- Shopping Bag ( 0 items )
Harvard Business ReviewBeating Microsoft at Its Own Game
Vermeer's FrontPage won the first stage of the race to establish a proprietary Internet software standard. Yet after having sold out to Microsoft, its founder now says antitrust action is the only way to keep the giant honest. What about open-source software?
by J. Bradford DeLong and A. Michael Froomkin J. Bradford DeLong is an economics professor at the University of California at Berkeley and a research associate at the National Bureau of Economic Research. A. Michael Froomkin is a law professor at the University of Miami in Florida and a nonexecutive director of the Internet start-up Out2.com.
High Stakes, No Prisoners: A Winner's Tale of Greed and Glory in the Internet Wars
Charles H. Ferguson
Times Books, 1999
In 1993, Charles Ferguson, an MIT-trained political scientist and high-tech industry consultant, had the brilliant idea of making a software tool for building online information systems. The tool, later dubbed FrontPage, was designed to be user-friendly so the nonprogrammers who were supplying on-line content could use it but sophisticated enough to provide flexibility in how to structure and present that information. Ferguson sunk his life savings into his new company, Vermeer Technologies, assembled an excellent programming team with his partner, Randy Forgaard, and raised venture capital to keep the company going. He pursued the enterprise with a mix of monomania and paranoia. When it was released in late 1995, FrontPage was far more advanced than similar products from any competitors—including Microsoft, which bought Vermeer soon after for $130 million.
Ferguson and his company pioneered an important way to interact with the Internet. Because of Ferguson's idea, because he backed it to the hilt, and because others saw what he was doing and imitated him, the public got easy-to-use authoring tools for Web sites a good deal earlier than it would have otherwise. We're all a little bit richer—and because Ferguson managed to carve out a small slice of this new wealth for himself and his colleagues, they're now a good deal richer.
As part of the deal for FrontPage, Microsoft bought Ferguson's silence for two years. That period is over, and Ferguson is now offering society a different kind of product: a detailed, extraordinarily honest account of his experiences in the world of high-tech start-ups. Ferguson names names and launches a great deal of criticism—much of it aimed at himself. But readers who get past the dirt will be treated to a multifaceted case study of how to compete in the software industry. Ferguson's narrative misses the mark only when he argues that Microsoft's strategy of gaining proprietary control over software standards is the only game in town. There is a different approach—nonproprietary, "open source" software—that's also worth considering. An Entrepreneur's Many Hats The story of Vermeer is actually the first of three major threads that run through this book. The second is an account of the rise of the Internet and the consequent dilemmas of high-technology corporate strategies, written by Ferguson in his role as a business analyst. The third is a related assessment of Microsoft and its increasingly dangerous power in the computer industry, written by Ferguson the policy analyst—who feels some guilt at having profited from and added to Microsoft's leverage.
Other topics, ranging from privacy law to productivity theory, are also raised, briefly discussed, and then dropped. One of the more interesting points Ferguson makes is that the private sector could not have generated the foundations of today's high technology; the research could have come only from the government and the universities it supported. This is a book rich in asides.
None of the major threads in High Stakes, No Prisoners is the best treatment in its class. For an entertaining and informative entrepreneur's-eye view of business in the computer age, try Andy Grove's Only the Paranoid Survive, which offers a more nuanced account of the high-tech industry. The pseudonymous Robert X. Cringely's Accidental Empires does a better job of placing the developments that created today's computer industry in their full and proper perspective. If you want a briefing on the dilemmas of high-tech corporate strategies, Carl Shapiro and Hal Varian's Information Rules: Competitive Strategy in the Information Age provides a better analysis of the features that make competition in high tech unique and interesting. And you can get Ferguson's point of view on Microsoft in much more detail by reading James Wallace's Hard Drive or David Yoffie and Michael Cusumano's Competing on Internet Time.
To a large degree, Ferguson is limited by space; he tries to cram three books into one. There is also a problem with narration. Ferguson sees the world in black and white; there are no shades of gray, no tolerance of ambiguity, no sense that things look different from different points of view. To his great credit, Ferguson recognizes, as few analysts are ever adult enough to do, that he has been very certain about some issues in the past—yet also very wrong. But this recognition of past errors doesn't help him to see uncertainty and ambiguity in the present. In his view, people are geniuses or idiots, corporate strategies are brilliant or misguided, organizations are healthy or fatally diseased.
