Most people can grasp very simple systems — say, a system with one stock, one flow, and one feedback loop. Cash flows in and out of an individual’s checking account (a stock) as that person makes deposits and writes checks. When the balance gets low, he or she must stop spending or start earning. But even the simplest organizations have multiple stocks and flows that operate in interconnected networks. Professor Forrester postulated that most industrial activity could be represented by five networks — materials, orders, money, capital equipment, and personnel — with a sixth, the information network, functioning as the connecting tissue between the other five. The complex interactions among the different networks, each of which has its own set of stocks and flows, and the feedback delays inherent in the information network make true cause and effect difficult to gauge.
For his first book, Industrial Dynamics (1961), Professor Forrester drew on the experience of his MIT students. They were typically managers, age 30 to 40, who were eager to bring their problems to Professor Forrester in the hope that his computer simulations could help them.
One of the consistent findings was particularly disturbing at first glance: The problems of most companies were not brought on by competitors or market trends, but were the direct result of their own policies. “People discover that their own policies inevitably generate their troubles,” says Professor Forrester. “That’s a very treacherous situation because if you believe these policies solve the problem, and you do not see that they are causing the problem, you keep repeating more of the very policies that create the problem in the first place. This can produce a downward spiral toward failure.”
On the other hand, once the links were revealed, companies could often fix the problem by changing some small but consequential practice that happened to influence all the other factors of the system. (Professor Forrester calls these “high-leverage” solutions.) For example, an electronics-component manufacturer suffered from inexplicable losses of market share. Professor Forrester’s model showed the culprit to be the company’s policy of buffering itself against downturns by waiting to hire more factory workers until there was a large backlog of orders. This had given the company a reputation for slow deliveries, which caused customers to lose interest, which led to falling orders — which, in turn, made the manufacturer even more cautious about hiring, and thus even more prone to backlogs. The solution was simple: Maintain a steady work force even during occasional downturns, while building up enough inventory to improve delivery times. (See Exhibit 2.)
Contemporary reviewers compared Industrial Dynamics to the works of Galileo, Malthus, Rousseau, and John Stuart Mill, and it is still required reading in many MBA programs around the world — a remarkable feat for a 40-year-old book packed with dense text and intricate diagrams. Its success brought Professor Forrester consulting engagements at major companies. But the assignments often frustrated him. Executives would listen politely to his presentation, and go on with the same problematic practices. Even at Digital Equipment, managed by his former students, he found system dynamics a tough sell.
“I was never successful in getting the board to believe the models would work,” he says. “The last time I tried, one of them said, ‘We agree that we’ve been successful following your advice, but it’s not because of your modeling. It’s just because you’re a better manager than we are.’ That excused them from having to pay attention to the source of my insights.”
Part of the problem was the computers of the day. Jay Forrester would run his models at the MIT computer lab and return to the company with a paper printout. Clients “could see the logic of the result, but they had not internalized the process of getting there,” he says. And managers did not enjoy hearing that they had caused their own troubles; they just wanted a solution. But another part of the problem was Professor Forrester’s own impatience: He chafed at the time it took to explain his solutions.