goal of The Beer Game is to keep costs low while meeting customer demand. As students make operational decisions and receive feedback from past decisions, they often are surprised by the unintended side effects and delayed consequences of their actions. Huge oscillations arise as most people order based on current inventory while ignoring the supply line of beer they’ve ordered but not yet received. Participants also learn how much better it is to be surprised in a board game than in the boardroom. By game’s end, management students gain insight into the complex dynamics of a supply chain—and deeper lessons about human behavior, including teamwork and trust.
The game can be raucous, but it’s also a sophisticated teaching tool. “Management flight simulators like The Beer Game are effective because they provide learning environments that motivate participants,” says John Sterman, the Jay Forrester Professor of Management at MIT Sloan. “Simulators provide experiential as well as cognitive lessons, and they compress time and space so that we may experience the long-term consequences of our actions.” The Beer Game has also proved to be a robust foundation for many subsequent SD simulation models, including one for managing severe anemia in dialysis patients.
Working with a cross-disciplinary team at the Mayo Clinic in Rochester, Minn., MIT Management of Technology alumnus James Rogers SF ’86 helped to create a biomedical system dynamics model that improves the protocol for hemodialysis patients who are being treated for anemia. Severe anemia is usually treated with injections of erythropoiesis stimulating agents (ESAs), which stimulate the cells in bone marrow that generate red blood cells. Standard dosing protocols, however, often lead to wide fluctuations in red blood cell and hemoglobin levels, which put patients at risk of serious harm, including heart attack, stroke and death. In simple terms, the standard treatment adds to the stock of circulating red blood cells, but only after a long delay. Because ESAs are typically administered based on the current hemoglobin level, they often miss the mark.
“The anemia control model is a lot like The Beer Game,” explains Sterman. “The standard treatment is analogous to ordering beer based on current on-hand inventory without considering the beer you’ve already ordered—it ignores the supply line of red blood cell precursors. The result, not surprisingly, is oscillation.” The Mayo Clinic team used their simulation model to develop and test an innovative dosing protocol that can be individualized for each patient and administered easily. The new protocol greatly reduces the amplitude of hemoglobin oscillations and decreases adverse events—improving patient survival and quality of life while reducing treatment costs.
Sterman’s recent work on the healthcare system includes helping to launch ReThink Health, a multidisciplinary collaborative effort committed to overcoming the paralysis that has thwarted efforts to build healthier communities. The ReThink Health team includes leaders in health, economics, politics, business, and energy, including MIT Sloan alumni Jack Homer, PhD ’83, Gary Hirsch, SB ’69, SM ’71, and Peter Senge PhD ’78, who is also a member of the MIT Sloan faculty. Over the past five years, ReThink Health has been the launchpad of promising SD models for the treatment of chronic conditions like heart disease as well as community-based healthcare initiatives.
The ReThink Health team has developed management simulators for communities like Atlanta, Georgia, Alemada and Contra Costa counties in California, and Pueblo, Colorado. “All the stakeholders in the community are included in the process,” says Sterman, “and everyone’s ideas are tested in the interactive SD model. People are often surprised by the results. They use the simulation to design more effective interventions, building on one another’s ideas to design better policies to promote sustainable, affordable improvements in health care and health.”
Meanwhile, Forrester has taken system dynamics into yet another realm—public education, which he considers the most exciting frontier for SD. Nearing his 96th birthday, Forrester is devoting much of his energy these days to The Creative Learning Exchange (CLE), a nonprofit organization dedicated to encouraging public education leaders to incorporate systems thinking and SD into classrooms across the United States and around the world. CLE provides free curriculum, books, and games that promote systems thinking. It also sponsors a biennial conference to help educators and students use SD in teaching and in school organizations.
“Children are capable of understanding the fundamentals as early as kindergarten,” Forrester says. “By fifth and sixth grade, students are working with models that we used to use only in graduate school.” Forrester doesn’t expect that children who are taught SD in schools will spend their lives in front of computers building SD models. Rather, he believes they will be more effective at interpreting and interacting with the world in which we live—fast-paced, evolving, and unpredictable. “Our collective life experiences are the basis of system dynamics. Educating more of the public about how these models work will increase the frequency with which we make good decisions and create more successful outcomes.”
System dynamics is by its nature disruptive. It’s about using models and simulations to test the outcomes of policies and the veracity of their component parts. Organizations are often wedded to existing procedures and system dynamics “outs” the flaws in the system. MIT Sloan Fellows are introduced to system dynamics during their year at MIT and many say it has been an indispensable tool in their work. A few of them have shared their experiences with us here.
