Google’s first marketing hire lasted a matter of months in 1999; his experience included Miller Beer and Tropicana and his proposal involved focus groups and television commercials. When Doug Edwards interviewed for a job as marketing manager later that year, he understood that the key word was “viral.” Edwards lasted quite a bit longer, and now he’s the first Google insider to have published his memoir of the experience. He was, as he says proudly in his subtitle to I’m Feeling Lucky, Google employee number 59. He provides two other indicators of how early that was: so early that he nabbed the e-mail address email@example.com; and so early that Google’s entire server hardware lived in a rented “cage.”
Less than six hundred square feet, it felt like a shotgun shack blighting a neighborhood of gated mansions. Every square inch was crammed with racks bristling with stripped-down CPUs [central processing units]. There were twenty-one racks and more than fifteen hundred machines, each sprouting cables like Play-Doh pushed through a spaghetti press. Where other cages were right-angled and inorganic, Google’s swarmed with life, a giant termite mound dense with frenetic activity and intersecting curves.
Levy got a glimpse of Google’s data storage a bit later and remarked, “If you could imagine a male college freshman made of gigabytes, this would be his dorm.”
Not anymore. Google owns and operates a constellation of giant server farms spread around the globe—huge windowless structures, resembling aircraft hangars or power plants, some with cooling towers. The server farms stockpile the exabytes of information and operate an array of staggeringly clever technology. This is Google’s share of the cloud (that notional place where our data live) and it is the lion’s share.
How thoroughly and how radically Google has already transformed the information economy has not been well understood. The merchandise of the information economy is not information; it is attention. These commodities have an inverse relationship. When information is cheap, attention becomes expensive. Attention is what we, the users, give to Google, and our attention is what Google sells—concentrated, focused, and crystallized.
Google’s business is not search but advertising. More than 96 percent of its $29 billion in revenue last year came directly from advertising, and most of the rest came from advertising-related services. Google makes more from advertising than all the nation’s newspapers combined. It’s worth understanding precisely how this works. Levy chronicles the development of the advertising engine: a “fantastic achievement in building a money machine from the virtual smoke and mirrors of the Internet.” In The Googlization of Everything (and Why We Should Worry), a book that can be read as a sober and admonitory companion, Siva Vaidhyanathan, a media scholar at the University of Virginia, puts it this way: “We are not Google’s customers: we are its product. We—our fancies, fetishes, predilections, and preferences—are what Google sells to advertisers.”
The evolution of this unparalleled money machine piled one brilliant innovation atop another, in fast sequence:
1. Early in 2000, Google sold “premium sponsored links”: simple text ads assigned to particular search terms. A purveyor of golf balls could have its ad shown to everyone who searched for “golf” or, even better, “golf balls.” Other search engines were already doing this. Following tradition, they charged according to how many people saw each ad. Salespeople sold the ads to big accounts, one by one.
2. Late that year, engineers devised an automated self-service system, dubbed AdWords. The opening pitch went, “Have a credit card and 5 minutes? Get your ad on Google today,” and suddenly thousands of small businesses were buying their first Internet ads.
3. From a short-lived startup called GoTo (by 2003 Google owned it) came two new ideas. One was to charge per click rather than per view. People who click on an ad for golf balls are more likely to buy them than those who simply see an ad on Google’s website. The other idea was to let advertisers bid for keywords—such as “golf ball”—against one another in fast online auctions. Pay-per-click auctions opened a cash spigot. A click meant a successful ad, and some advertisers were willing to pay more for that than a human salesperson could have known. Plaintiffs’ lawyers seeking clients would bid as much as fifty dollars for a single click on the keyword “mesothelioma”—the rare form of cancer caused by asbestos.
4. Google—monitoring its users’ behavior so systematically—had instant knowledge of which ads were succeeding and which were not. It could view “click-through rates” as a measure of ad quality. And in determining the winners of auctions, it began to consider not just the money offered but the appeal of the ad: an effective ad, getting lots of clicks, would get better placement.
Now Google had a system of profitable cycles in place, positive feedback pushing advertisers to make more effective ads and giving them data to help them do it and giving users more satisfaction in clicking on ads, while punishing noise and spam. “The system enforced Google’s insistence that advertising shouldn’t be a transaction between publisher and advertiser but a three-way relationship that also included the user,” writes Levy. Hardly an equal relationship, however. Vaidhyanathan sees it as exploitative: “The Googlization of everything entails the harvesting, copying, aggregating, and ranking of information about and contributions made by each of us.”
