Things are not going well for Microsoft, the world's largest software company, in either the courtroom or the marketplace. Microsoft was found guilty by a US court of being a "predatory" monopoly last April, and a judge ordered it to be split in two; the case goes to the appeal court at the end of February. Meanwhile, Microsoft is struggling to cope with the shifting landscape of the computer industry, where the rise of the internet, and its philosophy of "open" software standards agreed by consensus, has undermined its business strategy. Under the leadership of its co-founder, Bill Gates, Microsoft has weathered adversity before. But the confluence of these subtly interconnected threats presents an unprecedented challenge to its dominance of the industry. How did it get into this mess? And can it get out of it?
Windows on the world
The roots of Microsoft's woes lie in its efforts to defend the dominant position of its Windows operating system. Windows, which is installed on 90 per cent of all personal computers (PCs), is the centre of Microsoft's strategy, and the source of its power. It is a crucial layer of software that provides the "platform" on which all other programs sit, and through which they talk to the PC's hardware: keyboard, mouse, screen, network and so on. Software companies must make their programs compatible with Windows if they want to be able to sell them, and PC manufacturers must build Windows into their machines if they want anyone to buy them. Microsoft sits in the middle, controlling the standard and making a fortune in the process: around 100m new PCs are sold every year with Windows installed, and Microsoft charges an average of $50 per copy. Little wonder that Microsoft is the biggest software company on earth, or that Bill Gates (who owns a large chunk of it) is the world's richest man.
How did Microsoft end up in such a powerful position? Back in 1980, IBM was preparing to introduce the IBM PC, the original desktop computer that spawned the PC industry. This new computer needed operating-system software, and rather than write its own, IBM fatefully chose to license an operating system called MS-DOS from Microsoft, which was then a relatively obscure software company. But IBM, which dominated the industry at the time, made a big mistake. The licence agreement allowed IBM to sell computers with MS-DOS installed, but it also allowed Microsoft to license MS-DOS to other manufacturers. Before long, other computer makers had "cloned" IBM's design. With the addition of Microsoft's software, a cloned PC was just as good as a "genuine" IBM machine. The crucial ingredient-the software-was owned by Microsoft, not IBM. Within a decade, Microsoft had dethroned IBM as the computer industry's top dog. By this time, MS-DOS was showing its age, so Microsoft added Windows-a graphical system of windows, menus and so forth-on top of it. Since MS-DOS was already the industry standard, the success of Windows was assured.
Why has no challenger to Windows emerged? There are, admittedly, a handful of alternatives, notably Apple's Mac OS, IBM's OS/2, and more recently Linux, a free operating system developed by programmers collaborating over the internet. But none has been able to topple Windows from its dominant position, because software companies generally have little incentive to write software for operating systems that have a market share of just a few per cent. Write software for Windows, on the other hand, and you can be sure that it will run on the vast majority of PCs; and if you buy a PC with Windows installed, you can be sure of being able to run the vast majority of software available. Windows's dominance is thus self-reinforcing.
So far, there is nothing illegal about any of this. But Windows's near-monopoly of the PC operating system market gives Microsoft an enormous amount of power over both hardware and software producers. PC makers cannot afford to offend Microsoft for fear that it will refuse to sell Windows to them. Similarly, as Microsoft continually adds new features to Windows, software makers must watch out too. Suppose you run a software company that sells a particular program that runs on Windows. If Microsoft decides to add a similar program of its own to Windows, that software will then come as standard on every new PC with Windows installed. Even if Microsoft's free product is not as good as yours, your customers may decide that it is good enough, and your sales will plummet unless customers have a compelling reason to pay for your product. By choosing what new features to add to Windows, Microsoft has the power to crush software makers if it so chooses; since it is the biggest software company in the world, it can afford to spend as much as necessary to make its free add-ons to Windows just as good as any alternative offered for sale by another company.
Anti-trust investigates
In 1993 the US Justice department's antitrust division began an investigation of Microsoft's business practices to find out whether it was abusing its dominant position. In particular, the Justice department was concerned about Microsoft's ability to bolt on or "tie" extra software to Windows in order to kill rival programs produced by other firms. Another area of concern was a clause in Microsoft's contracts with PC makers, requiring them to pay for a copy of Windows for every machine they sold, whether or not Windows was actually installed on it. Without agreeing to this condition, PC makers would not be sold Windows at all, so they had little choice but to accept it. But it effectively prevented any competitor to Windows from entering the market, since any PC maker that chose to sell some of its machines with a different operating system installed would still have to buy Windows for each one anyway. Microsoft had, in effect, instituted a Windows "tax" on every new PC.
