by Tim Wu
It was not for want of effort or resources. Microsoft opened an entire new complex in Redmond, California, at some distance from its main offices, which it staffed with “content people” specially imported from New York and Los Angeles. The team was comprised of “black-clad creative types,” reported The New York Times, tasked with “building a new, chic kind of media business.” In reality, the new hires were given an impossible task: that of inventing both a new platform and an entirely new form of content—one that was not television, film, or a computer game but instead “interactive” and that capitalized on “convergence” via a new portal named “MSN 2.0.” That was the name of the interface being preinstalled on every new Windows machine and designed to make the Internet feel more like a television.4
Reviews of MSN 2.0—which had launched with the slogan “Every new universe begins with a big bang”—were unkind. The Times complained of complicated instructions that began “with a pushy multimedia commercial whose irritating 20-something host declares, ‘I see you have good taste!’ He also says, ‘I promise you, this won’t hurt a bit.’ Dentists say that, too.” The effort left the user with a feeble browser that could download and play a series of interactive shows that somewhat resembled video games but were much slower to load, and lacking any action to speak of. Microsoft later admitted the shows were “spectacularly unsuccessful.”5
If the Internet was still too new to achieve its content potential, perhaps Microsoft could for now hasten the convergence by making television more like the Internet. Following this logic they sank hundreds of millions into a new cable news network named Microsoft-NBC, or MSNBC for short. They had the vague notion that the network would eventually populate the Internet in some way. As Microsoft explained in a press release, “From now on, the promise of the Internet and the power of television become one.” Commentators weren’t entirely sure what to make of it. “Some very interesting and powerful creature will be created,” said media critic Steve Rosenbaum, “but I don’t know what it is.” If he didn’t know, it turned out that no one at Microsoft or NBC quite did either. When it finally launched, MSNBC elected to make itself a copy of CNN, with one exception: it featured a show named The Site, devoted to the Internet revolution.6
The synergy, whatever it was supposed to be, failed to materialize, and facing low ratings MSNBC executives turned to old-school cable thinking. Mining demographics, they noticed that left-leaning viewers lacked their own cable news network. MSNBC would therefore become the mirror image of Fox News; the “MS” in MSNBC would linger there, like an old store sign that the new owner forgot to take down. Microsoft’s only successful investments in this period were Slate, which billed itself as the very first online magazine and ultimately proved the viability of the concept, and the Xbox gaming console, which was the kind of nuts-and-bolts market entry that Microsoft had always done well. Most of the rest was a fiasco of the first order.7
In short, Microsoft couldn’t quite crack the nut; the idea of merging television and the Internet was just too simplistic. In effect, Microsoft treated the Internet as if it were merely a new channel on which it could broadcast content, as if what had been invented was an extension of cable television. But the impulse was not altogether misguided, for someone was going to make money selling all of the attention the Internet was capturing. It just wasn’t going to be Microsoft. The most successful contestant would be the one that truly understood the new platform and how it would be used.
