Bill Votes with a Veto
Bill Gates is an ingenious engineer, but I don’t think he is adept at business ethics. He has not yet come to realize that the things he did (when Microsoft was smaller) he should not have done when Microsoft was a monopoly.
—Judge Thomas Penfield Jackson1
The life of William Gates III has followed a unique arc. In the late 1970s Gates founded Microsoft with his friend, Paul Allen. Allen left for health reasons while the company was still young. Gates, along with college buddy Steve Ballmer, managed it for the next two and a half decades. Gates finally retired from his full-time position in 2006. In the process he became unbelievably wealthy—to the tune of $50 billion. Many of Gates’s associates also became extremely wealthy, Ballmer and Allen most notably. However, Gates’s wealth outshines everyone’s, making him comparable to only a small number of other business figures in history—robber barons in popular speak—such as Carnegie, Rockefeller, Bell, and Ford.
Gates has not always seemed certain how to manage his unique status in society and history. The social expectations, responsibilities, and worldly public persona often required by the media, public, and social stratosphere in which Gates lives were difficult for him to internalize and then project to the public. During Gates’s long tenure as CEO of Microsoft, a team of public relations experts tried to cover this particular weakness by managing or modifying how Bill appeared to the public. Yet every so often a crack would appear in Gates’s carefully crafted image. On occasion, one such public gaffe would illuminate the professional side of Gates’s life.
In volume 3 of Robert Cringely’s Nerds 2.0.1, a movie about the growth of the Internet, Cringely conducts an interview with Gates. During the time of the interview, Microsoft’s fortunes were growing, but the federal antitrust case against Microsoft had not yet been filed. Gates is not feeling defensive and is relaxed, even when Cringely asks Gates if Microsoft is losing money in its browser strategy.
Gates pauses, smiles, and then, speaking with just a hint of dry humor, says, “Well, nobody ever accused Microsoft of not knowing how to make money.” The smiling Gates then exhales in a little involuntary laugh, an introspective response to his own wit. There is nothing suave or planned about it—just a cross between a guffaw and a boyish giggle. After another pause, Gates provides an explanation of how Internet Explorer, or IE, helped the Windows business. The answer does substantively address the question, but it seems to dissolve in the revelation about the source. Gates is a socially awkward man who knows how to make money, and knows that he knows.
The Internet had not escaped Gates’s notice. However, for a time he did not see how Microsoft—or any other software firm for that matter—could make money in the browser business or any other Internet application. Gates had maintained a consistent position throughout 1994 and early 1995. He regarded the Internet as a useful piece of functionality and an additional feature of server software. Microsoft needed the functionality in its servers to compete with Unix systems. He saw no reason for Microsoft to develop browsers or experiment with prototypes of Internet applications to sell to end users. But by the spring of 1995, Netscape’s success had forced Gates to reconsider his assessment, and, as other chapters have mentioned, Microsoft’s behavior thereafter appeared to be focused on winning the market share for his browser at any cost.
Why did Gates behave this way? Why did Microsoft’s browser meet Netscape’s browser head on in the marketplace? What did Microsoft do and why did it upset so many in the industry? Contemporary commentators began calling the confrontation browser wars. The browser wars were more than just an ephemeral competitive episode in the Internet ecosystem. The strategies of both firms touched every single major participant in the commercial Internet, and shaped their actions. The episode also illustrates an important policy issue: how a large powerful firm can discourage the blossoming of innovation from the edges when it possesses the ability and incentive to do so. In this instance, the government took action, and the legacy of that intervention mattered for years thereafter.
What Happened inside Microsoft
By 1995 Microsoft combined aspects of a publicly traded company with many of the trappings of a privately held one. Gates helped take Microsoft public and had always owned a significant fraction of the company.2 Gates’s ownership coincided with considerable personal authority. He sat at the center of every major decision within the firm, and he retained the right to review any employee’s action.
Gates did not make money through commercializing one big invention, like Alexander Graham Bell. Rather, Microsoft made money by employing Gates’s uncommon savvy in the art of software commercialization. Gates made countless little decisions. The strategic and financial benefits accumulated over years, occasionally becoming manifest in a few well-known events, such as the rollout of the Office suite and Windows 95. Gates repeatedly approached situations with the intent to outthink and out-execute others. In pursuit of that approach, he became a devoted student of computer industry business models and lessons learned from other markets.
Gates’s approach was on full display in the key memo for the browser wars, “The Internet Tidal Wave,” which, as prior chapters described, Gates wrote after a night of surfing, arranged by the skunk works within his firm. Few CEOs could have written a single-spaced eight-page document that (a) summarized fifteen years of strategic positioning, (b) provided technical detail of that strategy, and (c) repositioned the firm’s priorities for the next epochal change. Gates had learned from the efforts to investigate browsers, and the memo showed that he grasped an impressive range of facts and theories.3
The timing for the memo shaped Gates’s analysis. Microsoft had set itself on a lucrative trajectory with Windows 95’s development, whereby Microsoft aspired to realize its ambitions through a two-pronged approach: first, provide every piece of ubiquitous software for PCs; and second, prevent any other firm from offering the same degree of functionality. Together these assured Microsoft a unique position: a piece of revenue from every PC purchase. As discussed in prior chapters, Netscape’s browser brought to life a version of Gates’s biggest nightmare—namely, the possibility that enormous functionality could be delivered to users through alternative channels—the Internet, in this instance. That functionality would be a substitute for software Microsoft provided.
Multiple versions of this nightmare would eventually grab hold of Microsoft’s strategic outlook. In one version, Netscape invited application developers today, and Netscape became an operating system tomorrow. In another version Netscape and related “middleware”—software that stood between the operating system and other applications—facilitated the entry of an operating system from another firm on PCs, or it facilitated entry of a range of new devices that accessed the Internet. In any of these versions, the strategic focus was the same; Netscape’s presence and prosperity reduced the value of Microsoft’s position on the value chain because it facilitated entry of substitutes.
While Microsoft’s actions later appeared to outsiders to follow a relentless logic, things looked different inside Microsoft. Gates initially proceeded over many tense discussions in which Microsoft executives disagreed with the substance of his nightmare. For example, the launch of Windows 95 was going well by the spring of 1995, and, accordingly, many inside Microsoft could forecast high sales with confidence. In that light those in charge of marketing and distributing the operating system did not share Gates’s vision of Netscape. They saw Netscape’s entry as unexpectedly good for sales of Windows 95 and Microsoft’s bottom line, because growth in Internet adoption would lead to more PC sales than originally forecast. In short, Netscape’s success appeared to immediately help the sales of Microsoft’s biggest product, Windows 95, not threaten it. (Indeed, this turned out to be a correct forecast of what the Internet and web did to PC sales over the next several years.) Employees with this vision took direct issue with the paranoia in Gates’s strategy and saw Netscape as a gift, and they believed Microsoft should seek to encourage it to blossom.
Was Gates prophetic or paranoid to change his mind about browsers? Gates had been right in the past, steering the company into its unfathomable position of profitability, and in most parts of Microsoft Gates had earned the benefit of the doubt. That will partly explain why he was able to induce many actions, though many employees did not fully perceive the threat, even at Microsoft’s largest and most profitable division.
He also could induce many actions because Gates had a comparatively free rein throughout the firm. Most employees did not perceive all that he did, or even part of it, unless it involved their own corner of the firm. Many employees looked only at their own division’s behavior.