In order to avoid being taken to court, Microsoft agreed in 1994 to sign a "consent decree" drawn up by the Justice department. The company promised not to tie software to Windows to stifle competitors, and agreed to modify its contracts with PC manufacturers. But the terms of the decree were vague, and still allowed Microsoft to add features to Windows if they were truly "integrated" rather than simply bolted on. Almost immediately, it became clear that this was an enormous loophole. The consent decree was challenged in court but was upheld in August 1995 in a ruling by Judge Thomas Penfield Jackson, who deemed it adequate to restrain Microsoft's bad behaviour. So Judge Jackson was infuriated when Gates said on television that the consent decree had had "no effect."
Within months, an apparently cut-and-dried case of Microsoft's bad behaviour emerged with the rise of Netscape, the software company that pioneered the web browser. Netscape's unusual strategy was to give away its browser over the internet. Academics, students and home users could use it for free, but companies that decided to adopt it were expected to pay for it. With few other browsers available, Netscape quickly became ubiquitous, and by the summer of 1995, more than 80 per cent of people browsing the web were doing so using Netscape's software. Microsoft, which had been slow to grasp the significance of the internet and had instead launched its own proprietary information network, called MSN, suddenly realised that Netscape might be about to steal its lunch.
Netscape offered programmers an alternative "platform" to Windows. Instead of writing software to run on top of Windows, they might start writing software to run within Netscape's browser (though Netscape itself would, in most cases, be sitting on top of Windows). It would then be Netscape rather than Microsoft that would have a stranglehold on the software industry. Bill Gates decided something had to be done. Microsoft duly launched its own browser, called Internet Explorer. The firm gave this browser away for free on the internet, even to corporate users; and it also included it with every copy of Windows. As a result, during 1996, Internet Explorer's share of browser usage started to climb. Bill Gates boasted that his free browser was "priced to sell." Faced with a competing browser that cost nothing and came as standard with every Windows PC, Netscape was in deep trouble. It seemed clear that Microsoft was in violation of the consent decree, and pressure started to mount on the Justice department to do something about it.
In October 1997 US attorney general Janet Reno announced that the Justice department was seeking an injunction against Microsoft for violating the consent decree. Microsoft, for its part, argued that under the terms of the consent decree it was allowed to develop "integrated" products, and that the browser and Windows were integrated, rather than merely tied. But in December Judge Jackson ordered Microsoft to offer PC makers a version of Windows that did not include Internet Explorer. Microsoft haughtily responded that this was impossible; PC makers could choose between a version of Windows that was two years old, before Internet Explorer had been added, or a version of Windows that did not work at all. This proved to be a crucial tactical error by the company, because it enraged Jackson. Microsoft backed down almost immediately, but its disregard for the court's authority convinced the Justice department that the company could not be trusted.
Meanwhile, Netscape's allies (and Microsoft's rivals) in Silicon Valley, notably Sun Microsystems, started lobbying the Justice department to launch a full-scale antitrust case against Microsoft. As John Heilemann recently revealed in his book Pride Before the Fall, Sun paid $3m to high-powered legal experts who helped draw up evidence against Microsoft for presentation to the Justice department. But with the stock market riding high on the back of a technological boom, Joel Klein, the chief of the antitrust division, was ambivalent about taking Microsoft to court. Instead, he tried to negotiate a settlement, and asked Microsoft to make a handful of concessions, including allowing PC makers to decide whether Microsoft's or Netscape's browser should be installed on new PCs. But the negotiations broke down, largely because Microsoft did not want the government telling it that other software should be included with Windows. And so on 18th May 1998, the Justice department, together with representatives of 19 states, launched a suit against Microsoft. The company was, said the Justice department, guilty on two counts: it was a monopolist that crushed its rivals and harmed consumers, and it exploited its monopoly in operating systems to stifle competition.
the trial of the century
The trial, which began on 19th October 1998, was disastrous for Microsoft. Despite the difficulty of getting testimony from many industry executives whose companies were dependent on Microsoft, the Justice department managed to portray the firm as a bully, routinely using its market power to strongarm its rivals. Jim Barksdale of Netscape claimed that at a meeting in June 1995, Microsoft officials had tried to discourage Netscape from producing a browser to run on Windows 95, the newest version of Windows. If it chose to compete with Microsoft in this area, said Barksdale, a Microsoft executive had threatened to "cut off Netscape's air supply." And by tying its own free browser to Windows, Microsoft had subsequently done just that. Microsoft dismissed Barksdale's account as hearsay. But worse was to come.