By the late 1990s, as Microsoft’s content initiative faded, word was spreading about a new company named Google, whose specialty was search. Search became a major application as the Internet got more populated, fast becoming wild and woolly, too vast to tune in to like a TV channel. There were a number of search engines running—Lycos, Magellan, AltaVista, Excite, Yahoo!, which originally launched as an Internet directory, to name a few. But it was soon clear that Google did search better than anyone else; its inventors were smart, its algorithm was wicked, and its code was tight. With a simple search box placed on a white page, it was both elegant and technically a quantum leap over what was then available. But all who followed the company knew it had one massive and potentially fatal problem. Google was more like an academic project than a firm, and as it gained in popularity, it was burning through cash at a terrifying rate, with nary a business model in sight. As cofounder Sergey Brin reflected later, “I felt like a schmuck. I had an Internet start-up—so did everyone else. It was unprofitable, like everyone else’s and how hard is that?”8
It is a fascinating moment to contemplate. Tech’s hottest firm stood at a fork in the road, pondering its future and, being an unusually deliberative company, tried to think through the consequences of each choice. Thanks to the growth of search, Google was capturing bushels of attention, but how to capitalize on that? As we’ve seen, there have always been two ways of converting attention into cash. The older is to charge admission, as to the spectacle, on the model of the opera or HBO or a book; for Google this would mean licensing or selling access to a premium product. The second is to resell the attention gained, that being the model of the attention merchant.*1
It didn’t take a genius to see that the safest and most immediate revenue stream for a firm with Google’s potential to harvest attention was in advertising. And that way it went. “Every day, millions upon millions of people lean forward into their computer screens,” John Battelle, cofounder of Wired, would write in 2005, “and pour their wants, fears, and intentions into the simple colors and brilliant white background of Google.com.”9 If radio had promised a “relaxed” audience, here was something better: people who really wanted something and ready to tell exactly what—“intentional traffic,” in the jargon. Google could serve up minds to advertisers in this particularly valuable mental state—open, desirous, and impressionable. Thanks to the attention merchant’s proprietary alchemy, advertising would make Google feel “free” to its users—as if it were just doing them a big favor.
It retrospect, it was a no-brainer. But the firm’s principal founder, Larry Page, was unusually sensitive to the corrupting potential of advertising. Before Google had come to its fork in the road, he had strenuously insisted in a piece coauthored with Brin that “advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers.”10 As an engineer and scientist, Page had wanted to build a clean, pure tool, free of commerce’s distortions. His aversion, moreover, was not purely academic: he had seen what an advertising-driven search engine looked like, peddled by a man whom both Brin and Page found distasteful—a character, now mainly forgotten, who was for a while Google’s greatest rival. His name was Bill Gross.
Two years earlier, in 1998, Bill Gross was at the then newish “Technology, Entertainment, Design” conference in Monterey, California, peddling his ideas for an advertising-driven search engine. Launched in the early 1990s, the TED conference staged short speeches by innovators and visionaries, with the intention of inspiring audiences; the flogging of products was strongly discouraged. Gross, who made a career out of ignoring such admonitions, went to TED to hawk his latest product. He proposed a new concept for Internet search: instead of some fancy algorithm, he said, why not just sell places in search results to the highest bidder?
While Gross’s presentation was engaging, the response was not friendly. TED is famous for its easy and adoring audiences, yet on this occasion hisses were heard. Afterward, in the hallway, the buzz was downright negative; as John Battelle wrote, the model “was in clear violation of every ethical boundary known to media.”11
But Bill Gross was not the type to be deterred. Issuing from a rough-and-tumble 1990s dot-com culture, he was famous for turning one idea (supposedly inspired by Steven Spielberg) into a company nearly overnight at his start-up incubator Idealab. He was given to saying things like “The old rules don’t apply when you’re an Internet company.”12 In retrospect what Gross meant by the “old rules” was often what oth
ers might call “ethics” or “valuation” or “generally accepted accounting practices.”
Many of Gross’s innovations took the attention merchant’s model to new extremes. eToys, an online toy retailer, spent a huge portion of its venture capital on advertising during its short life, on the assumption that nothing more was needed to succeed. Gross also founded FreePC, which, while it lasted, gave people low-end Compaq computers in exchange for detailed personal information and the right to monitor their Internet surfing and buying habits. His NetZero was a “free” Internet service that depended on ads and similar tracking practices, at least until its collapse.
