1.6 METCALFE'S USEFUL EQUATION

    Moore's Law goes a long way toward explaining why the digital age is increasingly populated by killer apps. What it doesn't tell you is why these applications seem to spread as quickly as they do. To understand that, you need Metcalfe's Law. Consider the telephone. How useful is it? Your answer depends entirely on how many other telephones there are and on how easily they can be interconnected. One phone is useless, a few phones have limited value. A million phones create a vast network, and a network is something, as the Communist governments of the former Soviet Union learned to their dismay, with tremendous power.

    Robert Metcalfe, founder of 3Com Corporation and the designer of the robust Ethernet protocol for computer networks, observed that new technologies are valuable only if many people use them. Specifically, the usefulness, or utility, of a network equals the square of the number of users, a function known as Metcalfe's Law. (See Figure 1.2) The more people who use your software, your network, your standard, your game, or your book, the more valuable it becomes, and the more new users it will attract, increasing both its utility and the speed of its adoption by still more users. If you and I can call only each other, to return to the telephone example, a phone is of little value. But if we can call nearly everyone else in the world, it becomes irresistible.

    The diagram of Metcalfe's Law in Figure 1.2 shows a magical point of inflection, the knee of the curve, at which a technology reaches critical mass. After that point, its value increases exponentially. How quickly a new application hits the knee depends on how much it costs new users to get access to the network (for instance, a telephone and monthly connection charges), since buyers will weigh this cost against the usefulness of the technology at the time of purchase. The lower the initial price, the more quickly critical mass is reached. And ironically, once critical mass is reached, the developer can in theory charge subsequent users more, because the network effect increases the application's value.  

    For the phone system, or the power system, the initial investment in network infrastructure was high, which kept the price of access high. In the case of railroads and telephones, initial developers failed to appreciate the value of interconnection (in essence, the power of the Metcalfe curve). Railroads struggled with multiple gauges of track, which limited connections between systems, until the late 1880s. It didn't even occur to telephone companies to put a dial on the phone until 1931, even though the high cost of employing people as switchboard operators limited the reach of the network. In the predigital age, Metcalfe's Law could take decades to unleash network power.

    Developers of today's digital technology are conscious of Metcalfe's Law, and they are developing counterintuitive rules necessary to optimize and exploit it. The most dramatic demonstration of Metcalfe's Law during the digital age has been the explosion in the early 1990s of the Internet, a network of computers and a set of standards that makes it easy for computers to share data. The Internet had existed in various forms for many years, but reached critical mass in 1993. (See Figure 1.3) From there, true to form, the Internet became the "it" technology, attracting not only users but billions of investment dollars, regular cover stories in popular magazines, and even Hollywood movie makers.

    Computer hardware, software, and networking companies had been building up their user bases for decades with closed, proprietary networking standards like IBM's Systems Network Architecture and its PC token ring network, document interchange, and hundreds of subsidiary "IBM solutions"-solutions, that is, for IBM. The Internet, on the other hand, has always been based on open, public standards, allowing it to grow faster despite its lack of a marketing function or, indeed, any organization whatsoever. The Internet became the dominant global computing network it now is by being the first to reach the knee in Metcalfe's curve, and the impact of that victory will be played out in the information technology industry for years to come.

    The notion that an open system-a system that gladly gives all its secrets away-could humble a giant like IBM seems the stuff of fairy tales. Meet Marc Andreessen, who was an undergraduate at the University of Illinois in the early 1990s when the World Wide Web, a new set of open standards for sending and receiving multimedia communications over the Internet, was in its infancy. In an effort to exploit these standards Andreessen wrote Mosaic, a program that allows users to browse through the various Web sites that were being created. To get maximum exposure for Mosaic, Andreessen lowered the access cost-that is, the price of the software-to zero. Even when former Silicon Graphics founder James Clark stole Andreessen and his team away to start Netscape Corporation, the operating model didn't change. Netscape Navigator, the company's rapidly evolving browser, is still being given away.

    The result? Netscape captured 80 percent of the browser market within months of its first product release in 1995 by giving away millions of copies of its software. Unlike telephones, giving away Navigator costs little in real dollars. Thanks to the Internet itself, users simply download the software, using their own phone connection, their own machines, and their own electricity. The marginal cost of each "copy" of Navigator that Netscape has given away is not effectively zero, it is actually zero.

    Eighty percent of a market for a free product doesn't sound like much of an achievement, but when Netscape stock debuted in 1995, it went from an initial asking price of $14 a share to $150 in a matter of days, giving the company a market valuation of more than $3 billion dollars. The rapid proliferation of Navigator encouraged faster development of Web sites, which in turn led to greater demand for Navigator. The company now derives revenues from subsidiary goods and services, such as advertisements on its own Web site and information broadcasting services, software tools for building corporate "intranets" (an intranet is a network similar to the Internet that operates only within a single organization's defined boundaries), and products that allow developers to build and manage the Web sites that Netscape's "free" users are visiting. Software giant Microsoft arrived late to the market for browser software and the related services it makes possible, but used its own market power (illegally, according to the U.S. Justice Department) to cut into Netscape's dominance in 1997. Still, Netscape reported more than $500 million in revenue for that year.

    Navigator hit the Metcalfe curve at warp speed, with critical mass and the predictable explosion occurring not in years but in months. In 1994, when we first demonstrated Mosaic to clients, the World Wide Web was an experiment. By 1997, nearly every network television advertisement included a Web address. What's more, entire industries had sprung up and matured by devoting themselves to creating supplemental software, including audio, video, and three-dimensional modeling tools, as well as services for developing, hosting, and managing Web sites for corporations. The Internet itself is their low-cost channel for advertising, product development, manufacturing, and distribution.