The first version of a browser, IE 1.0, actually arrived at Microsoft before the “Internet Tidal Wave” was written, because it came from actions taken by the earliest investigations into browsers and the Internet. Shipped with Windows 95 in August 1995 as an additional “plus pack,” IE 1.0 was a mildly changed version of the browser licensed by Spyglass, the firm that licensed Mosaic for the University of Illinois.
The next key actions took place behind the scenes before IE 1.0 ever shipped. In June 1995, before Netscape’s IPO and before the official release of Windows 95, Microsoft contacted Netscape. Their stated intent: they wanted to make an investment in the young firm and, in return, gain a board seat. In fact, Microsoft’s discussions with Netscape had multiple motives. Firstly, to explore whether it was possible to gain control over Netscape’s action through an inexpensive board seat or merger; secondly, to explore whether it was possible to gain insight into Netscape’s strategic intent through mere conversation about the purchase of a board seat; and thirdly, to engage in a conversation and gain information that might help settle debates inside Microsoft’s headquarters in Redmond, Washington, about whether any of Netscape’s outsized aspirations conflicted with Microsoft’s goals.4 Eventually the first meeting turned into several, and those talks would evolve into discussions about Microsoft buying all of Netscape.
There was little publicly said about the talks until several years later, when the antitrust trial in 1999–2000 (discussed later in this chapter) shed light on them, and the lawyers for both sides fought over interpretations of the ambiguous notes taken at the meeting. The meetings and conversation did have lasting consequences because they altered the view that each side took of the other. Netscape refused Microsoft’s last offer for an acquisition, which Microsoft’s management interpreted as a signal of their outsized aspirations. Microsoft’s management believed their last offer had been respectfully high, and no firm would turn it down except a firm who intended to invade Microsoft’s core market. In contrast, Netscape perceived Microsoft’s offer for Netscape as much too low. Netscape’s managers gained the impression that Microsoft did not understand the potential size of the market opportunity for Internet software. They concluded that Microsoft’s market forecasts were either too pessimistic or misguided.
Whose perceptions were right and wrong? A couple months later, it did not matter. On August 9, 1995, Netscape held its IPO. This event settled many matters, both establishing Netscape as a publicly traded company and providing the investment community with a public demonstration of the potential for Internet firms. More to the point, the IPO settled any lingering debates inside Netscape and Microsoft. The management at Netscape believed events had shown they had been correct to hold out. The IPO suggested many investors had a very optimistic view of Netscape’s potential. After the IPO, inside Microsoft, Gates was no longer met with any opposition to his view that Netscape’s aspirations posed a threat, as he continued to push for a coordinated set of actions across Microsoft.
Two Platforms and Two Positions
In August 1995 Microsoft released the first version of Windows 95. The next key moment would come in early December 1995. To understand that moment, it is essential to appreciate the strategic similarities and differences in the positions of Netscape and Microsoft. These differences would translate into strategic advantages and disadvantages in their fight.
The strategic similarities arose from similar forecasts. Both Microsoft and Netscape anticipated supporting developers who would write applications for the platform, and both firms anticipated that developers would prefer to write applications for the browser with the most users, and users would go to the browser with the most functionality. Thus both firms anticipated fighting to win a majority of users and developers. Both firms also had a similar forecast for the near future, anticipating that the majority of new users of browsers would arrive somewhere in the future. That forecast translated into similar strategic stances—namely, both firms sought to gain an edge in 1995 in anticipation of a competitive future in 1996 and 1997 and beyond. That forward-looking attitude would persist throughout the browser wars. Accordingly, from the fall of 1995 onward, every confrontation between Netscape and Microsoft took on extra importance for management, even in episodes where the short-term stakes appeared to be small.
The similarities ended there. Their different business interests would shape their actions in visible ways. Microsoft profited from selling PC software and server software. In contrast, Netscape sold browsers and a range of complementary products to help enterprises take advantage of the commercial web. It had no other revenue, nor any other business to support, though it had aspirations to grow many more.
As prior chapters have discussed, Netscape had moved first. It priced its browser with the intent to build its base of users. Netscape had employed an imaginative and genuinely novel pricing strategy. Calling it “free but not free,” Netscape gave away its browser to home users, but sold commercial licenses to businesses. Throughout 1995 the free browsers achieved Netscape’s goals, displacing Mosaic browsers already in use, and it began building word-of-mouth support for more adoption, while the licensing brought revenue into Netscape.
Netscape also faced a major strategic challenge: it needed partners. For Netscape to succeed, in addition to attracting programmers, many other software firms would have to help Netscape’s platform. It also needed firms to aid in the adoption of browsers at businesses. Netscape intended to attract such firms by committing to the open standards of the Internet and web. Many of these firms already did business with Microsoft in one way or another, so Netscape had extra work to do. Attracting and supporting many business partners stretched the young firm from the outset and continued to stretch it over the next few years. Netscape had to invent its support processes, hire and train employees, and build operations from the ground up.
Microsoft faced a rather different set of strategic challenges. The late start framed many of these. Microsoft had to find a way to slow down the adoption of Netscape’s browsers by developers and new users, and thereby give Microsoft sufficient time to develop a browser that matched Netscape’s in functionality. In the short run that meant Microsoft had to find a way to convince developers that they should not exclusively cooperate with Netscape.
Throughout the fall of 1995 Bill Gates studied this situation and debated with his colleagues. From that debate came two fateful strategic choices: First, he chose a strategy that depended on his programming team’s ability. They would have to build a browser at an extraordinarily fast pace, while keeping up with any new functionality available at Netscape. Second, Gates chose to financially subsidize the effort, hoping to turn the development race into a contest that stretched Netscape’s resources, which played to Microsoft’s superior financial strength. In short, Microsoft funded programmer time and personnel far in excess of the direct revenue generated in the short run.
This was a familiar playbook at Microsoft. Microsoft had become the largest software firm in the world because it was so good at out-executing and, when it helped other parts of its business, out-spending others. As the remainder of the chapter describes, in fact, Microsoft stuck with these two grand strategic principles over the course of the browser wars, and it certainly mattered.5 This persistence also accounts for one of the biggest myths about the browser wars. Observing only these two actions, it is possible to conclude (incorrectly) that these actions—and only these actions—led Microsoft to win the browser wars. Relatedly, it is also possible to point to only these actions and conclude Microsoft did not deserve any legal troubles. That is a misleading inference based on incomplete observation. As described later, Microsoft’s additional actions got them into trouble.
On December 7, 1995, Microsoft held a major press conference announcing their Internet strategy. Prior to that announcement, the executives at Microsoft used the preparation for the news conference as a “forcing function” to finalize their strategy.6 The event received considerable notice in the national news. In industry circles each word from this event was parsed for deeper clues. One key announcement on that day got more attention than any other. Microsoft announced that it would not charge any price for its browser. It went further, announcing that future browsers would be integrated into its operating system as a feature, ceasing to be an application in Microsoft’s products.
Several things followed: First, an integrated browser could take advantage of the ubiquity of Microsoft’s operating system. As long as the browser was displayed prominently and supported, it would have to be distributed to users—almost automatically—along with the Windows operating system, which was used by close to 90 percent of the user base among PCs at that time. Second, although IE was not as good as Netscape’s browser at that point, a free and integrated browser would eventually motivate Netscape to give away its browser too, reducing Netscape’s revenue. Third, and less obvious, Internet Explorer was made the default browser unless users made efforts to change it to something else, such as Netscape’s browser. Finally, and perhaps most important, there was no longer any public doubt about Microsoft’s intent and commitment.