Executives from AOL, America's leading internet-access provider, testified that it had chosen Microsoft's browser over Netscape's because Microsoft had offered them a valuable incentive: placement of an AOL icon on every Windows desktop. This had further eroded Netscape's share of the browser market, since all AOL users unwittingly adopted Microsoft's browser. (Netscape was eventually bought by AOL). A senior official from Apple, one of the few computer makers whose machines do not run Windows, explained that Microsoft had forced Apple to adopt its browser, rather than Netscape's, on all its new Macintosh computers. If Apple did not agree to this, Microsoft threatened to stop making its business software for Macs, without which they would not be taken seriously within large companies. Microsoft, it was claimed, also pressured Apple to stop giving away free multimedia software for Windows, because Microsoft wanted its own rival multimedia software to prevail. And emails between Microsoft and Compaq, another PC maker, showed that Microsoft had threatened to stop supplying Windows if Compaq started selling machines with Netscape installed. Other companies such as Intel and IBM also complained about unreasonable pressure from Microsoft.
Microsoft's defence was a mess. Rather than concentrating its fire on a few key areas, it unconvincingly challenged every statement made by its opponents, no matter how minor. The company refused to admit that it had a monopoly, dismissed its rivals' claims as gossip, and when presented with the damning e-mails of Microsoft executives (trying to stop Compaq promoting Netscape's browser), tried to reinterpret them. And as the government broadened its case beyond Microsoft's attack on Netscape, and presented examples of its bullying behaviour in other areas (such as its attempts to subvert Java, a software technology invented by Sun), Microsoft continued to behave as though the only issue at stake was the legality of tying. In its public statements, it protested that its "freedom to innovate," through adding new features to Windows, was under threat-and that by bringing the case, the government was thus threatening to derail years of growth and innovation in the computer industry.
Microsoft was on firmer ground when it highlighted the rapid pace of change in the software industry, which meant that competitors could, like Netscape, spring up almost overnight. The company also pointed out that the government had not demonstrated that Microsoft's actions had harmed consumers. Another common theme in Microsoft's defence was that the technological issues involved were extremely complicated, and were being misleadingly oversimplified by the court; Microsoft implied that only it really understood the finer technical details.
But the factor that did the most to undermine Microsoft's credibility in the courtroom was the videotaped deposition given by Gates himself. He was vague, evasive, and unhelpful in the face of a barrage of questions from the government's main lawyer, the charismatic David Boies. Gates claimed to be ignorant of his own company's strategy and decision-making processes, which could only mean he was forgetful, a liar, or did not actually control the company. Boies realised that Microsoft would not risk calling Gates as a witness after such an incoherent performance, since he would either have to defend the indefensible statements he had made on the video, or contradict his testimony. So instead of asking Gates to appear in person, Boies showed occasional clips of the video in court. Each clip was eagerly anticipated, briefly electrifying an otherwise tedious trial. And each clip made Gates look more foolish than the last.
When the first phase of the trial ended in February 1999, things were looking grim for Microsoft. Its lawyers were now assuming they would lose, and were laying the foundations for a subsequent appeal. Judge Jackson, however, made it tougher for them to win an appeal by announcing that he would issue his "findings of fact" first, and his verdict a few months later. The findings of fact would, of course, give a good idea of what his final verdict would be, and he hoped that this might encourage the two sides to negotiate some kind of a settlement in the intervening period.
The findings of fact, which were issued on 5th November 1999, were damning. Over 16 pages, Jackson disputed the widely touted idea that antitrust laws from the 19th century could not be applied to the computer industry. He outlined the "self-reinforcing cycle" of Windows's dominance. He flatly described Microsoft as a monopolist. And he said that Microsoft had used its monopoly power to stifle competition. His verdict, though unstated, was clear: Microsoft was guilty on all counts.
The verdict
A few weeks later Jackson appointed Judge Richard Posner, an appeal court judge and an acknowledged authority on antitrust, as a mediator. He hoped that Posner, whose previous writings and rulings implied that he might be sympathetic to Microsoft, would be able to negotiate a settlement between the two sides. He hoped Microsoft would be prepared to offer major concessions to avoid a guilty verdict, since such a verdict would make the company's monopoly, and its associated monopolistic behaviour, legal facts.
The inside story of the mediation effort has been recounted by Ken Auletta in his account of the trial, World War 3.0: Microsoft and Its Enemies. Ultimately, although Microsoft offered a number of concessions that went far beyond the negotiations that had preceded the trial in 1998, the mediation failed. In part this was due to Posner's failure to involve the 19 state attorney generals in the negotiations; he wrongly assumed that the Justice department spoke for all of them. When the states introduced their own additional demands at the last minute, the negotiations collapsed.