What makes the book worth reading is Ferguson's multiplicity of vision; he brings the perspectives of both an aggressive entrepreneur and a reflective analyst. In the early 1990s, Ferguson the industry analyst discovered the advantages of the Silicon Valley start-up system, and he had the gumption to dive in with his own company after his big consulting clients passed on the idea. As Ferguson the entrepreneur launched Vermeer, he was powerfully influenced by Ferguson the analyst's views about how to create an entrenched market position in the information age. Meanwhile, Ferguson the entrepreneur's experiences in the start-up world educated Ferguson the analyst in the finer points of gaining capital and breaking into an industry. Those experiences also help Ferguson, as a government policy analyst, discuss the challenge from Microsoft.
Industry analysts and consultants usually have little understanding about how difficult it is to run a business. And policy analysts usually don't know how the private sector really works and what government interventions in the market will actually do. Ferguson the entrepreneur has a solid grasp of those issues. Entrepreneurs usually have a hard time articulating their intuitive vision of industry dynamics. Even when successful in business, they can rarely explain what led them to follow their strategies for creating a position in the market. Ferguson can and does. The combination gives the book a wealth of perspective and a streetwise credibility that few books about the Internet economy possess. For example, when Ferguson tells us that government R&D people are "light-years ahead" of their counterparts in industry in both brainpower and vision, his judgment cannot be dismissed as the ramblings of an academic who has never taken a product to market. Putting Your Money Where Your Mouth Is Ferguson prefaces his account of Vermeer by describing how, in 1992, he and a fellow consultant, Charles Morris, came to the conclusion that "the key prize in high technology is proprietary control of an industry standard."1 When IBM, for example, decided to simply license technology from Microsoft and Intel for its new PCs, it unwittingly gave those companies the chance to establish their own particular technologies as the basic features of this high-growth product. In short order, IBM became one of many manufacturers of low-margin Wintel clones, while the two standard setters made huge profits.
When Ferguson came up with his idea for an authoring tool the following year, he didn't forget that insight. As he and Forgaard assembled their programming team, he didn't tell the team just to design an excellent software program for building and maintaining Web sites. He and his colleagues also made sure to write the code so it would set standards for the emerging Web site industry. He saw that path as the only way to make a lot of money for the company—in itself an extremely tough task because of the nature of software markets.
In most software categories, there can be only one reliably profitable product: the one with the largest market share. Software production involves enormous economies of scale, both on the production end (software costs almost nothing to duplicate) and on the demand end (if your colleagues use Microsoft Word, you're likely to do so, too). As a result, having even the second largest market share may not be enough for a company to recoup the cost of writing the software.
If the market leader falls short in innovation, however, it will see its position in future product generations supplanted by a more innovative rival. But by setting the standards for a product category, the leader can raise the costs for users who want to try a rival's product instead. That's because a software product category often relies on or spins off new categories that are complementary to the original product. Ferguson knew that in order to become an industry standard, Vermeer's authoring tool would have to interact with software that handles the flow of on-line information over big Web-server computers. Server software vendors hadn't yet settled on a standard for the interface between servers and Web site designers. If Vermeer was successful enough, it could induce the next generation of server software to accept some Vermeer standards. That would make it hard for any rival authoring software to succeed, because it would have to unseat not just Vermeer but the next generation of server software. Assuming Vermeer did a reasonable job in updating its tool, users wouldn't be able to replace Vermeer in the future.
With that kind of pressure, it is no wonder that the drive to become the standard becomes all consuming for vendors. Software makers like Vermeer become obsessed with speed to market, secrecy (to prevent competitors from recognizing what their product will be), and initial market share. The need for dominant market share also encourages software bloat. As Ferguson perceived it, "You can't become the industry standard unless your product covers every major portion of the market." He and Forgaard even delayed FrontPage's release and made the software code more complicated so it would offer additional Web functions and would work better with the different server software available.
When developers win acceptance for their standards, they can lock users into their product. And that's what Vermeer's code was poised to do. By contrast, Netscape didn't appreciate the need to lock in—or even identify—customers for its enormously popular Internet browser program, says Ferguson in a particularly acerbic section of the book. As a result, Microsoft, as it has done in several other product areas, eventually displaced Netscape's browser by offering better features in future product generations.