One other dimension to Microsoft’s actions was less appreciated at the time and became public only very slowly. Starting from the winter of 1996 Microsoft had begun to use an additional and broad tactic—namely, beginning to place pressure on every business partner—nearly every major firm in computing—to help it to push IE into widespread use and to prevent Netscape from gaining widespread use.7
Later it would be left to the court to decide why Microsoft had added this last tactic. In one view, Gates had not seen the progress he wanted in early 1996 and reacted in a panic. He had taken a risk and pushed beyond ethical limits. In another view, Gates was so focused on winning the browser wars that he lost sight of broader ethical and legal limits, and he simply misjudged where the line should be drawn by letting his self-interest interfere with his conversations with his legal team. A related third view stressed a feature of Gates’s personality in competitive situations: his stubborn unwillingness to concede points to others. In this view Gates misunderstood the legal ramifications of his own actions and misinterpreted the complaints as merely sour grapes over Netscape’s competitive misfortunes.
Not all of this was known at the time, but information about the strategy leaked out over time. For example, almost a year and half after the Internet strategy news conference, in July 1997, Jupiter Communications published a retrospective analysis of Microsoft’s Internet strategy, including analysis about the earliest moments in the browser wars. This article made public what others had inferred about Microsoft’s behavior: it was pursuing a strategy of “embrace and extend.” The strategy—allegedly based on an internal memo at Microsoft—was described succinctly:
Phase 1: Identify the market leader;
Phase 2: Emulate the market leader;
Phase 3: Steal the vision, provide a migration path;
Phase 4: Integrate, leverage, and erode.8
The quote reflected the popular view that Microsoft was focused on winning the browser war by matching any and all features found in its rival. It also reflected popular perceptions that Microsoft took a blunt and brazen approach.
Three Versions
Gates was a unique executive with an idiosyncratic approach to making strategic choices. His approach reflected a combination of technical smarts, personality traits, and acquired business sense. First, Gates liked technical topics for their own sake. For example, he taught himself to program and debug DEC’s computers when he was a teenager. Second, Gates drew on an almost unique mix of drive, persistence, and patience when pursuing his goals. For example, he furiously drove himself to learn things he believed he needed to know, and he negotiated incessantly until he won the points he wanted, while, in contrast, he could wait eons for results from strategic investments. Third, and least well known, Gates used his authority inside his company to teach and conduct lengthy investigations. He would convey a view, address questions and consider them in light of the available facts, and, in many cases, argue back and forth with employees. Every longtime Microsoft executive had experienced these pop quizzes, and, as a result, could talk intelligently about the firm’s strategy as well as every one of the Microsoft’s major competitors’ strategies.
All three predilections shaped Microsoft’s actions during a release of software. Gates arranged its product launches with the intent to learn from market experience. Many observers came to label this behavior a “three-version strategy.” In the first version Microsoft experimented with features and learned about market demand. In the second, they responded to feedback by improving its features. In in the third, they put all the lessons together, often to produce a hit product. Prior to the browser wars, Gates had personally coordinated the execution and updating of the strategy many times.
The three-version strategy was well known by the time the browser wars began. Most commentators assumed Microsoft would continue to follow the three-version strategy in an attempt to reach parity with Netscape. Since version 1.0 came from the University of Illinois, that meant parity had to be achieved by version 4.0. Knowing that others would expect this behavior, Gates anticipated that developers would look at the gaps between the functionality of Netscape’s and Microsoft’s browser after the releases of 2.0 and 3.0. Others would expect the gaps to decline with each release. That set up the situation: Netscape and Microsoft raced to add features that impressed users and developers. Each firm added more features that users valued, such as a range of bookmark tools and editing tools. Each added shortcuts and tools for developers, making their activities easier, and expanding the range of what they could offer to users.
Although a race can benefit users by adding features, some aspects of this race were not benign. Both firms tried to introduce proprietary features into their browsers, making it harder for vendors and users to switch to alternatives. For example, both distributed editing tools that generated web pages that included proprietary extensions to HTML that only their browser could read, and which their rival could not. Each of these was a step toward multiple proprietary flavors of HTML, one from Microsoft and one from Netscape, in addition to the version that the World Wide Web Consortium supplied.
HTML did not become proprietary, however. Tim Berners-Lee announced his opposition to such proprietary features, and he announced that the World Wide Web Consortium would never endorse such features. A large number of developers listened to Berners-Lee and refused to use these features. Chastened, neither firm tried this tactic again.
In the service of this race one other tactic also gained prominence: buy what could not be developed quickly. This played to Microsoft’s advantage in cash. For example, a start-up, Vermeer, had a well-developed authoring and editing tool, easy use, and was perfect for many new users. The founders decided in early 1996 that continuing as a stand-alone firm was less viable than merging with another. It made itself for sale, creating a bidding war between Microsoft and Netscape. Bidding eventually settled on $130 million. Netscape gave up, and Microsoft rebranded the software as FrontPage.9
The war did eventually unfold as forecast. Microsoft did catch up in three versions. IE 1.0 and 2.0 did not measure up to Netscape’s browser, but 3.0, released in August 1996, closed the gap considerably. For all intents and purposes, 4.0, released almost a year later in September 1997, closed the gap entirely.
Moving Developers
Microsoft’s strategy needed time. It needed time so the functionality of its browser could catch up to Netscape. It needed time so it could have a chance to gain new Internet users before they committed to another browser. This translated into a simple strategic imperative: Microsoft had to slow down user adoption of Netscape’s browser. How would it do that?10 In so many words, Microsoft could ask for the slowdown and buy it from business partners. Microsoft already did business with many industry participants, and many of them depended on maintaining good relationships with Microsoft. Nobody would refuse to take the call from Bill. Yet this course of action had an obvious drawback as well; many developers were not in a mood to do Microsoft any more favors.
Prior to the browser wars many developers respected and (begrudgingly) admired Gates for his achievements and quick mind. However, just as many despised him for being rich and ungenerous and self-centered and unwilling to suffer fools among his employees and business partners. He also was accused of being far too willing to imitate another firm’s good ideas without giving credit. Many also had walked away from negotiations with him with observations about his lack of social graces, his combativeness, and his unwillingness to yield a reasonable point to anyone else. Initially, many developers’ contempt for Gates helped Netscape, because Netscape recruited employees who wanted to develop browser technology and applications, and who wanted to see that technology bypass Microsoft’s products and APIs.11
Microsoft pressed ahead in 1996. Every important firm delivering Internet services received a phone call either from Bill or someone on his staff—all major ISPs, all PC manufacturers, all major Internet vendors, all major software firms, and all major consulting firms. Firms were offered deals to promote Microsoft’s browser while simultaneously making it difficult for Netscape to distribute its browser. Many reached a deal with Microsoft in 1996 or 1997. A few examples of what happened at companies doing computer assembly can illustrate how and why.
Consider Dell Computer. Dell had positioned itself as a low-cost provider of PCs, and it profited from high volumes of sales to business users. It needed the full cooperation of its business partners, particularly Intel and Microsoft, to produce a wide range of designs at large scale. It also had received preferential pricing and promotional deals from Intel and Microsoft—as good as or better than any other deal in the business. Dell needed such pricing to continue with its strategy of being a low-cost provider of PCs.
Dell cooperated early and readily with Microsoft’s demands. Without any fanfare, Dell stopped putting Netscape’s browser on its systems even when users asked for it. Not coincidently, Dell continued to get its favored status. This capitulation was well known and became the status quo quite early in the browser wars.