It was now clear that Jackson would issue a guilty verdict against Microsoft. But what action, or "remedies," would he recommend be taken against the company? There were two options: behavioural remedies and structural remedies. Behavioural remedies, the softer option, would involve placing legal restraints on Microsoft's behaviour: fining it, preventing it from acquiring other companies for a few years, requiring government lawyers to be present in all meetings, and so on. But Judge Jackson's previous experience with Microsoft, and its cavalier attitude to the 1995 consent decree, convinced him that behavioural remedies would be ineffective; the company would find a way to get around them. So the judge was more inclined towards a structural remedy: in other words, breaking up Microsoft.
There are a number of ways this could be done. One option would be to create two or more companies (dubbed "Baby Bills"), each of which would be identical to Microsoft, but smaller. Each company would then be free to develop Microsoft's software, including Windows, in its own way; the result, in theory, would be two or more competing versions of Windows. The problem with this option, however, is that these versions of Windows might quickly become incompatible with each other, which would fragment the industry. So a more attractive option would be to split Microsoft in two. One half would get Windows, and the other half would get all of Microsoft's other software, notably its business software, and the Internet Explorer browser. The idea is that this would establish a firm line between the operating system and other products, and prevent tying in future. And, in time, Microsoft's two halves might start to compete with one another.
It was this second option that Judge Jackson called for in his verdict, which was issued on 7th June 2000. As expected, he found Microsoft guilty on both charges brought by the government. And, as expected, Microsoft immediately appealed. Anxious to resolve the case quickly, given the pace of change in the industry, Jackson tried to get the case "fast tracked" to the Supreme Court, since the appeals process would otherwise take years. But in September 2000 the Supreme Court voted not to hear the case immediately, and instead ruled that the appeals process should be allowed to take its course. Both sides have since filed briefs outlining their positions to the appeal court, and the case will be heard at the end of February; a ruling is then expected in April or May.
Beyond the break-up
A break-up looks less likely following the election of George W Bush, who has hinted that he is opposed to one, and criticisms of Judge Jackson. So a behavioural remedy, if anything, now looks more likely. But this may not matter, for the landscape of the industry has changed dramatically since the trial began.
Though they are still reluctant to say so in public, Microsoft's competitors have been emboldened by the government's action. Several PC makers are now selling PCs with Linux, rather than Windows, installed as standard-something that would have been unthinkable before the trial. More significantly, the industry's centre of gravity is shifting away from the PC desktop, where Windows is dominant, and towards the internet, a realm of open standards and protocols that do not belong to any particular vendor. Growth in PC sales is slowing, while handheld computers such as the Palm Pilot, and a new generation of "smart" mobile phones, are in the ascendant. Indeed, handheld devices (rather than PCs) are expected to become the dominant means of accessing the internet within a few years. But Microsoft's attempt to promote a version of Windows for handheld devices has not been a success, since few manufacturers have signed up for it. Similarly, Microsoft has failed to make Windows the standard operating system for television set-top boxes. No doubt the evidence of Microsoft's behaviour towards PC makers, exposed during the trial, is partly to blame in both cases.
Even Microsoft has conceded that the PC is no longer as important as it was. The company's new ".NET" strategy, unveiled after the verdict last June, assumes that the future of software is as a service provided over the internet, rather than a product installed on a PC. This approach, which has long been championed by Microsoft's rivals, including Sun, Oracle and IBM, puts far more emphasis on powerful server computers than on cheaper PCs. And most servers, despite Microsoft's best efforts to beef up its software, do not run on Windows. Microsoft is encouraging programmers who now write PC software for the Windows "platform" to switch, and write internet software for the .NET platform instead. If it works, this strategy will insulate the company from the decline in the importance of PCs and Windows. It will also make a break-up of the company, if it happens, far less painful, since Windows will no longer be at the centre of everything Microsoft does.
Like IBM before it, Microsoft will probably go from being a big company that dominated the computer industry to being simply a big company. When the last big shift in the industry occurred, with the rise of the PC, IBM initially looked well-placed to benefit, but soon dropped the ball. Similarly, Microsoft's over-reliance on the supremacy of Windows now looks like a fatal error.
Breaking up Microsoft might do Gates a favour. He is increasingly involved in philanthropy and has been distancing himself from the day-to-day running of the company. If he retired it would draw a clear line beneath the legal battle and would give the company room to redefine itself. Gates has donated billions to medical research and in January, when asked for his vision of the world in 2020 by the Wall Street Journal, he said more about healthcare and agriculture than he did about computers. Is this a ploy to soften his image, or a change in priorities prompted by a bruising encounter with the law? We will soon find out.
