Linux For Suits

August 2000

Monthly columns about volatile industries are always a bit risky, because you write them a month or two before they come out. This means every topic is a bet that the world two months hence won't be much different thanright now. This time I'm betting on a real long shot: that when you read this, Microsoft will still be one company -- and will eventually prevail against the government''s antitrust case. I'm also betting on a much more sure shot: that the trial's outcome won't matter much to the world we're building on open source infrastructure.

Microsoft is a risky subject in any case, but it's especially risky when you have anything nice to say about them and your audience is a bunch of Linux folks. It's like going to Miami to talk up Fidel. So I'll try to be as rational and brief as I can -- at least about the nice stuff.

I'll start by positing a thesis: that Microsoft success has many good reasons and one bad one. All the good reasons reduce to orthodox marketing, which is not about "delivering messages" and "capturing eyeballs," but rather about finding what customers want and giving it to them, even if it's talking paper chips. The bad reason is behavioral. Microsoft is often brutal, to its partners as well as its enemies.

The first reason explains the success of Microsoft's two cornerstone product families, Windows and Office. Both are what you get when a large, aggressive desktop software company does what orthodox marketing demands (in the words of Theodore Levitt): "satisfy the customer, no matter what." For decades, Microsoft took that advice to a whole new level -- while its opponents ignored it.

In 1996, when I spoke briefly at a small retreat of computer industry types, my topic was "Markets are Conversations" (which later became the first thesis of The Cluetrain Manifesto). In that talk I said Microsoft succeeded by being the only PC productivity software company that made sure customer feedback informed -- and even drove -- engineering. Afterwards one of Microsoft's top people took me aside and said, "That's the first time I've heard us credited with Marketing 101. The truth is, we do it to a fault. That's why our products are so bloated. Excel is full of stuff that exactly one customer asked for." Later an Apple executive told me, "We hate our customers. That's why we try to scrape them off on resellers." Then he told me the story of a frustrated tech support group that chartered an airplane from Austin to San Jose, drove to Cupertino and presented the CEO with a long list of problems that engineering had been refusing to fix. Bear in mind that both Microsoft's and Apple's true customers are intermediaries -- resellers and OEMs -- not users. That means listening to those users is highly optional. Microsoft deserves credit for at least being curious, whether or not their tech support sucked.

Later, when I consulted a Seattle company that had a close working relationship with Microsoft, they agreed that Microsoft did a good job of closing the loop between tech support and engineering. But they added this: "Since most tech support calls are from dumb users, you get dumb features." Exhibit A: the talking paper clip. This also squares with Steve Jobs' four-word indictment of Microsoft: "they have no taste." (Which says as much about Steve as it does about Bill.)

So how could a company that listens so closely to users fail to understand the antitrust rules by which increasingly dominant companies are supposed to compete?

Because Microsoft isn't a competitive company. It's a combative one. The difference seems obvious until you consider that both use almost identical vocabularies to describe the world. Those vocabularies are supplied by what linguists call conceptual metaphors. Markets have many conceptual metaphors, but the two that matter most for Microsoft are sports and war. Competitive companies understand business in terms of sports. Their conceptual metaphors are business is sports and markets are playing fields. Combative companies understand business in less civilized terms. Their conceptual metaphors are business is war and markets are battlefields.

By these metaphors, "competition" and "competitiveness" mean very different things, even though both use nearly identical vocabularies. A competitive company will "attack," "defend," "flank," "command," "dominate," "control," "retrench" and so on. So will combative companies. The reason is that sports is nothing more than civilized warfare, and leverages the entire vocabulary of war.

But the difference in meaning is both subtle and absolute: In a war, all is fair. In sports, fairness is all. War has no rules. (Ask the people of Sarajevo and Hiroshima.) Sports is full of rules. Battlefields are rarely level. Playing fields are always level. War is barbaric. Sports is civilized. Both have winners and losers. In sports, the losing team may get "killed" by the final score, but it lives to play again. In war, people die. The loser submits utterly to the winner.

Business is not regulated by rules of war. The "competitiveness" about which business law speaks does not comprehend the right of one company to "cut off the air supply" of another. Microsoft tried mightily in The Trial to convince Judge Thomas Penfield Jackson that the computer and Internet worlds were battlefields, and that every compeitor's very existence was under constant threat. Microsoft failed because Judge Jackson -- and the entire regulatory system -- conceives business as sports, not war. There was no way Jackson could buy the argument. There are rules for competition. Microsoft broke them. So Judge Jackson wants to break Microsoft. Literally.

To a combative company like Microsoft, a breakup is a death penalty. Utterly unacceptable. But Judge Jackson doesn't have the power to execute Microsoft, which belongs to the appeals court or the Supreme Court. He does, however, have the power to save Microsoft, if he makes enough mistakes.

It is significant that all through this ordeal Microsoft has refused to settle, or even to admit the obvious fact that they play hardball.

There was hardly a more ludicrous moment in the trial than when Robert Muglia, a Microsoft executive, insisted that when Bill Gates writes in an email that he is "hardcore about NOT supporting" a hunk of Java technology, the words mean something other than what they say. Jackson bought none of it, and finally yelled "No! and "Stop!" at Muglia before leaving the bench to compose himself. "By asking the judge to swallow such absurdities as Muglia's testimony," wrote Joe Nocera in Fortune, "Microsoft instead gave Jackson an excuse to swallow nothing."