Such protocol generated an embarrassing moment for Michael Dell, in a March 1998, Senate hearing. Senator Orrin Hatch asked Michael Dell why his company promised to give customers whatever they wanted and yet did not offer a Netscape browser to customers. Dell offered an explanation that stressed his licensing agreements.12 Hatch shot back:
Could you explain why, when committee staff called Dell’s 1-800 number, five different sales representatives—George, Brad, Jason, Bobby and Jeff—told them that Dell could not distribute Netscape because of Dell’s licensing agreement with Microsoft?13
Michael Dell did not and could not provide a satisfying answer. His firm was doing well by cooperating with Microsoft.
Other firms that did not cooperate immediately suffered a painful route to cooperation. One firm, Compaq, having heard from many business customers who wanted Netscape browsers, featured it prominently. Unfortunately for Compaq, as a “reward” for listening to its consumers, Microsoft made an example of the company.
In 1996, for example, employees at Compaq removed the IE icon from shipped versions of computers. Their buyers were primarily businesses, and businesses had told them that they wanted the Netscape browser. Compaq removed IE in order to keep the icons less confusing and thus orient the computers toward the business needs of their buyers, while providing the applications users wanted. Microsoft believed Compaq had a business obligation to display IE, and it sent a letter to Compaq threatening to cut off its operating system license in sixty days if a removed IE icon was not put back on all new systems.14
Compaq capitulated quickly. At the time, it left everyone in the industry with the strong impression that Microsoft had chosen to make an example of Compaq. If nothing else, it also publicly demonstrated the drawbacks to being a business partner that did not play by Microsoft’s rules in all respects.15 Why did it leave that impression? As was frequently pointed out in public forums, the actions were public, but what lay behind them were not public (and it remained that way until the antitrust trial). This was but one of several alleged strong-arm tactics that most computer company executives refused to discuss with news organizations for fear of retaliation from Microsoft.16 Relatedly, no senior executive at Microsoft ever apologized, nor disavowed the action, nor did the firm ever give back any of the strategic gains it reaped from the action, which left the impression that the negotiating method was not an accident.
Despite the example Microsoft made of Compaq, the executive team at Microsoft decided to take further strict action with its business partners. Thereafter, Microsoft inserted clauses into operating system licenses that included but were not limited to restrictions on assemblers. Assemblers had to support an “out of the box” experience for users as they first fired up their systems, and it had to follow Microsoft’s script.
That sounds innocuous, but it was not because Microsoft wanted something that its business partners did not want to do. For example, one contracting provision prevented original equipment manufacturers (OEMs) from adding help screens for users of the Netscape browser. That is not a misprint—help screens for Netscape were forbidden, even though that reduced the quality of the user experience and drove up OEM service expenses (OEMs fielded the phone calls from irate buyers who could not operate a browser).
The restrictions generated strong feelings. Consider this statement from an executive at Hewlett-Packard, who wanted to include a help screen on its PCs to help a user figure out how to use Netscape:
We must have the ability to decide how our system is presented to our end users. If we had a choice of another supplier, based on your actions in this area, I assure you [that you (Microsoft)] would not be our supplier of choice. I strongly urge you to have your executives review your decisions and to change this unacceptable policy.17
By what right did Gates have the ability to tell OEMs not to help its users? His lawyers had inserted the clause and the OEMs could not refuse, lest they lose all their business.
Gates pressed everywhere, even at Apple. The newly returned CEO, Steve Jobs,18 cut a deal to make IE the default browser, which Jobs announced at MacWorld in Boston on August 6, 1997. Jobs received boos for this deal at the industry conference in which it was announced. But he preferred boos from his loyal buyers than the irritation of—and delays from—the provider of some of the Macintosh’s valuable software applications.19 Witness, once again, Redmond’s negotiating leverage. It never actually delayed the delivery of any software for the Macintosh. In conversations with Jobs, Gates merely had to allude to the possibility, even though development inside Microsoft (in truth) proceeded apace.
Microsoft also turned its attention to firms providing Internet access. While Microsoft successfully engineered to have many firms providing Internet access support its browser, there was one firm it never had been able to push around, America Online (AOL). AOL showed what a firm needed to bargain successfully with Microsoft—something Gates wanted.
As earlier chapters described, for many users AOL acted as both Internet access firm and portal, and through aggressive marketing had started to become the largest ISP in the United States. In 1996 and 1997 AOL focused their marketing efforts on new users to the Internet, where they believed they could add the greatest number of users. Executives at AOL did not trust Gates. They considered Microsoft a competitor for new users, and there was merit to the perception. Microsoft’s initiatives in MSN resembled AOL’s, were designed to compete directly with AOL for the same customers, and contained many features that imitated AOL’s. When Microsoft called to make a deal, AOL pushed back and negotiated hard.
AOL controlled the default browser settings for its users, who themselves tended not to switch these defaults. Although AOL users were not frequent Internet users, AOL had made itself compatible with the Internet. Its defaults could move significant amounts of “browser usage share” for a large fraction of online users. AOL argued that these defaults would not change unless Microsoft gave in on the thing AOL wanted—the right to have AOL’s branded icon on every desktop. The icon would link to directions that helped new users get online.20
Why was Microsoft in a position to trade position on the desktop? Once again, Bill has told Microsoft’s lawyers to insert a clause into contracts with OEMs about how the operating system had to look when it first booted up for a new user. Such was the power of Microsoft’s contract restrictions on assemblers. Microsoft could offer AOL the right to show up on a user’s system after it booted up, even when the assembly was being performed by IBM, Dell, HP, and others.
Repeatedly Gates refused to make this deal with AOL, even though his browser division wanted it. He worried that his employees at MSN would view it as a broken promise to protect and promote MSN by giving it favorable placement on the desktop. Accordingly, AOL refused in return to make Internet Explorer the default browser for their users.
The standoff broke when AOL threatened to make a deal with Netscape in March 1996. Only when the Netscape deal neared completion did Microsoft budge.21 Accordingly, AOL received the best deal of all. Along with hundreds of millions of dollars, it got the key item that Steve Case, CEO of AOL, had held out for. Case got a button on the desktop. No other ISP ever successfully managed to again get such a placement. In return, Microsoft moved enormous market share. Although it took a while to implement, by January 1998, over 90 percent of AOL’s users—by then, the largest access provider in the United States—were using Internet Explorer.22
In the midst of these debates in Microsoft, Gates revealed that he fully understood these trade-offs. He wrote:
We have had three options for how to use the “Windows Box”: First, we can use it for the browser battle, recognizing that our core assets are at risk. Second, we could monetize the box, and sell the real estate to the highest bidder. Or third, we could use the box to sell and promote internal content assets. I recognize that, by choosing to do the first, we have leveled the playing field and reduced our opportunities for competitive advantage with MSN.23
Circling back to the main theme, these aggressive tactics and obsession with beating Netscape yielded high short-term costs—in terms of employee time, managerial attention, and reputation costs—and far exceeded the short-term benefits in terms of revenue. While many contemporary news outlets highlighted the drama of the confrontation, these actions could not be rationalized without recognizing the trade-off between short-term costs and Bill Gates’s perceptions about long-term benefits for Microsoft. Gates authorized such an approach because he wanted users and developers not to work with anyone else other than Microsoft, regardless of what users or developers wanted.
