Manager’s Journal Wall Street Journal 12/13/00

What's the best way to organize a group of people so as to maximize productivity and innovation? Does centralized or even dictatorial leadership yield better results than diffuse responsibility? Should you organize your group as a single unit, or divide it into a few large units, or into many small ones? Should you foster secrecy, or free communication, or competition-among your units, and between your group and other groups?

These are general questions applying to groups of all sizes and kinds-from whole countries, to industries, to industrial belts, to companies, to small offices. Obviously, the answers depend partly on the idiosyncrasies of individual leaders. For example, Microsoft's success has surely had something to, do with the unusual qualities of Bill Gates. Conversely, even a business organized on the best principles won't thrive with an idiot as CEO. Nevertheless, one can still ask: All other things being equal, what form of organization works best?

I believe that we have much to learn about group organization from human history. History tells us of thousands of societies, organized in diverse ways, with widely varying outcomes in productivity and innovation. What can we learn from these natural experiments of history that will help us all get rich? What emerges is what I call the "Optimal Fragmentation Principle."

To understand this principle, consider the contrasting technological histories and political organizations of China and Europe. Why has China lagged far behind Europe technologically since the Renaissance? The usual answers contrast China's supposedly ultra-conservative Confucian tradition with Europe's supposedly stimulating Judeo-Christian tradition and capitalism. But those answers dissolve under scrutiny. In the early Renaissance, Confucian China led the world in innovation and technology, with inventions ranging from cast iron and compasses to gunpowder and printing: The real question is:

Why did China then lose its lead to late-starter Europe?

The demise of China's ocean-going shipping illustrates the answer. By the year 1400, China had by far the biggest, best and most numerous ships in the world, sailing in gigantic fleets to Indonesia, India, Arabia and East Africa. But those fleets came to an end during an episode of isolationism. Like other countries, then and now, the Chinese government included a pro-navy faction and an anti-navy faction. After 1433 the anti-navy faction convinced the emperor that spending so much money on ships was wasteful. Anywhere else in the world when one country happened to be in an isolationist phase and not building ships, some neighboring countries would still be building them. But because China was huge and unified, Chinese explorers had nowhere else to turn for support when the emperor said "no." The emperor's order to dismantle the Shipyards applied to all of China and was irreversible. Other imperial decisions, stopped the development of clocks and water-driven industrial machinery throughout China.

Contrast that outcome with ship development in Europe. Columbus, an Italian, made nine unsuccessful applications for support to various princes of Italy, Prance, Portugal and Spain before Spain's king and queen finally gave him three ships. Columbus's resulting "discovery" of America, and the wealth that subsequent Spaniards amassed, stimulated 11 European countries to build ships and jump into the colonial scramble.

That is, Europe's political fragmentation gave Columbus many second chances and fostered fleet expansion through competition. Yes, there were also European princes who temporarily said "no" to ships, printing or guns. But Renaissance Europe was divided into 2,000 competing principalities, so there was no one with the authority to abolish a whole technology throughout Europe, as happened in China. Whenever one state tried something that proved valuable, other states saw the opportunity and adopted it.

And if China illustrates the disadvantages of excessive unity, the Indian subcontinent, which was hyperfragmented politically, illustrates the disadvantages of excessive disunity. Innovation proceeds most rapidly under conditions of some optimal, intermediate degree of fragmentation.

Now, let's apply this principle to modern industries. We Americans sometimes fear Japan's supposedly greater industrial productivity. Actually, America's overall productivity is higher than that of Japan, but that average figure conceals differences among industries related to differences in their organization. Let's consider an instructive case study of the American and Japanese food-processing industries, carried out by McKinsey & Co.

The efficiency of the Japanese food-producing industry is a mere 32% of ours. Two factors have doomed Japanese food processors to be organized into non-competitive, inefficient, small companies enjoying virtual local monopolies: a Japanese cultural trait and government policies.

The cultural trait is that Japanese are fanatics for the freshest possible food. A milk container in the U.S. bears one date the expiration date. In Japan the container bears three dates: the date of manufacture, the date it reached the supermarket, and the expiration date. Milk production in Japan starts at one minute past midnight so that milk reaching the market in the morning can be labeled as today's milk. If the milk were bottled at 11:59 P.M., the container's date would have to say that the milk was made yesterday, and no Japanese consumer would buy it. As a result, a milk producer in northern Japan cannot compete in southern Japan because transportation delays would add one day to the date on the container, a curse of death for milk sales.

These local monopolies originating in a Japanese cultural preference are reinforced by the Japanese government. which hamstrings competition from foreign producers by imposing arbitrary food import restrictions like a 10-day quarantine. Imagine the apoplectic reaction of a Japanese consumer, already skittish about one-day-old milk, on being offered month-old food shipped from overseas and then quarantined. Hence Japanese food-producing companies are not exposed to competition with each other nor with foreign imports, don't acquire economies of scale and don't learn the best international methods of food processing.

Similar organizational problems cripple the productivity of Japan's soap, beer and computer industries-but not its differently organized steel, metal, car "and electronics industries, which have higher productivity than their U.S. counterparts and-inspire well-earned envy in the U.S. Did you ever wonder why your TV and car, but not your computer or soap, are Japanese made?

Finally, let's apply these lessons to the U.S. The companies of Silicon valley compete fiercely with each other, yet there is also much flow of ideas, information and people between them. In contrast, the companies of Route 128 in Boston have been more secretive and insulated, like Japanese milk producers. As a result, Silicon Valley has overtaken Route 128's early lead as our center of innovation

Microsoft vs. IBM - - - - clipped out - - - - - - -

So we don't have to guess about organizational principles, set up a business on that basis, and wait 20 years to see whether our business goes bankrupt. Instead, we can learn by studying history’s thousands of completed natural experiments in organizing human societies. The approach teaches us that excessive unity and excessive fragmentation are both ultimately harmful; the best organization is to break up your business into groups that compete and generate different ideas but maintain relatively free communicate with each other. That's also the genius of our federal system of government, in which our 50 states compete fiercely and try their own experiments, but in which institutional barriers to interstate flows of people and ideas are non-existent.

Mr. Diamond is a professor of physiology at UCLA Medical School. His book "Guns, Germs, and Steel" (Norton, 1997) won the 1998 Pulitzer Prize for General Non-fiction