FrontPage's code was so impressive that Vermeer hardly needed to sell any software before the big competitors came calling. Even a company like Microsoft, willing to commit substantial resources to win market share, may find a well-established proprietary standard difficult to dislodge. Rather than assign a team of programmers to invent a tool to compete with FrontPage, Microsoft found it expedient to pay a huge premium to acquire Vermeer directly.
Vermeer essentially beat Microsoft at its own game of lock-in. And without that aggressive software code, Ferguson admits, Vermeer wouldn't have stood a chance. The start-up lacked even a minimally adequate distribution operation and sold only a few hundred copies of FrontPage before the acquisition. As Ferguson colorfully describes, the pressures and paranoia of the situation were so great that he was barely on speaking terms with his venture capitalists and his recently hired chief executive officer. Ways to Contain Microsoft The story of Vermeer is interesting and entertaining, but would-be entrepreneurs should take note: being well informed about start-up financing and other hurdles may not help. The venture capitalists still have the money, and you still need it. As Ferguson explains, an entrepreneur has a hard time getting venture capitalists to compete against one another because the VCs know they'll need each other's help in other, bigger deals. This coziness has been tempered slightly by the recent growth of new funds seeking to bankroll start-ups. But the basic power imbalance will remain as long as the number of would-be start-ups grows at least as quickly.
Even Ferguson's confident analysis of proprietary lock-in, which also informs his analysis of the high-tech industry and Microsoft, may have its limits. It's true that FrontPage has held up well since it was released in 1995. Vermeer's special software for interacting with servers has indeed made it expensive for users to switch to rivals. It also appears to be difficult to install the server-side features of FrontPage on any machine that is not running Microsoft's Internet Information Server, so FrontPage users may be locked into other Microsoft programs as well.
Indeed, that lock-in, coupled with Microsoft's monopolies in key software platforms, makes Ferguson the policy analyst worry that the company will extend its dominance over desktops into the software markets of the Internet world. In conjunction with the financial and programming resources it can throw at any new software market, Microsoft can use its control over related products to drive competitors out of business. That's what it did when it administered "cashectomies" to Borland's flagship database program and to Netscape's browser. By including a database application in its Office suite and a browser in Windows, Microsoft essentially offered these products for free and drastically reduced its rivals' cash flows. And where Microsoft does confront an established standard, it can either license the standard—and then add advanced, proprietary features to win control—or buy the standard outright, as it did with Vermeer. As a result, Ferguson advises the U.S. government to break up Microsoft into two competing companies: one for operating systems and one for application software.
Yet perhaps Ferguson's insight into proprietary architecture is outliving its usefulness. Perhaps what worked—gloriously, for Ferguson—in the 1990s will not work ten years from now. In reading High Stakes, No Prisoners, we were struck by Ferguson's near neglect of a serious threat to Microsoft's dominion. As we look at Microsoft today, its principal competitive restraint—in the operating system business at least—isn't Sun or Apple (which draw more of Ferguson's attention), but open-source software, particularly the "freeware" operating system Linux. Ferguson's discussion of open source is limited to one page, in which he concludes that Linux is unlikely to have more than a small niche. He cannot see how open-source software can ever "attract the levels of investment and infrastructure required to mount a serious challenge" to Microsoft's Windows.
That Ferguson discounts the open-source movement is remarkable given Vermeer's own debt to it. Ferguson's original idea for an on-line tool included, in addition to the authoring tool, the software to transfer and interface with information sent over phone lines. He soon discovered that the open-source World Wide Web, which had been recently developed by Tim Berners-Lee at Geneva's CERN research center, already provided the underlying code for transferring information. And the open-source Mosaic program, developed by students at the National Supercomputing Center in Illinois, provided a basic interface. Netscape soon emerged to displace Mosaic, but Berners-Lee's Web standards have been able to resist all proprietary encroachment—including Microsoft's launch of the proprietary Microsoft Network. As CEO Bill Gates finally recognized in 1995, Microsoft had no answer to information distributors and users who asked why they should use MSN instead of just setting up their own Web servers on the Internet. While Ferguson recognizes Berners-Lee's great achievement for humanity, it may be telling that he says he found Berners-Lee "unrealistic" when he first met him—and that Ferguson slams the World Wide Web consortium that Berners-Lee joined as a "rather useless, nonprofit Web standards group."