More broadly, these examples illustrate how Gates made the browser wars a top priority, and expended costs far above any direct revenue affiliated with Internet-related businesses. Microsoft effectively pushed its browser out to all kinds of PCs, not just new versions of Windows, and it prevented Netscape’s browser from becoming the default in widespread distribution. Microsoft made it difficult for Netscape to distribute its browser to the typical mass-market user, thereby slowing down Netscape’s adoption.
Netscape had had to put up with something no mass-market application software firm ever had faced. Much of the adversity Netscape faced distributing its product had nothing to do with the products’ merits and functionality. It had to do with reluctant business partners, and partially or fully closed distribution channels, none of which would have been reluctant or closed had Gates not worked so hard to achieve that outcome.
From High-Flying Entrant to Collapse
Netscape aspired to become the primary commercial hub for developing browser applications and the primary window for the user experience of the growing World Wide Web. Yet as Gates had hoped, the competition continued throughout 1996 and 1997, and it stretched Netscape’s resources and eroded its competitive position. Despite the stellar résumés and high-profile founders, any young firm taking on this role would face enormous challenges scaling their operations and executing its goals. Like any young firm, Netscape lacked routines grounded in years of experience. Flaws were bound to surface as they grew, and they did. Complaints began to surface about Netscape’s overwhelmed support staff, and lack of returned phone calls from many employees.24
Marc Andreessen, Netscape’s CTO and media-appointed wunderkind, did not make the situation any better by talking aggressively in public. His business inexperience showed. He satisfied every mid-twenty-year-old’s desire to gain an adrenaline rush from bold quips quoted in major news outlets, but in doing so made the strategic situation worse. At one point, for example, he predicted that Windows would be reduced to a “poorly debugged set of device drivers.” On other occasions he referred to Microsoft as the “beast from Redmond,” “Godzilla,” and “the Evil Empire.” His quips were very entertaining for readers who despised Microsoft, and were therefore widely reported and repeatedly quoted.25 To Microsoft employees, they came across as disrespectful, condescending, and combative. Although doubters were increasingly few in number, Andreessen’s quotes also confirmed Bill’s theory that Netscape wanted to bring down Microsoft.26
Most of these types of flaws are forgiven in young firms, and in a normal situation an inexperienced organization merely irons out their rough edges over time. In due time sober venture capitalists and experienced board members know how to quietly inch talented but mouthy founders out of their firms and bring in reliable managerial talent, or patiently teach an inexperienced executive to behave better. Netscape, however, faced a situation that did not permit room for forgiveness. Netscape found itself in a race to develop features with one of the world’s most efficient developers of packaged software, and in a setting where Netscape faced a growing distributional disadvantage.
By the middle of 1997 the three-version strategy began to play out as Gates had hoped. Contemporaries began to predict that plenty of potential users would still be available and amenable to adopting IE 4.0. By mid-1998, plenty of households and businesses were just starting to use the Internet. To many of these new users IE 4.0 was just fine for their needs, and many tried it first because it was the default browser in the new PC they bought or the AOL account they used, and, as well, in plenty of other places.
The increasing realization that Microsoft had caught Netscape sucked the enthusiasm out of many developers who had committed time and energy to the Netscape browser. As Netscape’s revenues declined in visible ways and as public support waned, Microsoft’s success seemed inevitable to many observers. Finally, the Netscape coalition simply collapsed, with hundreds of developers each month switching exclusively to Microsoft and ceasing to add to and maintain their code for Netscape’s browser.
In a short time, Netscape’s value declined precipitously. Its market value effectively became reduced to the traffic attracted by its home page—which was still considerable by the norms of start-ups, but small by the heights the firm had previously obtained. The decline happened so quickly and irreversibly it was shocking to behold. Only three years earlier Netscape’s IPO had symbolically given birth to the commercial Internet. Now it was a “destination site,” kept alive by the inertial habits of many users who had made its website their default start page for surfing.
More indignities for Netscape followed. Seeing an opportunity to move traffic on Netscape’s web portal to include different content and a different advertising network, AOL bought Netscape in November 1998, with the company exchanging at a value of more than four billion in overinflated stock.27 The deal finished the following spring, effectively ending the existence of an independent browser firm, and nobody knew how long it would be before another arose.
It took a few years. By 2002 it was said that well over 95 percent of the browsers in use were from Microsoft. The code for alternative browsers remained moribund, and reemerged later as the Firefox browser.28 It was as if the progeny could not resist the fight; in the middle of the next decade Firefox became a competitor to IE 6.0, the dominant descendant of IE 4.0.
Legal Problems29
Antitrust law includes many arcane details, but at a broad level it is straightforward. Antitrust law makes it illegal for a dominant firm to use its negotiating leverage to gain an advantage when it competes for a new market. Implementing that law accomplishes two goals at once: first, it encourages dominant firms to compete only with new or improved products and services; second, it discourages a dominant firm from using distributional restrictions on rivals or other advantages obtained through negotiating leverage with existing business partners. Discouraging firms from investing in the second encourages them to invest in the first, and that benefits everyone.
Here, Microsoft was at risk. Gates’s priorities for browsers were no secret, and many of Microsoft’s actions had received attention in the news. Many participants in PC and ISP markets suspected that Gates had talked to and put pressure on everyone. A few fights with business partners, such as the fight with Compaq, had made it into public knowledge. It was reasonable to wonder whether Gates was using Microsoft’s negotiating leverage with a variety of market participants to avoid innovating without using that leverage.
In popular conversation two issues became intermingled. First, had Microsoft stepped over the line morally? Second, had they violated a legal definition, one founded on antitrust law? These questions were animated by the brazen way Microsoft had competed in the browser wars, fueling the sense that fouls hung in the air. Confusion characterized popular discussion.
Microsoft’s bolting of IE to the operating system became a major source of confusion. It seemed to be technically unnecessary. The competing application could and did deliver similar functional performance without any bolting whatsoever. This action was raised frequently in popular conversation to illustrate Microsoft’s unfair advantages.
Although brazen, this specific action was not a legal foul. It did users a favor by making the operating system better. For good reasons, no court or lawmaker ever has made it illegal to enhance a product while charging users nothing for the enhancement.
If there was a foul with IE, it lay in something simpler about its design. For example, Microsoft made it virtually impossible for users or suppliers to remove IE if they wanted to do. This was an unnecessarily restrictive design and a transparently self-serving departure from prior practice. Indeed, its triviality illustrates the point: no detail was too small. Microsoft would go the extra mile to limit user options and choice if Microsoft foresaw a benefit for itself.
Another illustration comes from the way Microsoft treated Java, a piece of middleware. Specifically, SUN Microsystems invented a computer language for networking, called Java, and wanted to do some experimentation with developers and users. Hyped well beyond any realistic possibility, Java attracted considerable attention from developers. Netscape was interested too, because Java could potentially improve the workings of their browser. Many developers attempted to write code for Java and test out its capability in applications that users valued. Unlike Netscape, SUN was an established firm with an experienced division for supporting developers, so developers expected to get the best of both worlds—frontier software and first-rate technical support.
Gates thought of Java’s promise to developers to deliver services on the desktop in much the same way he thought of Netscape’s browser. If circumstances were auspicious either one could facilitate entry of applications that substituted for applications in the existing value chain. In short, it could threaten the revenues of his core product. Gates did not want that bud to blossom, and he took deliberate actions to make it more challenging for developers to experiment with Java in ways SUN desired, even when such experiments had a strong chance of benefiting users.