This looked like Muhammad Ali's rope-a-dope strategy to me. Stay on the defensive, dodge constantly, and look for the other guy to make mistakes. The prosecution has more than the burden of proof here. They have the burden of their own rules. If they break enough of them, Microsoft reckons, the final ruling goes to the defendant, who maintains the presumption of innocence throughout. And all Microsoft loses is time and fees. Indeed, Microsoft attorneys have been saying for some time that they expect to win on appeal.

But this isn't just a story about Microsoft. It's also about Netscape. The judgement against Microsoft was based mostly on the company's behavior toward Netscape, whose air supply it allegedly wished to cut off. Nobody, however -- including Microsoft -- gives due credit to Netscape for choking itself.

True, Microsoft was out to get Netscape. I remember Jim Clark, on stage at the first Netscape conference in 1996, responding to a question about "polarizing" remarks made by Netscape people about Microsoft. "They're out to kill us!" Clark yelled back. "That has a polarizing effect!" Quite true.

But Netscape was born with enormous advantages. It made the first popular browser, which quickly achieved a nearly unanimous market share. Browsing was understood on Netscape's terms. The whole company essentially took sides with the Net, which offered enormous strategic advantages in an increasingly Net-driven world. And they used some of those advantages very cleverly.

One was buying and quickly ubiquitizing LDAP, the Lightweight Directory Access Protocol. In an interview for Linux Journal in 1996, Craig Burton said, "Microsoft built its entire services strategy on what it thought was a titanium vise. One side was an object-oriented file system called OSS which collapsed the directory into the file system. This was Cairo (now Windows 2000). The other side was a distributed application development framework called OLE (now ActiveX), which they owned lock, stock and barrel. They would squeeze those together and the Netscapes of the world would squoosh like jello. But the Internet blew the jaws of that vise apart. When Microsoft tightened the jaws of that vice, they bent wide open. The world has shifted, and Microsoft is not going to dominate it." One of those shifts -- a big one -- was LDAP. There were others, mostly involving the growing popularity of the Net itself, that went in Netscape's direction.

For a year or two, Netscape looked like it could do no wrong. But then it seemed to get high smoking its own publicity exhaust. It publicly implied that its browser obsoleted Microsoft's (and everybody's) operating systems (Marc Andreessen said an OS was "just a device driver"). It also turned the browser from a tool of Demand (browsing) to an instrument of Supply. They added channels during the "push" craze. They portalized their Web site. They turned the Netsite bar into a search term window for a whole new name space, to be populated by the identities of companies that paid to be put there. Worst of all, they bloated the browser from a compact single-purpose tool to an immense contraption that included: an authoring tool, a newsgroup reader and an email client. Worst of all, they charged money for their browser and gave useless customer service, even to their paying customers.

It was painful to watch. Netscape was family for me (a close relative was highly placed in the company). I was such a partisan that I paid for that damned browser the whole time they charged for it, and I never got useful customer service. One time a tech support person told me that the company really didn't care about the browser because the big money was in server software.

And then there was engineering. By the time Netscape was sold (at top dollar) to AOL, the dirty secret was that its browser code was terminal, and had been for a long time. The Mozilla project was launched with a mess of unfinished code that has since been largely abandoned. In other words, Mozilla.org had to start approximately from zero a year or more after it appeared to start from an advanced position. Jamie Zawinski (one of the company's first and best-known engineers) put it bluntly: "Netscape was shipping garbage, and shipping it late." Not exactly competitive.

Today we forget that the browser was born free. At its birth, Mosaic was little different in spirit than Linux, SendMail Apache or any other free software project at its inception. That's why making browsers officially free (in cost if not in code) was a brilliant and insightful move by Microsoft that was utterly lost on Netscape, which was unwilling to play smart chess by making a reciprocal sacrifice. It was a suicidal mistake. The company we loved and rooted for is history.

Its ghost may still have revenge on Microsoft. It was Netscape, more than any other company, that pushed the antitrust case against Microsoft, and then served as the exemplary victim of Microsoft's rapacious behavior. If the Feds bust up Microsoft, Netscape deserves more credit than any other company.

But let's have some perspective here. What happened between Microsoft and Netscape may make interesting stories -- whether told in terms of sports or war . But the story that matters can't be told in those terms. It needs a better metaphor.

My favorite, of course, is conversation. As a topic of conversation, Microsoft has been shrinking for the last five years, regardless of what happened to Netscape or what revenge the Feds carry out. There are too many other, more interesting, subjects to talk about.

Craig Burton, Brian Behlendorf and Tim O'Reilly all prefer the environmental metaphor. That's what I've learned from conversations with each of them. The wide-open world each of them envisions is a habitat where, development tools, methods and strategies matter more than the companies that employ them.

Programmers have put the building metaphor to good use. They are the ones who will design, architect and build this new world, using free software tools, Linux building materials and open source buildilng methods -- along with everything else lthat gets the job done.

If Microsoft can help, whether as one company or as three, that's great. If they can't, it's probably because they're too busy looking for enemies. And giving us less and less to talk about.

Doc Searls is Senior Editor of Linux Journal and co-author of The Cluetrain Manifesto.