As Ferguson argues, the volunteers who develop Linux are unlikely to make big investments in a product that gives them no financial return—they need to keep their day jobs. But the open-source movement has interesting properties that suggest to us it is worth taking seriously. The movement has evolved a number of practices that at least ameliorate and may even overcome the problem of inadequate incentives.
One is the now familiar way in which open-source authors gain status and participate in a sort of on-line gift exchange. As an idealistic movement, open source benefits from the input of dedicated, even obsessed, volunteer programmers who are willing to cooperate to keep big companies from controlling the computing environment.
Ferguson's own experiences suggest another, more important, advantage to freeware: free open-source software can lure even for-profit software developers into helping out. When Vermeer was developing FrontPage, it had to make major decisions about the interface with server software. Open-source Apache software was (and still is) a major competitor in the unsettled market to supply the big servers that handle Web transfers. Eager to make money, Ferguson and his developers had no interest in simply donating their authoring software to the existing Apache program—they wanted proprietary control. But they also did not want to go to the trouble of developing their own full-fledged server software to go with FrontPage. Their solution ended up meeting both goals: By developing extension software that made Apache (and other major server software) compatible with FrontPage, they took a big leap into establishing proprietary standards. But Vermeer also made Apache competitive in the next stage of Web development, when ordinary users would make their own Web sites.
If Apache had not been open source—if Netscape or Microsoft had owned the server market—Vermeer would have had to think hard about adding its own server to FrontPage. And FrontPage would have had a harder time establishing its own standards in authoring software; it would have been susceptible to a cashectomy. By relying on open-source software to handle the server end, Vermeer limited its upside potential but protected its flank.
From the user's perspective, open source provides at worst a mixed blessing. It's true that proprietary software packages tend to adopt new features faster. But most software users apparently take advantage of only a small subset of new features, and the hastily added bloat makes proprietary software prone to crashes. Freeware, by contrast, tends to be stable. And for-profit developers can still step in with proprietary add-ons for specialized users who need those features.
Indeed, one of the great strengths of the open-source movement, albeit one not always endorsed by its hard-core devotees, is the extent to which it creates opportunities for market-driven complementary products and services. Such products are often much more difficult to provide for proprietary software. Red Hat, for example, can make a successful business by providing easy-to-install Linux software and technical support in a way that is all but inconceivable for a supplier of help with Windows.
And open-source software offers major advantages in security. In proprietary software markets, manufacturers have a hard time making sure that only those users with valid licenses are running programs. Sophisticated antipiracy systems have emerged, but they often go haywire and cause enough hassles so that legitimate users have all but vanished. Since user respect for copyright tends to decline sharply after the first year or so of an application's existence, for-profit developers make a point of releasing frequent minor upgrades that can be more trouble than they're worth to install. Because freeware is free, developers can release upgrades at a more respectful pace.
Open-source developers can also work in a more comfortable environment, without the secrecy and paranoia that plague for-profit ventures like Vermeer. And because no one can control open-source standards, the complementary providers are assured access to the underlying software. The point of open source is to provide a common (and low-cost) platform for competing products to bloom. If the industry can harness these energies for a determined assault on Microsoft, the giant might be cut down to size without the help of antitrust action. The traits that made Charles Ferguson a successful entrepreneur both strengthen and weaken this book. Ferguson the analyst, by providing a framework for organizing the narrative (and his company's strategy), gives the founder of Vermeer a much more intelligible voice than entrepreneurs usually possess. He also produces clear, well-written prose. But Ferguson's analytical mind leads him to be overconfident of everything. His intolerance for ambiguity, and his refusal to admit uncertainty beforehand and in retrospect made Ferguson an aggressive and decisive entrepreneur. Those characteristics also make his book fun to read. But they make readers wary about just how strongly they should trust his conclusions. Does he really want us to believe, for example, that in 1995 no one in the entire Justice Department understood high tech and Microsoft?
Ferguson concedes that his track record as a consultant and policy analyst has been uneven. He has proved farsighted on some issues, but also thought at one time that Silicon Valley start-ups were no match for established companies, that Japan posed a grave threat to America's high-tech industries, and that IBM's hiring of Lou Gerstner was a disaster. Will his next book, in a decade or so, begin with an admission that he failed to see that open source had achieved a state of maturity that vastly undermined the value of proprietary control of industry standards?
1. Charles H. Ferguson and Charles R. Morris, "How Architecture Wins Technology Wars," HBR, March-April 1993. Reprint R00102