Ask Jack Dorsey, the co-founder of the social network Twitter and the mobile-payment start-up Square, what his two companies have in common, and he has a quick answer: “They’re both utilities.” Mark Zuckerberg might agree: he spent years trying to convince people that Facebook is not a social network but a “social utility.”
It’s an intriguing choice of words for such of-the-moment entrepreneurs. Utilities tend to be boring, slow-growing beasts. They also—and this is the more important point—tend to be monopolies that are either regulated heavily by governments or owned outright by them.
Indeed, once they get beyond a certain size, technology companies do become wary of the word. Google has been called a utility by lots of people, but you won’t hear the company’s executives using the term (at least, I couldn’t find any examples). And Zuckerberg, when asked in 2010 whether, as a utility, Facebook ought to be regulated, said he hadn’t meant the word that way at all: “Something that’s cool can fade. But something that’s useful won’t. That’s what I meant by utility.”
Yet there are lots of useful things in the world—clothing, breakfast, this issue of The Atlantic—that no one would ever think of calling a utility. Yes, there is an innocuous class of computer software known as utilities. But what companies like Twitter, Square, and Facebook—not to mention Google, Amazon, and Apple—aspire to, and in some cases have achieved, is a status similar to that of traditional utilities like Ma Bell. They attempt to position themselves such that customers can’t get around them, or can’t afford to leave them. And when they succeed, they start appearing to some customers, would-be competitors, and regulators like scary monopolies that somebody needs to do something about.
The connection between attractive business opportunity and monopoly is not new. Pursuing a “short run” monopoly, the economist Joseph Schumpeter wrote in 1942, is what profit-seeking enterprises do—in the process, driving significant innovation and economic growth. In the 1970s, the business-school discipline of strategy arose as the study of how to build and defend these short-run monopolies—a sort of mirror image of the antitrust classes long found in law schools. “Strategy is antitrust with a minus sign in front of it,” says the Columbia Law School professor Tim Wu, who has taught both subjects. That is, strategy tries to maximize what antitrust tries to minimize.
What is new is that the path from looking for an edge to being attacked as a monopoly has gotten a lot shorter—and that gaining a monopoly seems such a plausible goal within some of the fastest-growing parts of the economy. Standard Oil had been in business for 36 years when the Justice Department sued it for antitrust violations; AT&T for 97. By comparison, Microsoft was just 15 when federal regulators started looking into its business practices, 23 when Justice sued. Google, a mere 14 years old, is already under antitrust investigation.
Today’s technology entrepreneurs are well aware of the tight link between profit and monopoly. Few are as open about it as the PayPal co-founder and early Facebook investor Peter Thiel, who has described monopoly as the natural goal of any smart tech entrepreneur. But everybody gets the basic idea. “There’s a joke in Silicon Valley,” says the UC Berkeley economist Carl Shapiro: “ ‘You know you’ve really made it when you’ve got antitrust problems.’ That’s the sign of success.”
The modern theory of monopoly began its rise in the mid-1980s, when a handful of scholars—Shapiro among them—noted some salient characteristics of a fast-growing new industry. Many information-technology businesses, observed Stanford’s W. Brian Arthur, benefit from increasing returns: as they make more of something, the cost per piece keeps falling. This is especially true of software, for which the cost per piece moves quickly to zero. (Increasing returns had been deemed in the late 19th century to be the mark of a natural monopoly, an industry that would inevitably be dominated by one entity.)
Another trait that characterized many technology businesses, these same scholars observed, was lock-in, or prohibitive switching costs. Companies that committed to getting their mainframe computers from, say, IBM would eventually find switching to another provider hugely expensive and disruptive. (Later, with the PC, Microsoft was able to shift the lock-in from hardware to software.)
But most intriguing of all was the enormous power of network effects. A telephone “without a connection at the other end of the line … is one of the most useless things in the world,” AT&T President Theodore N. Vail wrote in the company’s annual report in 1908. “Its value depends on the connection with the other telephone—and increases with the number of connections.” In 1980, Bob Metcalfe, an inventor trying to persuade people to buy his $5,000 Ethernet cards, which connected computers in a local area network, came up with a formula that expressed the value of a network as the number of connections squared. The specifics of “Metcalfe’s Law” have frequently been challenged, but the basic idea that networks add value exponentially as they grow has not.
For society as a whole, though, these phenomena can have a dark side. In a famous paper, the Stanford economic historian Paul David described in 1985 how the ubiquitous QWERTY keyboard layout had been devised mainly to prevent jamming of primitive typewriter mechanisms. Later, as typewriters improved, there were repeated attempts to supplant QWERTY with configurations that allowed for faster typing. But by then the layout’s high switching costs had made it an impregnable standard. Economic forces, wrote David, “drove the industry prematurely into standardization on the wrong system.”
Brian Arthur borrowed the term path dependence from physics to describe this predicament, and wove it into an alarming story about how technology standards develop and get locked in even when there may be better options. The saga of Sony’s Betamax, which purportedly delivered better picture quality but lost out to VHS as the dominant videocassette, was told and retold. Research by other economists soon muddied the tales of both VHS and QWERTY—they may not have been such obviously inferior technologies after all. But the behavior of Microsoft, which by the early 1990s was using its market power to shove aside innovative competitors like Apple, WordPerfect, and Lotus, certainly seemed like an example of technology heading down the wrong path.
In May 1998, professing to be heavily influenced by Arthur’s work, Assistant Attorney General Joel Klein sued Microsoft. The central offense involved Netscape, the maker of a new kind of software called a browser that was used to navigate a new communications network called the Internet. As it had done before, Microsoft built a copycat product, Internet Explorer, and used its operating-system dominance to wrest control of the market.