Graham Rong, SF ’06
Senior Industrial Liaison Officer
“In a recent Global Entrepreneurship Lab (G-Lab) project at MIT Sloan, we worked with a leading European company specializing in security products. The company wanted to produce a system for identifying counterfeit auto parts, a formidable global challenge. In fact, loss from fraud in the auto parts industry exceeds $2B a year—not surprising when you realize that automotive manufacturers produce somewhere in the range of 200,000 different parts and that nearly 10 percent are being counterfeited.
But the damage far exceeds that $2B in measurable revenue loss. Fraudulent parts, which often don’t perform according to expectation, can seriously damage brand equity. If your BMW is always breaking down, neither you nor your mechanic might ever realize it’s a result of cheaply made fraudulent parts. Bar codes are not enough to prevent theft, because they too can be counterfeited. So our G-Lab team used system dynamics to design a fool-proof track and trace model using radio-frequency identification (RFID) tags and tag readers. Certified Parts manufacturers issue a warranty on a part only if the information on the part’s RFID tag matches that in the electronic product code (EPC) database.
The system dynamics model tests that every stakeholder along the supply chain will derive value from the new system—in this case, automobile manufacturers, parts makers, dealers, consumers, as well as the company producing the security solution. The model we used proved that this innovation would increase revenue, secure the brand, and prevent theft and that all links in the chain would benefit. This is the first study ever conducted that uses system dynamics to analyze RFID applications. Modeling and analyzing the causal loop for the individual company and the industry as a whole made it possible to understand the interdependence of the key variables and their ramifications.”
Daniel Ross, SF ’11
Chief Operating Officer
Wholesome Wave Charitable Ventures
“With little access to fresh, locally grown produce, low-income people in this country face the grave consequences of diet-related diseases like obesity and diabetes. And farmers—especially owners of small, local farms—are losing out on business with those underserved consumers. At Wholesome Wave we are using system dynamics to understand how we can change the food system to make it work better for both low-income consumers and smallholder farmers.
I was first introduced to system dynamics in the MIT Sloan Fellows Program during SD class with Anjali Sastry and during Leadership Lab (L-Lab) with Peter Senge. We were working on a project involving the diverse community use—and environmental impacts on—the Gulf of California in Baja California Sur, Mexico. I saw that SD was a highly effective way of helping very diverse stakeholders grapple with the large, complex system that was central to all their lives. Everyone, from local fishermen to the millionaire owners of luxury hotels on the beach, could see the impacts made clear in the model.
During that experience, I learned how to leverage SD to change minds and systems. When I took over as COO of Wholesome Wave, I wanted to see how we could use SD to align our efforts to the key levers of the food system. I worked with staff to rough out some models. One key outcome was the very effective scaling of the Double-Value Coupon Program (DVCP) for use at local farmers’ markets. The program gives consumers using food stamps twice the buying power when they redeem their stamps for healthy fruits and vegetables at farmers markets. Government programs, foundations, and corporate partners provide the matching funds.
The program has been so successful that more than 400 markets and 60 community-based organizations in 27 states are participating, and we were able to add a provision to the Farm Bill that included $100M for nutrition efforts at farmers markets. Ninety percent of DVCP consumers say they have boosted their consumption of fresh fruits and vegetables, while participating farmers report increasing revenues as a result of the program. We are now working with fellow alum Jeroen Struben, MIT Sloan PhD ’06, who is on the management faculty at McGill University, to develop a full simulator model [see illustration] showing the impacts of investment into nutrition and community food systems. We think it’s a model that can have enormous impact on the nutrition problem here and around the world.”
Devon Kinkead, SF ’08
Founder and CEO
“In my experience, system dynamics can be a game-changer. I've used system dynamics to develop the insights needed to make high stakes, non-intuitive, strategic decisions. Here’s an example from my last start-up. It was a capital equipment company that faced all the usual high tech manufacturing challenges. For one thing, it was an incredibly cyclic business—up 80%, down 40%, up 120%, down 60%. Fortunately, the net growth worked out well over time, but the board of directors did not enjoy or understand the wild ride.
What we needed was a big picture, and system dynamics gave us that. I consulted my former MIT Sloan professor Nelson Repenning and over the next six or seven months developed an understanding of what systems thinking was about. Eventually, we built a model of the system we were living in, one that mapped the ebbs and flows of supply and demand through the system. The board grew comfortable with the model and confident that we knew what was happening and could predict business conditions better. Our environment wasn't just ‘one damned thing after another’ but a system that could be understood and predicted. So we did something that was counterintuitive: we borrowed money when we didn’t need it so that when we were next in the predicted down cycle, we could invest through it and come out the other side with stronger products to gain market share.
It turned out to be a good move, a move that was made possible by gaining a high level understanding of the system in which we were operating. We began to restructure the company so that we wouldn’t have to empty the building every few quarters, outsourcing variable costs and setting fixed costs at a level that we knew we could afford in the downturns. SD was absolutely key to running that business through its cycles intelligently and strategically. We brought the company to a successful exit ten years after entering the semiconductor capital equipment business.
Realizing how powerful a tool system dynamics was, I began to see the possibility of solving other problems. I am working with a nonprofit agency called HomeStart dedicated to ending the cycle of homelessness by moving people from shelters to permanent housing. With Nelson’s help, we are rethinking homelessness in Greater Boston. We’re pretty convinced that, as a society, we don’t fully understand this system and that we’re throwing money at the wrong parts of it.
Working with one of Nelson’s grad students, we want to devise a system dynamics model that illustrates the problem of homelessness in all its component parts. Using SD, we hope to refocus the problem and get a new perspective on solutions. To reverse this scourge, we need to reach a deeper understanding of the root cause of homelessness and the economics of prevention, and SD will help us do that. After evaluating the SD model, we’ll make specific policy recommendations to Massachusetts Governor Deval Patrick and to the Department of Housing and Urban Development (HUD).”
Martin Capriles, SF ’05
PTO Capital LLC
“After finishing the Sloan Fellows Program in 2005, I headed back to my home country of Venezuela, where I was working for CEMEX. I was responsible for all commercial operations of the company’s Venezuelan subsidiary.
At the time, we were facing a lot of major constraints in the market—the fall of a major bridge in the country, the risk of being nationalized by the government, storage limitations, as well as the displacement of the demand away from our major production facilities, complicating logistics. This pressure was also reflected in internal, interoffice tensions. The challenge was to refocus the company as a whole to fulfill the market on a just-in-time basis. We needed to get everybody on the same page—quickly.
To achieve this, I went back to what I learned at MIT and organized a Beer Game scenario, John Sterman’s renowned system dynamics model (see above). I started working with top management, then moved to the directors, then to the managers of our facilities and other key personnel. The idea was to put the production team in sales, sales in supply, and so on, so that each of them would be able to visualize through the game the pressure the other faces, relate that to the real world experience, and conclude the necessity to work together.
The debriefs were outstanding and the results became quick victories. This simple exercise allowed the company to work as one. We also established a logistics system based on system dynamics that resulted in the optimization of our resources, which allowed us to prepare for even more challenging scenarios.”
Patricia Winand, SF ’13
Senior Vice President
Sales & Marketing Americas
Close the Loop
“The first time I stopped to think about loops was when I studied system dynamics at MIT. For me, SD was love at first sight because I am a very spatial person, and my brain responds best to visual stimuli. Having someone graphically explain a complex problem made it so clear, and it turned out to be one of my best learning experiences at MIT Sloan.
The beauty of system dynamics is that it is a wise combination of simplicity and analytical power. I enjoyed learning the technique, but I must admit that I was not sure how I would use the tool in the real world. The professor who taught the course was a master at the subject, but my million-dollar question was “how can a mere mortal do it?”
After graduation, I realized that the course had changed my mindset and that I was using SD automatically when analyzing problems and identifying their root causes. I joined a company called Close The Loop (any reference to system dynamics is purely coincidental), which operates in the clean technology, recycling, and environmental services space. My mission as SVP, Sales & Marketing Americas, is to develop differentiated offerings leading to sustainable competitive advantage, to redesign the company’s business model, to expand in South America, and to provide vision and direction on the marketing program priorities—a tall order.
So, I organized our first brainstorming meeting to develop the company’s future business model using the best mapping tool ever—system dynamics. One of my team members had said, ‘loop means a lot to us,’ so I thought they would naturally enjoy system dynamics thinking. They did.
Thanks to system dynamics, we’re going forward with a plan that the entire team understands, is committed to, and can explain to all the disparate players up and down the value chain.”
We’re already at work on the next MIT Sloan Fellows Program Newsletter. Please drop us a line at firstname.lastname@example.org if you have ideas about themes and news items for future issues.
Stephen Sacca, Director
Marsha Warren, Associate Director
Mary Marshall, Associate Director
Marc O’Mansky, Assistant Director
Libby Dilling, Marketing Associate
Saul Horowitz, Program Coordinator
Elizabeth Mackell, Program Assistant II
Davin Schnappauf, Program Assistant II
MIT Sloan is a registered trademark. All trademarks are the property of their respective owners.
© MIT Sloan. All rights reserved