(6 of 9)
So is Microsoft a classic monopoly? Not exactly. For one thing, it fosters competition, at least within some well-defined boundaries. As Gates likes to point out, nearly 70,000 independent vendors produced software in 1993, more than three times as many as did five years earlier. Nor has Microsoft lowered prices to drive competitors out of business and then raised them to gouge consumers. "Microsoft cut prices when it didn't have to," says Jesse Berst, publisher of the industry newsletter Windows Watcher. "It's something that Apple could have done if they'd had the courage, and IBM too."
What Gates understands better than anyone else is that control over the standards that others must adhere to is the great lever of wealth and power in the digital age. Just as the value of a telephone network increases with each new phone added to the system (because it gives everyone on the network one more person to call), so does the value of a computer system increase with each program that runs on it. This is what Stanford economist Brian Arthur calls the law of increasing returns-a twist on the classic law of diminishing returns.
Diminishing returns, for those who slept through Economics 101, is the rule that explains why the second candy bar never tastes quite as good as the first and why doubling the amount of fertilizer on a crop field doesn't necessarily double the yield. But in the software industry, the second candy bar often tastes better than the first, and the third tastes even better.
So desperate are programmers for a stable computer standard that they will latch on to the first one that works-whether or not it is the best. A case in point, says Arthur, is ms-dos. "Microsoft dos for any serious user is a crummy operating system," he says. "But Microsoft got there first and played its market advantages extremely intelligently."
Using a strategy Arthur calls "target, leverage, link and lock," Microsoft proceeded to convert dos users to Windows users, Windows users to Word users and so on down the product line. Microsoft's customers, of course, were free to switch to WordPerfect for Windows or Lotus 1-2-3 for Windows or any other competing products. But by the time those programs were ready, Microsoft already owned the markets. "You could argue," says Arthur, "that Microsoft is the product of clever strategy, mediocre technology and a hell of a lot of increasing returns."
The links that bind users to Microsoft software run deep-down to the level of "plumbing" invisible to the user but of critical importance to the programmer. For example, Microsoft has for several years been using a tool called ole (for object linking and embedding) that makes it easier for users to move information from one Windows application to another-to shuttle an expense report from a spreadsheet to a word-processing document, for example. Apple is developing a competing system, called OpenDoc, that is supposed to run equally well on Windows, Mac and IBM's OS/2. But if a programmer uses OpenDoc instead of ole, he runs the risk that Microsoft will at some future point make a subtle change in Windows that would render his programs useless. "Everything Microsoft is building is linked to ole," says Jerry Michalski, editor of the newsletter Release 1.0. "Everywhere you look there are these dependencies."