The trial that ensued was such a headline-grabbing sensation that it is remarkable how few people even remember the outcome (I didn’t until I looked it up). Microsoft lost the case decisively in U.S. district court, where Judge Thomas Penfield Jackson ordered that it be split in two. That wasn’t the end, though. Jackson had openly criticized Microsoft in interviews, and the court of appeals blasted his behavior as “rampant disregard” for judicial ethics, vacated parts of Jackson’s ruling, and handed the case over to a different judge. Microsoft was still found to have violated the law, but the remedy fell far short of a breakup or even serious punishment.
There were reasonable economic arguments for leaving the company be. In its 2001 ruling vacating Jackson’s decision, the appeals court described the quandary at the heart of the case. Citing the literature on network effects, the judges acknowledged that, in some tech markets, “once a product or standard achieves wide acceptance, it becomes more or less entrenched.” That made software companies a bit like utilities—where competition is not “within the field” but “for the field.” On the other hand, citing Schumpeter at some length, the judges noted that, in technologically dynamic markets, “such entrenchment may be temporary, because innovation may alter the field altogether.”
For the next few years, Schumpeter looked like a pretty good guide. Microsoft continued to mint money from Windows and Office, but it missed out on almost every new digital opportunity. Apple rose from its deathbed on the strength of better products and a new way of selling music online.
Then there was the Internet, which seemed mainly to reward upstarts that built good products. Google, the first of these upstarts to break through in the new millennium, hardly looked like a trust worth busting. By and large, it didn’t make use of network effects. It also didn’t lock users in. Yes, Google enjoyed increasing returns to scale. But if another company were to come along with a dramatically better search technology (just as Google once did), it might well take over the market.
Other businesses that arose in Google’s wake, however, did focus on creating network effects. Facebook’s business is built entirely around them—its value for users resides in the presence of other users, period. As you build up an online identity on Facebook, the costs of switching to another network rise, too. Other, less universally ambitious businesses, from Yelp to Evernote to Airbnb, similarly have tried to establish themselves at the center of unique new ecosystems. And in recent years—with Gmail, Google Places, the Chrome browser, the Android smartphone operating system, and other products—Google has been trying increasingly hard to build a network that customers will find tough to abandon.
So what, if anything, should we do about all these budding monopolists? The prevailing attitude among regulators over the past decade has been: very little. One reason is that firms don’t always know how to transform the network effects they’ve been creating into empire-entrenching profits. Microsoft’s approach—charging for its software—is not open to most of today’s firms, at least not at first. To build a network big enough to exploit, you usually need to offer something for free. So we get lots of “freemium” businesses that charge only heavy users. We get Apple, which has learned to build networks around music and movies and apps, but makes most of its money selling the devices that access those networks. We get Amazon, which is building similar networks around books and other media, and is selling its devices for no profit. And we get Google and Facebook, which have such staggering numbers of users that they are able, by smart targeting, to make serious money selling ads—although so far they’ve had less success with ads on mobile devices than on computers. With the situation so dynamic and uncertain, we may be tempted just to shrug and let the market take us where it will.
But to assume that this hyperfast Schumpeterian capitalism is the new normal, and that government or civic action can only get in the way, is to ignore or misread history. (If you need to do some remedial reading, as I did, I recommend Tim Wu’s The Master Switch and Richard R. John’s Network Nation.) We have repeatedly seen new communications technologies go from open and chaotic to closed and, in many cases, censored. This happened with newspapers, radio, TV, movies, even the telephone system. Just because the Internet has succeeded in blowing up some of these established communications monopolies and oligopolies doesn’t mean it won’t create its own.
Declaring that the new utilities of the Internet age thus require regulation can be problematic. As Wu puts it, “Regulation is often a way to keep a dominant firm in place.” At this stage, a set of rules on how search engines are supposed to work would probably just entrench Google. But other tools are available. We can, for example, agree on principles of how the online world should develop. The best-known of these is “network neutrality”—the idea, coined by Wu, that broadband providers shouldn’t be allowed to favor some Internet businesses over others. Still incipient is the principle, espoused by the open-source advocate Doc Searls and others, that customer data should belong to customers rather than to retailers, social networks, search engines, and the like. Third-party providers should store your information, and share it only when you want them to. This would lower your switching costs, and could upend the business models of Google, Facebook, and others that profit from hoarding user data. It might also deliver a better consumer experience at a lower cost.
And let’s not forget antitrust enforcement. It is now fashionable to dismiss it as pointless—look what happened with Microsoft. But look more closely, and it’s possible to spin a credible tale of antitrust lawyers enabling disruption and innovation. The Justice Department may have given up its 13-year struggle with IBM in 1982, but pressure from Washington helped push the company to unbundle its software from its hardware, speeding the development of an independent software industry. The 1980s breakup of AT&T created fertile ground for the growth of the Internet. The spectacular rise of Google and Facebook, and the resurgence of Apple, were possible at least in part because Microsoft didn’t feel free to strangle them as it had Netscape a decade before.
The virtual world is so new that it still seems boundless, and its continued exploration and colonization are providing new tools and new entertainments for us all, along with great fortunes for the explorers and colonizers. Yet, if we are inattentive, today’s landgrabs could crowd out opportunity, innovation, and even economic growth in the future. It’s one thing to acknowledge that the way ahead is uncertain, and that what looks like a permanent monopoly today might be roadkill tomorrow. It is quite another to assume that everything will work out for the best as long as market forces are allowed to work their magic.
So all praise to today’s would-be utilities and monopolies, as they try to build enterprises that own their markets and that we can’t do without. But when they actually succeed, don’t think we shouldn’t be sniffing around in their business. At a certain point, it becomes our business, too.
This article available online at: