“The revelation of this scheme shows just how deeply fraud is embedded in the digital advertising ecosystem, the vast sums being stolen from brands, and the overall failure of the industry to stop it,” BuzzFeed News reported.9 But the fact that only Android apps were targeted highlights how vulnerable the Google platform—and, in turn, everyone who uses it—is to fraud and data breaches.
Though this story failed to make the front pages (after all, when it broke in fall 2018 the media had more pressing issues on their hands), the revelation was enough to prompt Mark Warner, a senior Democratic senator from Virginia, to write to the Federal Trade Commission, calling on it to address “the prevalence of digital advertising fraud and in particular the inaction of major industry stakeholders in curbing these abuses.”
It was one of many, many letters that he and other senators have written in recent years, trying to get Big Tech to change its behavior. But simply making a few half-hearted efforts on the margins—adding a few human watchdogs here and there, or reiterating their supposed commitment to quality content over propaganda—is akin to trying to treat an aggressive cancer with a multivitamin. Why? Because these problems—filter bubbles, fake news, data breaches, and fraud—are all at the center of the most malignant—and profitable—business model in the world: that of data mining and hyper-targeted advertising.10
The Aura of Science
The Cambridge Analytica scandal, whereby it was revealed that the Facebook platform had been exploited by foreign actors to influence the outcome of the 2016 presidential election—precipitated a huge rise in public awareness of how social media and its advertising-driven revenue model could pose a threat to liberal democracy. But the surveillance business model itself was pioneered at Google, not Facebook, and its founders were aware of both the possibilities and the perils as early as 1998, when Brin and Page were still coming up with the name for their new venture. By the time they’d settled on google instead of googol (Page was the one who simplified the spelling—or, in one version of the tale, simply mistyped), and immortalized it in a cheerful logo, the powers that be at Stanford had begun getting very curious about this mysterious project that was siphoning off so much computer power, and reminded Page and Brin that, as academics using university resources to conduct whatever research they were getting into, they should feel some obligation to publish their findings. Page and Brin disagreed. They were too busy working on perfecting their algorithms: those complex mathematical equations that crunched data into answers.
Algorithms have the aura of science—they are based on math and quantitative information, after all. And yet, they are all too human, in that they reflect the particular ideas and biases of the people who program them. Some are better than others, of course. And already, back in the late nineties, there was a sense that the ones Page and Brin had invented were very, very good—or at least very, very valuable and thus something they should keep to themselves.
“People [at Stanford] were saying, ‘Why is this so secret?’ ” Terry Winograd, who was mentoring both Page and Brin at the time, recalls wondering at the time. “ ‘This is an academic project. We should be able to know how it worked.’ ”11
This attitude highlights a fundamental difference between the academics and entrepreneurs. Academics are rewarded for revealing the findings of their research—ideally in a peer-reviewed journal—so others can learn from them. Entrepreneurs, on the other hand, need to keep proprietary secrets that could yield big money. It was becoming abundantly clear that Page and Brin were decidedly the latter.
Page, in particular, was wary of being scooped, citing the cautionary tale of Nikola Tesla, the brilliant Serbian scientist who in 1931 had been celebrated on the cover of Time magazine for his innovations in robotics, electricity, and radio, but who had died in poverty because he failed to commercialize his ideas. (Tesla is largely remembered today because Elon Musk paid the otherwise forgotten engineer tribute by naming his electric car after him.) Page vowed that he would not go the way of Tesla.12
Still, Winograd prevailed, and in 1998, while still at Stanford, Page and Brin published an academic paper entitled “The Anatomy of a Large-Scale Hypertextual Web Search Engine.”13 Mostly, it explained the inner workings of their search engine. But it also foretold an existential conflict over how search could actually make money. The issue centered around data mining—which was Sergey Brin’s area of expertise—and, more specifically, the fact that advertising driven by user data would turn out to be the way people would get very, very rich through this new invention.
Ironically, this was something that the Google founders were adamantly opposed to in the beginning. Not the data mining itself, but its marriage with search and targeted advertising. Data mining simply involved analyzing large amounts of data to discover trends and patterns in the aggregate.14 But the idea of tracking people’s individual behavior—what they searched for, which result they clicked on, and so on—and then building a database about who those people were so that the information could then be sold to the appropriate advertiser seemed anathema.
“If you read Larry and Sergey’s original paper that they wrote at Stanford, where they talked about creating a search engine, they specifically said that advertising would inherently corrupt the search engine if you sold advertising. So they were opposed to the notion of having advertising on Google,” recalled Douglas Edwards, one of the firm’s earliest software engineers.15
Indeed, this view is laid out in black and white. “Currently, the predominant business model for commercial search engines is advertising,” Page and Brin wrote on page 18, section 8, appendix A, titled “Advertising and Mixed Motives.” But, they added, “The goals of the advertising business model do not always correspond to providing quality search to users.”
In the appendix, they go on to say, “We expect that advertising funded search engines will be inherently biased towards the advertisers and away from the needs of consumers. Since it is very difficult even for experts to evaluate search engines, search engine bias is particularly insidious.” This was an interesting statement, given that Google has subsequently declared that everything they do, including some of the things that have caused the greatest controversy, is for the benefit of users. Interesting, too, how this statement underscores the inherent complexity of the technology itself—complexity that would give Googlers plenty of room to obfuscate later on when they were asked hard questions about the very bias of which Page and Brin were clearly aware.
Of course, the future implications of this were not fully apparent back in 1998. Still, it’s important to understand that even then, Page and Brin were somewhat worried. They had concluded that the risks of malfeasance in commercial search were not insignificant. They even considered whether search should be left in the public domain, where it wouldn’t be as easily manipulated as it might be under an ad-based business model. But the pair ultimately concluded that the downsides of private sector search were “likely to be tolerated by the market,” or, in other words, people either wouldn’t know or wouldn’t care that they were being manipulated.
And for a while, they didn’t.
A Million Clicks a Day
By the time Brin and Page published their paper, their friends and colleagues at Stanford were banging out more than ten thousand search queries a day. Though their site was still relatively unknown outside of a small circle of academics and tech geeks, its traffic was growing exponentially. Stanford wasn’t going to kick them out of the lab, since there was a certain pride associated with the new invention, but Page and Brin needed money to take their venture to the next level. Luckily, they had the Stanford network right at their fingertips.
They approached a professor, David Cheriton, who connected them with Andy Bechtolsheim, a Stanford alum who had gotten rich cofounding Sun Microsystems.16 Bechtolsheim saw immediately that the fastest way to make money was via advertising, a
nd in particular the targeted advertising that would appear alongside search results. Bechtolsheim recalls thinking, “Well, we’ll have these sponsored links and when you click on a link, we’ll collect five cents. And so I made this quick calculation in the back of my head: ‘O.K., they are going to get a million clicks a day at 5 cents, that’s $50,000 a day—well, at least they won’t go broke.’ ”17 Then he handed Brin and Page a check for $100,000, got back into his Porsche, and drove away.
He was the first. Subsequent early stage investors would include some Silicon Valley legends, including Amazon founder Jeff Bezos, who put in $250,000 in 1998. “There was no business plan,” said Bezos to journalist Ken Auletta in a New Yorker article. “I just fell in love with Larry and Sergey.”18
And so Google, the company, was born. That first year, they embodied the caricature of a young Silicon Valley start-up, complete with pilgrimages to Burning Man, fancy perks to lure top talent, and plenty of hanky-panky. “Sergey was the Google playboy,” remembers Charlie Ayers, the first company chef, whom Page and Brin hired when the company had only twelve employees.19 “He was known for getting his fingers caught in the cookie jar with employees that worked for the company, in the masseuse room. He got around. HR told me that Sergey’s response to it was, ‘Why not? They’re my employees,’ ” Ayers recalls in Adam Fisher’s book, Valley of Genius.20
In the meantime, big things were happening. The biggest of all was that the huge amount of search data that Google was generating had begun to create a self-reinforcing cycle, whereby users begat users, because in the end, search wasn’t so much about the brilliance of a particular algorithm, but the amount of data it had to work with.21 In other words, Google seemed to have captured the holy grail for any fledgling Internet company: the network effect, which simply means that the more people use a product or function, the better it becomes. The mechanism was almost ridiculously simple—a reputation for the best search would draw more users toward their site, and more users meant more data—which would build not only a better search engine, but also eventually revenue.
Google had up until that point been earning most of its money from licensing its search technology to various content sites, slowly but surely accumulating the traffic that would beget more traffic. Moreover, all of this user behavior would now be tracked via “cookies,” digital tags that saw where people went and what they did online. The cookies didn’t ID people by name or address, but they did provide information for a database that would prove extraordinarily useful for the company’s nascent advertising business. After all, the more you knew about what people were doing, and when, and where, on the Internet, the more you could make selling that information to advertisers who wanted to reach them exactly where they were.
Even as Google was gaining traction, there were still plenty of bigger search engines with more users out there, most notably Excite, which had been funded by yet another Sun Microsystems alum, venture capitalist Vinod Khosla. And of course, there was Yahoo. The latter was arguably the shiniest dot-com start-up in the Valley at that point, the Swiss Army Knife of websites that dabbled in everything from search to email to content aggregation and creation. Yahoo decided that rather than focus on the nuts and bolts of search itself, it would focus on simply being a “portal” to the Web, and outsource search to Google.