By 2003, AdWords Select was serving hundreds of thousands of advertisers and making so much money that Google was deliberating hiding its success from the press and from competitors. But it was only a launching pad for the next brilliancy.
5. So far, ads were appearing on Google’s search pages, discreet in size, clearly marked, at the top or down the right side. Now the company expanded its platform outward. The aim was to develop a form of artificial intelligence that could analyze chunks of text—websites, blogs, e-mail, books—and match them with keywords. With two billion Web pages already in its index and with its close tracking of user behavior, Google had exactly the information needed to tackle this problem. Given a website (or a blog or an e-mail), it could predict which advertisements would be effective.
This was, in the jargon, “content-targeted advertising.” Google called its program AdSense. For anyone hoping to—in the jargon—“monetize” their content, it was the Holy Grail. The biggest digital publishers, such as The New York Times, quickly signed up for AdSense, letting Google handle growing portions of their advertising business. And so did the smallest publishers, by the millions—so grew the “long tail” of possible advertisers, down to individual bloggers. They signed up because the ads were so powerfully, measurably productive. “Google conquered the advertising world with nothing more than applied mathematics,” wrote Chris Anderson, the editor of Wired. “It didn’t pretend to know anything about the culture and conventions of advertising—it just assumed that better data, with better analytical tools, would win the day. And Google was right.” Newspapers and other traditional media have complained from time to time about the arrogation of their content, but it is by absorbing the world’s advertising that Google has become their most destructive competitor.
Like all forms of artificial intelligence, targeted advertising has hits and misses. Levy cites a classic miss: a gory New York Post story about a body dismembered and stuffed in a garbage bag, accompanied on the Post website by a Google ad for plastic bags. Nonetheless, anyone could now add a few lines of code to their website, automatically display Google ads, and start cashing monthly checks, however small. Vast tracts of the Web that had been free of advertising now became Google part- ners. Today Google’s ad canvas is not just the search page but the entire Web, and beyond that, great volumes of e-mail and, potentially, all the world’s books.
Search and advertising thus become the matched edges of a sharp sword. The perfect search engine, as Sergey and Larry imagine it, reads your mind and produces the answer you want. The perfect advertising engine does the same: it shows you the ads you want. Anything else wastes your attention, the advertiser’s money, and the world’s bandwidth. The dream is virtuous advertising, matching up buyers and sellers to the benefit of all. But virtuous advertising in this sense is a contradiction in terms. The advertiser is paying for a slice of our limited attention; our minds would otherwise be elsewhere. If our interests and the advertisers’ were perfectly aligned, they would not need to pay. There is no information utopia. Google users are parties to a complex transaction, and if there is one lesson to be drawn from all these books it is that we are not always witting parties.
Seeing ads next to your e-mail (if you use Google’s free e-mail service) can provide reminders, sometimes startling, of how much the company knows about your inner self. Even without your e-mail, your search history reveals plenty—as Levy says, “your health problems, your commercial interests, your hobbies, and your dreams.” Your response to advertising reveals even more, and with its advertising programs Google began tracking the behavior of individual users from one Internet site to the next. They observe our every click (where they can) and they measure in milliseconds how long it takes us to decide. If they didn’t, their results wouldn’t be so uncannily effective. They have no rival in the depth and breadth of their data mining. They make statistical models for everything they know, connecting the small scales with the large, from queries and clicks to trends in fashion and season, climate and disease.
It’s for your own good—that is Google’s cherished belief. If we want the best possible search results, and if we want advertisements suited to our needs and desires, we must let them into our souls.
The Google corporate motto is “Don’t be evil.” Simple as that is, it requires parsing.
It was first put forward in 2001 by an engineer, Paul Buchheit, at a jawboning session about corporate values. “People laughed,” he recalled. “But I said, ‘No, really.’” (At that time the booming tech world had its elephant-in-the-room, and many Googlers understood “Don’t be evil” explicitly to mean “Don’t be like Microsoft”; i.e., don’t be a ruthless, take-no-prisoners monopolist.)
Often it is misquoted in stronger form: “Do no evil.” That would be a harder standard to meet.
Now they’re mocked for it, but the Googlers were surely sincere. They believed a corporation should behave ethically, like a person. They brainstormed about their values. Taken at face value, “Don’t be evil” has a finer ring than some of the other contenders: “Google will strive to honor all its commitments” or “Play hard but keep the puck down.”
“Don’t be evil” does not have to mean transparency. None of these books can tell you how many search queries Google fields, how much electricity it consumes, how much storage capacity it owns, how many streets it has photographed, how much e-mail it stores; nor can you Google the answers, because Google values its privacy.