Gross would turn his “highest bidder” search idea into a company named GoTo.com, which got its start seven months before Google and was, for a while, its main competitor. Unlike Google, GoTo took payment from advertisers in return for higher search rankings in the search results; “relevance” depended on dollars spent in this scheme of search payola. Or, as Gross spun it, “We’re not letting a blind algorithm decide.” GoTo’s approach was a relative of AOL’s paid-for walled garden, and its trick of baiting users with search and delivering ads was branded misleading and unethical by critics; one complained, “If a middle school student does a search for ‘nutrition’ on GoTo, the first 221 sites listed are bidders” (mainly sites selling diets). When asked by The New York Times whether his search systematically favored commercial results, Gross answered that anyone could pony up for the same privilege. “Mr. Gross insists that nonprofit groups, universities and even medical organizations will find it in their interests to pay for placement,” the Times reported. As Gross said, “Even if it’s philanthropic, it wants your attention.”13
Among Gross’s most exercised critics were Brin and Page, who found his methods repulsive. In their seminal 1998 paper describing Google’s search algorithm, the two graduate students blasted the paid search model for creating a problem not easy to detect. “Search engine bias is particularly insidious,” they wrote. “It is very difficult even for experts to evaluate search engines.”14 Mixed incentives were the enemy, and that is precisely what being in the pay of advertisers produced. For this reason, search, the men believed, should be transparent and dispassionate.
More than anything, these sentiments reflected the personality of Larry Page, the main driving force behind Google. Page does not ooze charisma; he has a famously flat and unassuming manner; if you saw him in a meeting, at which he’d likely be eating lunch from a cafeteria tray, you might easily mistake him for some insufferably opinionated middle manager. Yet he’s a man of unusually bold brilliance and almost preternaturally instinctive—once he’s made up his mind. He is in his heart of hearts a purist, devoted to an engineering aesthetic whose beauty is hard to deny once you’ve understood it.
Google’s unspoken motto has always been “We do it better”; its founding culture a radical meritocracy, where brilliant code and ingenious system design count above all. The early Page had no patience for less technical sides of the business, including things like marketing, administration, even raising or making money. Not that he thought he was above all that; rather, he believed, with some reason, that companies often wasted most of their time and energy on such mundane matters, at the expense of the true goal.*2 That, of course, was inventing and building beautiful, perfect tools that were outrageously better than those that came before.
Two features distinguish Brin and Page’s favorite creations. First, they were always courageous new solutions to seemingly intractable old problems; ideally they were counterintuitive, even crazy seeming, yet once executed, utterly magical. Second, they were fast. Google demanded that its things work faster and better than anyone else’s, and without the compromises so many other products put up with. Combined, these two qualities came close to defining “Googley,” an adjective used on campus for things that reflected the firm’s idea of the good.15
However they could express this ideal they did. Why not have a campus where the food is both delicious and free, so you didn’t waste time in line at the cash register? Why not give away gigabits of free storage to users? Scan the world’s books? Drive trucks mounted with cameras through the streets everywhere and create a pictorial map? But nothing should be Googley-er than the flagship itself: and so Google search had to be not merely better than the rest but a thousand times better than those ugly and slow stepsisters, like Yahoo! and GoTo. The terms of Google’s self-styled triumph were indeed aesthetic. Where Google was fast, unadorned, and untouchable—Shining! Incorruptible!—it saw its rivals as essentially “sucky”—slow, festooned with advertising, and, ultimately, venal.
All of this made the prospect of advertising very difficult for Google and for Page in particular. To the purist, few things are smellier than advertising, and once you step in it, there is no easy way to fully wipe it off. Yet the time had come when Google had to show it could make money somehow. As pure as they were, Page and Brin also admired efficacy. Neither much respected idealists who never got anything done, the high-minded academics and programmers whose eyes and ideas never saw daylight. Google had come too far to founder on its own ideals.
And so, with the optimism of Silicon Valley engineers, Page and the other technical minds started to consider how they might invent their way out of their revenue problem. The question they would ask was whether there was some way of accepting ads that didn’t compromise their product. Could they come up with a “Diet Coke” solution—all of the taste, none of the calories? It went without saying that those “sucky” competitors freely taking ads weren’t as inventive. If Google took a crack at it, the company might create a form of ad that actually improved the product, instead of degrading it.
