First in a series of posts that lay out a process for building sound, useful theories in simple, practical ways. We welcome your feedback as we develop them.
—John Maynard Keynes, The General Theory of Employment, Interest and Money
Anyone who bought an American-made car in the 1970s is likely to have a few unhappy memories associated with the experience. A door handle would fall off, sometimes before the car had left the dealer’s lot. The heater fan would quit working at the first cold snap. The problems went beyond the occasional. The “fit and finish,” as car people say, was generally poor. So was the mechanical quality. Many hapless buyers found themselves taking the vehicle back in for countless repairs and adjustments.
Usually these fixes worked well enough, and buyers made the best of what they had purchased. But sometimes the repairs just piled up, and the car would be declared a “lemon.” When a technician repaired or replaced one component that had broken in a lemon, another component and then another seemed to let go. Multiple system failures eventually conspired to make repair next to impossible. It was a frustrating situation for both manufacturers and car buyers. Buyers wanted a car that worked. Manufacturers wanted to stop having so many cars towed to their service centers, usually with irate customers in the passenger seat of the tow truck.
From one point of view, it’s not surprising that lemons were common. A typical car contains nearly 30,000 individual items, everything from big parts like steering wheels to tiny parts like clips, covers, wires, and bolts. Many of these, like starter motors and seat assemblies, are prebuilt and shipped to the assembly plant. Still, a typical auto manufacturing line will receive around 2,000 unique parts from several hundred suppliers in as many as 17 countries. Doesn’t the complexity of taking so many things from so many different sources and turning them into a working car a miracle in itself? Isn’t an occasional lemon a small price to pay for the privilege for the promise of a well functioning automobile?
That seemed to be the prevailing wisdom at the time. Indeed, many manufacturers had so much trouble getting the job done well that the rubric lemon began to be slapped on whole brands. There was the Chevrolet Corvair (“unsafe at any speed,” wrote Ralph Nader) and the Chevy Vega. There was the Ford Edsel in the 1950s—a car almost no one wanted to buy, partly because of its poor quality—and, later, the Ford Pinto, whose gas tank was liable to explode. In that era, carmakers believed that quality in manufacturing was essentially a random process. You did the best you could, but you accepted the fact that a significant minority of cars you made would not work right. Life was like that.
Life was like that—until it wasn’t.
The change began in the late 1950s. Inspired by the work of two Americans, W. Edwards Deming and Joseph Juran, Toyota began developing a new approach to building cars.[i] Part of the Toyota Production System, as it came to be known, focused on reducing cost and increasing productivity by eliminating waste. But part focused on improving the quality of every part in the car so that the finished product would work precisely as designed. Two essential theories informed this view. The first: Instead of accepting that manufacturing was inherently probabilistic or subject to randomness, leaders of the quality revolution insisted that every defect has a cause.[ii] If a part is improperly made, there is a reason. If parts or subassemblies don’t fit together as they should, there is a reason. Things appear to be random only when we don’t understand their causes. The second theory was that every time you do something, it is a test to see whether doing it that way leads to the outcome you want. If it does, you will repeat the process just as before. If it doesn’t, you should modify the process in a controlled manner until you get the desired result.
Toyota applied these theories to the production of its cars, and the results were dramatic. It developed a reputation for the best quality in the industry. It grew to be the leading importer of cars to the US, and would later become one of the biggest auto companies in the world. Nissan and Honda followed Toyota’s lead in the quality revolution. So, at varying rates, did other Japanese, Korean and (eventually) European and American automakers. Year after year their products have become higher in quality and lower in performance-adjusted price. And the lemon has virtually vanished from view.
So here is the moral of the tale. Designing and manufacturing cars is an intellectually challenging problem. Thousands upon thousands of people try to resolve all the issues that come up, over and over again. Yet engineers and managers cannot teach all those people the optimal way to do their job and then continually monitor compliance. Rather, they developed the two theories mentioned above to guide employees’ work, and then they trusted the employees to improve their product in accordance with the theories. The theories helped them teach each other how to do the job better. Over time, lemons disappeared.
The aim of DisruptiveInnovation.Org is to do something similar for theories: to improve the theory of disruptive innovation and to improve all of our collective understanding of how theories are (and should be) built.
[i] See Womack, Jones and Roos, The Machine that Changed the World: The Story of Lean Production--Toyota's Secret Weapon in the Global Car Wars That Is Now Revolutionizing World Industry and the deluge of books and articles that have since been published.
[ii] Surprisingly, there has been a venerable tradition in economics, management, and social science research generally that rejects causality as its stalking horse. See, Campbell, D. T. (1988) “Can we be scientific in applied social science?” Elster, J. (1983) Explaining Technical Change. Cambridge, England: Cambridge University Press. See also, Osterloh, Margit and Frey, Bruno S., “Corporate Governance for Crooks? The Case for Corporate Virtue” (July 2003), where the authors argue that management research has the tendency and power to become self-fulfilling prophecy. This means that effects posited by causal mechanisms are actually caused by the behavior of the readers of new management research. ZEW Working Paper No. 164. http://dx.doi.org/10.2139/ssrn.430062