After difficult negotiations Gates signed a contract with SUN, as if Microsoft intended to make Windows compatible with SUN’s preferred version of Java. However, not long after the ink was dry, Microsoft took actions designed to confuse users and developers about what was possible with the Windows-compatible Java, inviting a public fight with SUN over some features Microsoft implemented, making it unclear what version of Java worked with Windows. That slowed down everyone’s experiments.
Later, in the antitrust trial, there was a reductionist legal debate about details. Did Gates negotiate with SUN in bad faith? Lawyers for Microsoft also stated that it had the right to defend its business, and it was under no obligation to hurt itself.
Both arguments missed the key point—namely, that Microsoft’s tactics seemed to cross a line about using technical complementarity to slow the distribution of software users valued. Gates’s actions sowed fear, uncertainty, and doubt about whether Microsoft’s and SUN’s versions would work together. That did not help vendors and users experiment with Java. Every established firm, aspiring entrepreneur, and venture capitalist in the country watched these events and got the message. No experiment would be allowed to win or die on the technical merits if Gates didn’t like where the experimentation might lead. If Gates could do something about it, he would. The promise of originality, technical potential, or user benefits did not matter. Only Gates’s self-serving judgment did.
Such behavior violated common understanding of how the technical meritocracy operated in the computer market. As discussed in earlier chapters, the technical meritocracy was quite a familiar concept for many participants in the Internet, and part of that familiarity was inherited from the computer industry and part of it came from the experience in the precommercial Internet. In technical meritocracies every entrant had to have their chance to reach users, whether that supplier was an insider or outsider, established or entrepreneurial, big mouth or unknown. In markets the technical meritocracy played out in competition. Products and services were supposed to compete against one another on the merits. The market had to give users access to new products and services, and young firms had to have time to grow and experiment.
The US meritocracy did not grow and operate on its own, and this was the sticking point for Gates. Gates did not believe his leading firm had an obligation to nurture the technical meritocracy, to allow it to thrive, not if it came at the expense of his firm. It only had an obligation to be self-serving, and that trumped all else, including the meritocracy. Adding to that, even his firm’s tactics used his market distributional strength as a tool, shaping the technical meritocracy to his preferences. As long as it stayed within the letter of contracting law, Gates asked, what could be wrong with that?
It would be imposing retrospective bias to characterize the actions of government antitrust prosecutors as if they knew the issues before investigating the actions. At first the legal case developed slowly. Government lawyers did not just come out and accuse Microsoft of committing anticompetitive fouls by restricting user choice and limiting distribution, or making it harder for rivals to reach consumers. Rather, in 1996 and throughout 1997, government lawyers raised questions about whether Microsoft was abiding by a settlement that the firm and the government had reached in 1994 over antitrust issues in Microsoft’s contracts with assemblers.30 The answers were unsatisfactory, and that kept the investigation alive. One case led to another.
The legal confrontation became unavoidable after a day in court in which Microsoft behaved as if it believed the government lawyers had no basis for raising additional legal issues. This defiance dared the government lawyers to investigate further and fight back, and, after considerable deliberation, they did. By the time Netscape’s coalition began to collapse, the lawyers at the antitrust division in the Department of Justice believed they had enough material to bring a federal antitrust case. On May 18, 1998, Janet Reno, the US attorney general, announced that the Department of Justice was filing suit.
Microsoft’s lawyers indicated that they believed they did not think they had done anything wrong, and they spoke as if they anticipated winning any courtroom confrontation. Popular opinion speculated that Gates welcomed this confrontation.
The Trial31
The trial began in May 1998, and would last for many months, with the later appeals dragging into the next millennium. It involved many eye-opening revelations and spectacles. The first occurred at Gates’s deposition, which had been recorded prior to the trial. Depositions are normally perfunctory, boring legal events. This was anything but that. Under the cold light of hard questions, the lead prosecutor, David Boies, reduced Gates to human proportions.
Boies made Gates sound unprepared for skeptical questions, as well as uncooperative and internally inconsistent. Gates denied his firm had market power, and he denied recalling e-mails that showed that he fully comprehended his negotiating advantages with business partners. Boies made it appear that Gates lived in his own self-serving world, effectively deploying market power against anything he regarded as a threat, no matter how far off in the future the threat lay, no matter what anyone else wanted, including his own users.32
During the actual trial, things became even worse for Microsoft. One subpoenaed e-mail after another revealed coarse and selfish language. The substance also did not help, as the e-mails exposed the defensive motives that characterized Microsoft’s internal debates.
The prosecution eventually achieved its principal legal goal. It persuaded the presiding judge, Thomas Penfield Jackson, that management had both motive and ability to diminish the competitive process. As for motive, it possessed defensive intentions, and these had motivated the firm to take multiple actions to alter the competitive process in its favor and not compete on the merits. As for ability, it also possessed many levers to influence other firms in computing, and Gates had not hesitated to pull those levers.
Jackson’s two rulings strongly favored the prosecution’s view of events. The “Court’s Finding of Facts,” dated November 5, 1999, provided a coherent narrative of the many ways in which Microsoft had used its existing monopoly in operating systems to restrict the distribution of Netscape’s products. The judge then encouraged the government and Microsoft to negotiate. They did not reach resolution after many months of secret negotiations. Judge Jackson then issued his second judgment on April 3, 2000. In it Microsoft was found guilty of multiple antitrust violations. Before any hearings before Judge Jackson on remedies—the DOJ proposed breaking Microsoft into multiple firms—the case was appealed, as expected.33
All the publicity led to tons of collateral damage to Microsoft’s and Gates’s public image. News commentators wondered if the company’s internal culture had any sense of restraint. Many prior business partners wondered whether their deals had been enacted in good faith. By association, it tainted many legitimate activities at Microsoft, reducing the value of years of brand building.
Microsoft’s legal defense team next did what any good law team would—it attempted to reverse the judgment before an appellate court. The lawyers accomplished many of their goals, winning on some points and losing on others. More importantly, the court agreed to place some limits on the range of punishment Microsoft could receive.34
Yet by losing just a little, Gates lost a lot. The appellate court did not conclude that Microsoft’s actions fell outside the domain of antitrust law. It did uphold the finding that Microsoft possessed market power and a monopoly. It did not exonerate Microsoft from accusations that it used its market power to restrict the distribution of products from rivals. It did find that established firms did not have free rein, even with underdeveloped products. The court said:
It would be inimical to the purpose of the Sherman Act to allow monopolists free reign [sic] to squash nascent, albeit unproven, competitors at will—particularly in industries marked by rapid technological advance and frequent paradigm shifts.
This ruling had enormous consequences for the tone and texture of industry competition. Had Gates won all points at the trial or on appeal, then Gates would have been free to continue to authorize Microsoft to behave as it had. In that case, other powerful executives in computing would have been tempted to adopt Gates’s strategic playbook for limiting distribution and subverting the experimentation of others.
As for whether the trial undid the fouls in this particular event, however, ambiguity carried the day. Although Microsoft did not escape punishment altogether, it did avoid severe punishment. That occurred through a sequence of highly unlikely events.