The idea of advertising that actually adds value is something of a Holy Grail in both tech and the advertising industry. Engineers knew exactly what ads were doing to web pages—slowing them down, taking up screen space, and diverting the user’s attention from what she really wanted to do; in some cases they were doing truly horrible things, like deliberately inserting bugs. But there had always been the theory that advertising, if not deceptive, could serve the useful purpose of informing the consumer—what marketers call “discovery.” Outside of the web, not all advertising was or is despised by customers. Print advertisements can be as beautiful as editorial content—for instance, the lavishly produced ads in Vogue; some can usefully connect a unique buyer and seller—as with classified listings in the newspaper. And some viewers of the Super Bowl tune in mainly for the commercial spots, which are famously the industry’s most innovative and witty. Could Google invent some web version of desirable advertising?
With a primitive advertising system devised by 2000, Page tapped a top engineer to make it better. Eric Veach was a brilliant Canadian mathematician, who, among his other qualifications, absolutely detested advertising. They got to work (long hours being another Google tradition) and by mid-2001 Veach and his team were convinced they’d solved the riddle. Borrowing some of Gross’s ideas (like auctions), they had invented an advertising system that turned prevailing wisdom about reselling attention on its head.
The attention merchant had always tried to reach as broad an audience as possible, bombarding them with as many ads as they’d stand before going into total revolt. Such had been the strategy in television since Pat Weaver first introduced the commercial break: cram fourteen to sixteen minutes of ad time into every hour, along with a further three to seven minutes of other forms of branding, eating up roughly one third of the broadcast; the content was devised essentially to cue up the ads, with a little cliffhanger before each one. It was in some ways little different from a mass mailing. It was worth the cost just to catch a tiny percentage at the optimal moment, when the viewers were receptive or in need of what was being sold, and also for the sake of building “brand awareness” among others who might turn into consumers. Everyone else was, as it were, collateral damage.
An annoying ad today might make you try Anacin tomorrow. But G
oogle had something different: access to people who were plainly expressing a commercial intent: when someone types in “drug rehab,” “male pattern baldness,” or “find a mortgage,” there’s little doubt what’s on their mind. The new system would therefore show relevant advertisements only at those moments, and—quite radically—stay out of your way otherwise. In short, Google took the decision to leave a lot of attention on the table; the theory was that advertisers would value surgical strikes at least as much as carpet bombing. Imagine a television channel that read your mind and showed you just one commercial an hour, keyed to your interests, reasoning that it would make you like that channel better: that was the early Google’s idea.
The technological implementation was called “AdWords,” and it would create auctions for text-only advertisements that would run alongside the search results when a specific search term (say, “plaintiff’s lawyer” or “mortgage”) was entered. A further brilliant stroke was to assign each ad a “quality score,” based on how often it was actually clicked on. Google insisted on running only ads that, in addition to being relevant, people had actually liked. And so, if you were searching for “Dead Sea scrolls,” you wouldn’t end up with advertisements for “mortgages” or “diet aids”—most likely, you wouldn’t see any ads at all.16
As soon as AdWords launched, it “accelerated past the outer moons of Jupiter on its way to some distant galaxy made entirely of money,” to quote Douglas Edwards, Employee Number 59.17 That happened for two reasons. First, Google, even while running ads, remained better than Yahoo!, Lycos, GoTo, and the rest, thus eating into their market share. Second, more importantly, Google’s advertising system finally delivered on metrics that satisfied advertisers and might even have elicited a nod from Arthur Nielsen in his grave. Among the various reasons for the dot-com crash was the fact that, over the late 1990s, AOL and everyone else had trouble proving that online ads necessarily did anything at all (hence AOL’s need to fabricate revenue). But with Google’s direct proof of clicks and tracking of customers, advertisers could finally see a direct link between their ads and eventual purchases. The beauty, once again, was that Google users were advertised to at a moment often very close to purchase. There was no need to prove the search engine could build brands—instead, it found users who were like travelers coming out of an airport, in a hurry, looking for a taxi, and there was the advertisement—“click.” With these innovations, the money started pouring in, first by the millions, then by the hundreds of millions, and finally billions upon billions. Google would soon become the most profitable attention merchant in the history of the world.*3