Specifically, Judge Jackson had been all too aware of the attention the trial received in the news, and, showing poor judgment, gave long and frank interviews to reporters during the trial.35 Although he embargoed those interviews while the trial continued, he allowed them to be published after it was over. The interviews made for entertaining reading, particularly because Jackson thought Gates and his colleagues needed to be taught a lesson, and he used colorful language to describe why. Judge Jackson compared Bill Gates to Napoleon Bonaparte for displaying arrogance from unalloyed success. He also remarked on the lack of maturity among the executives at Microsoft and compared their declarations of innocence and lack of repentance to what he had heard in a prior trial involving gangland killings.36
These quotes were published prior to the appeals court’s final judgment, as the appeals court proceedings were ongoing. That violated longstanding practices at the court. In its June 2001, decision (which is another great read), the appellate court severely reprimanded Jackson for breaching protocol.37 The reprimand showed some examples of quotes the appellate court found inappropriate. In addition to the remarks about Napoleon and gangsters, this long story from Judge Jackson to reporters came under severe rebuke:
He told the same reporters that “with what looks like Microsoft intransigence, a breakup is inevitable.” … The Judge recited a “North Carolina mule trainer” story to explain his change in thinking from “[i]f it ain’t broken, don’t try to fix it” and “I just don’t think that [restructuring the company] is something I want to try to do on my own” to ordering Microsoft broken in two:
He had a trained mule who could do all kinds of wonderful tricks. One day somebody asked him: “How do you do it? How do you train the mule to do all these amazing things?” “Well,” he answered, “I’ll show you.” He took a 2-by-4 and whopped him upside the head. The mule was reeling and fell to his knees, and the trainer said: “You just have to get his attention.” The Judge added: “I hope I’ve got Microsoft’s attention.”38
The appeals court also found some of the judge’s reasoning less than adequate. It sent the open legal questions, including the determination of the eventual punishment, back to another judge.
This outcome was quite fortunate for Gates. No other judge had sat through Gates’s deposition, nor had any other judge watched one Microsoft witness after another wilt under David Boies’s questioning.
Another unexpected event then played a role. Although the Bush administration had no appetite to fight Microsoft’s case further, initially it could not act on those predilections without inviting very bad publicity, as well as a revolt from the states’ attorney generals who had signed on. However, the horror of the 9/11 terrorist attacks on the World Trade Center in New York City provided motivation and/or cover to redirect the priorities of the Justice Department. The new lead prosecutor negotiated a settlement with minimal bite, which Microsoft readily agreed to, and after some tussle, got the new judge to approve it.
The new judgment did one thing well: it appointed a technical committee and set up a process for review. That made it far more difficult for Microsoft to claim legality for questionable tactics and defer a trial, something Gates had done in the past. Now a court or a court-appointed committee heard all complaints quickly. Even so, since the market had already moved on, and the Netscape coalition had effectively disbanded, there was not much to adjudicate. Ambiguity ruled the day.
Ambiguity also trailed Gates’s career. He had become singularly associated with the questionable behavior at the center of the firm’s legal problems. After the final court judgment, Gates gave up his CEO position and moved to a newly created position, chief software architect. Steve Ballmer became CEO. The reasons were never fully explained in public, but indications were that—after the finger pointing stopped—even Gates’s friends in Microsoft thought this was a way for the firm to move forward, at least on a symbolic level. Yet Gates continued to be present and contribute to decisions at every strategic level, so a skeptic might reasonably wonder if the move accomplished much in practice.
The case left the news after 2001—except for periodic stories involving Ballmer settling every private antitrust case Microsoft faced within the United States. The private suits in the United States cost over $4 billion, a small fraction of Microsoft’s cash on hand. Later, Microsoft faced another set of issues from the European Union, which increased the final bill and hassles, and the concerns of that administrative process seemed to continue almost indefinitely.
Finally, fulfilling the dystopian predictions of Microsoft’s biggest critics, the case also faded from view due to Microsoft’s actions. Without a competitive reason to continue to invest in the Internet, Gates showed he did not care much about it. Microsoft went back to its older strategic priorities and slowed or eliminated investments in the Internet. Accordingly, Microsoft failed to execute an Internet strategy that took advantage of its dominant position in browsers. Once so prominent, Microsoft’s aspirations in the Internet faded from public view, largely unfulfilled.39
Resisting Innovation from the Edges
Microsoft’s experience illustrates an inherent tension with self-governance of commercial markets undergoing creative destruction. Self-governance relies on restraint by those who possess the incentives and ability to alter the competitive conditions in their own favor. While some of the largest firms stand to gain from the successful blossoming of commercial markets spurred on by innovation from the edges, others can anticipate the potential losses, just as Microsoft did. As Microsoft’s actions illustrate, such established firms tend not to sit by passively, and their actions do not necessarily support the competitive processes underlying entrepreneurial capitalism.
This example also illustrates a related issue: the losses to society from self-governance are not felt in events that occur, and, instead, are felt in events that never occur but could have. To understand why, consider whether Microsoft’s decisions altered the economic impact of innovation from the edges.40 Specifically, think about what would have happened had the web become a competitor to Microsoft instead of merely new functionality to sell more PCs, as it might have, had Netscape had access to less clogged distribution. In short, compare what actually did happen to what might have happened without fouls.
As a baseline, first recognize what did actually happen in a broad summary. Competition between platforms—one organized by Microsoft and one organized by Netscape—began petering out in late 1997, when Microsoft’s fouls drew developers away from Netscape’s browser and SUN’s Java. Competition then collapsed in 1998 when developers abandoned Netscape’s platform altogether. After this collapse, Microsoft slowed several of its Internet initiatives. After 2001 IE went years without another upgrade, and IE 6.0 became heavily criticized on developer forums. Developers were stuck, and many complained, but they had nowhere else to go.
For the sake of speculation, consider the possibility that the lack of aggressive tactics might have allowed competition to continue for an additional, yet modest, amount of time. Let that be just a couple more years, say, to the end of 1999 or middle of 2000, while the Internet continued to diffuse to new users, or until the dot-com bubble burst.
More competition would have yielded considerable gains. Users would have experienced what would have been invented had developers had a cooperative business partner, such as Netscape, for a little while longer, and a competitive rival, such as Microsoft, trying to keep pace. Since the late 1990s was an era of easy money and the biggest venture-capitalist-led entry boom in the history of computing, there would have been additional experimentation in the form of new entrepreneurial entrants. For example, an additional eighteen months would have generated at least two—perhaps three—more rounds of competitive browser upgrades and complementary inventions in security, search, identity protection, transactional automation, and other applications.
Users might have experienced what Microsoft would have done in a features race, had it possessed a sense of competitive urgency after 1998. More competitive urgency would have generated more innovative features out of Microsoft, especially in comparison to what its browser division did after competition ended.
Innovation from the edges did not blossom as much as it could have. At a crucial time some entrepreneurial purveyors of the technology were not able to experiment, or able to take advantage of unrestricted distribution channels. The lack of experimentation made the commercial Internet a different place than it could have been and prevented society from benefiting from the gains innovation from the edges could have yielded.
The broad lesson is sobering and applies to today’s market as much as it did to these events in the 1990s. Creative destruction does not necessarily emerge on its own, nor automatically. Market events need not be either creative or destructive if the leading core firms act on their incentives, and use their abilities to control events for defensive purposes. Said another way, the process of creative destruction emerges only after the failure of an established firm’s attempts at defensiveness. It is not enough that established firms need to perceive the threat as a real one. If they can merely block it with actions that alter distribution for a rival, then blocking can reduce incentive to respond by innovating.
This example also illustrates the broad lesson about what open platforms do better than proprietary platforms coming from established firms. Established firms will hesitate to cannibalize their own leading products and will tend to follow their own planning, trying to make greater returns on their existing sunk investments. In contrast, creative destruction led by entrepreneurial entry that appeals to users will pay no heed to an established firm’s existing profitability, or to its planning, or to the vision or investments underlying it. Said another way, platforms led by entrepreneurs can be especially effective at creating and supporting new value for users when they become unexpectedly destructive in the eyes of an established firm.
1 Quoted in Opinion for the United States Court of Appeals, 2001.
2 By 2006 that percentage decreased to 10 percent when Gates stopped working in the firm day to day.
3 As earlier chapters described, Ben Slivka wrote “The Web Is the Next Platform.” The fourth and final draft is available at http://www.usdoj.gov/atr/cases/ms_exhibits.htm, exhibit 21. Gates’s memo is exhibit 20.
4 See Cusumano and Yoffie (1998).
5 Notice that the logic also implies the opposite, which did come to pass. After the war ended Microsoft also reduced the speed of development and level of funding, illustrating that the fast execution was motivated by strategic considerations in a competitive race.
6 A “forcing function” is an event that leads management to settle open matters, and sometimes the event is under the control of the management and sometimes not. See Cusumano and Yoffie (1998) for a lengthy description of this event, and an argument that Microsoft’s management deliberately scheduled the news conference to make it serve as a forcing function.
7 The most succinct description of the allegations is in Tim Bresnahan’s (2003) summary of the DOJ case and Henderson’s (2001) declaration for the remedies portion of the trial. These stress control of information, control of interfaces, tying, bribing potential middleware suppliers, using control over original equipment manufacturers, and bribing and/or coercing other third parties. The chapter is necessarily a simplification.
8 Jupiter Communications (1997).
9 See Ferguson (1999). Despite selling the firm to Microsoft, the sale did not stop Ferguson from devoting the last chapter of his book to an analysis of the competitive problems Microsoft’s actions posed.
10 This description is necessarily a shortened version of events. For a full recounting of the sequence of actions, see the court’s Findings of Fact, especially pages 40–177. http://www.justice.gov/atr/cases/f3800/msjudgex.htm, accessed November 2012.
11 See Cusumano and Yoffie (1998).
12 As quoted in McWilliams (1998), Michael Dell said,
We negotiate vigorously with all of our vendors to obtain the best possible terms that we can for our own company and our customers and our shareholders. And based on the terms that we have negotiated … the incentives that have been put in place for us to sell one product or another are going to dictate our actions.
This answer is both true, and not, related to explaining why Michael Dell promised customers anything they wanted but acted differently than promised, maintaining an exclusive relationship with Microsoft, even when customers directly asked for Netscape.
13 Quoted in Chandrasekaran (1998).
14 Hardwick (1996). Or see the court’s Findings of Fact, pages 99–101, http://www.justice.gov/atr/cases/f3800/msjudgex.htm, accessed November 2012.
15 The dispute was (apparently) settled through a few phone calls, but only after the threatening letter had been sent, which makes one wonder how much of the public discussion was simply putting lipstick on a pig. Thereafter, the Netscape and IE icons appeared on both desktops for a short period, but Compaq renegotiated its contracts with others. See McCullagh (1999).
16 See Chandrasekaran (1999).
17 After the introduction of these restrictions Hewlett-Packard sent a letter to Microsoft with the strongly worded lines. See also, United States v. Microsoft Corp., 84 F. Supp. 2d 9 (D.D.C. 1999), 210–15.
18 He was the interim CEO at the time of this announcement.
19 The deal also settled several patent disputes, gave Apple cash in exchange for stock, and committed to make Microsoft Office available for five years. After the announcement about the default browser, Jobs said to applause, “Since we believe in choice, we’re going to be shipping other Internet browsers, as well, on the Macintosh, and the user can, of course, change their default should they choose to.”
20 Microsoft also made certain cash payments to AOL. See page 71, and pages 135–48, court’s Findings of Fact, http://www.justice.gov/atr/cases/f3800/msjudgex.htm, accessed November 2012.
21 Page 142, court’s Findings of Fact, http://www.justice.gov/atr/cases/f3800/msjudgex.htm, accessed November 2012.
22 Page 147, court’s Findings of Fact, http://www.justice.gov/atr/cases/f3800/msjudgex.htm, accessed November 2012.
23 Quoted from page 141 in court’s Findings of Fact, http://www.justice.gov/atr/cases/f3800/msjudgex.htm, accessed November 2012.
24 This is documented in some detail in the account of Cusumano and Yoffie (1998), as well as in the court’s Findings of Fact, http://www.justice.gov/atr/cases/f3800/msjudgex.htm, accessed November 2012.
25 See Chapman (2004), 189, for a short compilation of many of these quotes.
26 This noisiness contrasts with a “stealth” strategy, in which a young firm publicly denies having aspirations to compete with an important firm, and denies have aspirations to become as important as the leading firms in its industry, in the hope of not generating defensive investments, so that the young firm can later surprise them in a more vulnerable position.
27 Junnarkar and Clark (1999). Also see pages 148–49, court’s Findings of Fact, http://www.justice.gov/atr/cases/f3800/msjudgex.htm, accessed November 2012.
28 Near the end of the browser wars, Netscape had declared its code “open source.” That invited suggestions from outsiders, though these did not arrive nearly fast enough to improve the browser dramatically. It also left the code in a general purpose license. After AOL lost interest in improving the browser, another group of ex-Netscape programmers and others coalesced around improving the code and starting an open source browser, which later became Firefox.
29 This is necessarily a summary of a longer set of arguments. This section draws on a number of perspectives. See, e.g., Eisenach and Lenard (1999), Liebowitz and Margolis (2000), Evans et al. (2000), Evans (2002), Bresnahan (2003), Rubinfeld (2004), and Page and Lopatka (2007), not to mention the court’s documents, as already cited.
30 A brief summary of the events leading to case can be found in Rubinfeld (2004).
31 For a narrative focused on the trial and its public face, see Auletta (2001), or Brinkley and Lohr (2000).
32 See, e.g., the account in Chapman (2004), 178–86.
33 See the website maintained by the Department of Justice, http://www.justice.gov/atr/cases/ms_index.htm, accessed November 2012.
34 The appeals court decision came down on June 28, 2001. See http://www.justice.gov/atr/cases/f257900/257941.htm, accessed November 2012.
35 Two books resulted from those interviews, e.g., Brinkley and Lohr (2000), and Auletta (2001). These interviews are dealt with on pages 106–20 of the appeals court decision, and were treated quite harshly for giving the appearance of partiality. Opinion for the United States Court of Appeals, for the District of Columbia Circuit, June 28, 2001, http://www.justice.gov/atr/cases/f257900/257941.htm, accessed November 2012.
36 See e.g., Brinkley and Lohr (2000) and Auletta (2001).
37 Pages 106–20, Opinion for the United States Court of Appeals, for the District of Columbia Circuit, June 28, 2001, http://www.justice.gov/atr/cases/f257900/257941.htm, accessed November 2012.
38 The court is quoting from Brinkley and Lohr (2000), and others. See page 112, Opinion for the United States Court of Appeals, for the District of Columbia Circuit, June 28, 2001, http://www.justice.gov/atr/cases/f257900/257941.htm, accessed November 2012.
39 For more on why and how this occurred, see the summary offered by Bresnahan, Greenstein, and Henderson (2012), as well as the chronicling of the debates inside Microsoft offered by Bank (2001).
40 One of the more hotly debate points from the trial focused on the losses from Microsoft’s actions. The questions were never settled because the remedy phase of the trial never was conducted. See, e.g., Eisenach and Lenard (1999), Liebowitz and Margolis (2000), Evans et al. (2000), Evans (2002), Bresnahan (2003), Rubinfeld (2004), and Page and Lopatka (2007).