That was Google’s big break, which came in June 2000. In addition to the millions of dollars Yahoo paid them, the deal granted Google a major concession—users of Yahoo’s new search function would see Google’s logo and a message that Google was, in fact, powering the search.22 But the most valuable part of the deal for Google wasn’t the money or the brand recognition—it was the data. Not only would it provide the raw material that made the search engine itself smarter, and thus help it continue to attract new users, it would also fuel what would soon become the equivalent of a money-printing machine—the online advertising system known as AdWords.
Selling advertising, as we’ve learned, had not been part of the original plan (to the extent that there was an original plan). But while the millions from the Yahoo deal were nice, Google needed more than just a onetime hit of cash to make its investors really happy; they needed a consistent revenue stream. What good were all these users, anyway, if you couldn’t make money from them? The company was burning through $25 million a year, and its VCs were getting anxious about the business proposition.23 “There was a lot of pressure to generate revenue,” recalled early Googler Douglas Edwards, and so Page and Brin decided that maybe they’d been too restrictive in their moral philosophy. Maybe it wouldn’t be so bad after all to put paid advertising at the top of the search results. The two decided “advertising doesn’t have to be evil—if it’s actually useful and relevant,” said Edwards.24
The idea for what eventually became AdWords was the brainchild of a man who was already becoming a Silicon Valley legend for his Midas touch: the serial entrepreneur, investor, and visionary Bill Gross. A brilliant Caltech grad, Gross had founded Idealab, an “incubator” that rolled out start-ups the way movie studios rolled out films.25 (At press time, Gross had launched 150 companies and done forty-five IPOs or merger and acquisition deals.)
When Google was founded in 1998, Gross had just started a company called GoTo.com, a kind of Yellow Pages of the net, which allowed businesses to compete in an auction for the advertising space next to the most relevant Web pages. For example, advertisers of, say, mountain bikes, would bid on the search terms mountain bikes, and the highest bidders would get to display their ad alongside the results. When GoTo first debuted its technology at a TED conference, people were impressed but also horrified—Gross was tampering with the supposedly objective search results by allowing advertisers to pay for ad space that made them look higher up the rankings than they were. This was so counter to the prevailing ideology at the time that people actually hissed during Gross’s demonstration.26 Undeterred, Gross went on to cut huge deals to provide his service to AOL.com and other companies, netting $50 million in revenue.
Meanwhile, the Google guys—and even more important, their investors—were watching intently, and what they saw was dollar signs. “GoTo prospered, and Google executives took notice,”27 wrote Wired cofounder John Battelle, author of The Search, which outlines the development and commercialization of Google’s search function. Google investor Michael Moritz in particular had become worried that the existing revenue model of licensing search technology to other companies simply wasn’t going to work, calling it “a brutal path.”28 Why try to make money deal by deal, rather than by leveraging the power of big data and advertising?
Here again, the network effect was key; more data meant better search results, which meant more advertisers, which meant more traffic clicking through to more ads, which meant more data, and so on. He pushed Google to look closely at GoTo’s technology as a model, and they did. “People started reading about how much money was being brought into various other companies by search advertising, and it was kind of decided that we were leaving money on the table,” noted Ray Sidney, Google employee number five.29
He had a point. The network effect was proving to be formidable. Google had powered 3 million searches a day in August 1999—and by the summer of 2000, that number was up to 18 million. Add in the searches they were powering for Yahoo, and that number rose to 60 million. But those searches and the user data they yielded hadn’t yet been fully monetized by targeted advertising—people simply didn’t click on sponsored links the way they would if the links paid for by advertisers actually appeared to be part of the organic search results, as they did in the GoTo system. So the Google founders “cleverly fastened on the proposition offered by GoTo,” as Google VC Moritz put it. Had they not “adopted some of the advertising techniques that were working for others, [Google] would have ended up a small, but nice, high-end company.”30
The result, as we now know, was a company that would
come to dominate the entire search and advertising industries, and set a business model standard for the platform technology companies as a whole. Targeted advertising would become the key way to make money not just for Google, but for Facebook, Amazon (which uses targeted advertising to lure users to its own site and prompt them to click on things they may have been interested in before), and a host of other companies. It created an entirely new model for how to get people to buy things.
Surveillance capitalism would become the best and fastest way for companies to grow—not just the platforms themselves, but all the other sorts of businesses that would advertise through them. That’s a key point. While it would only be a matter of time before companies like Google would enter all kinds of other industries—from healthcare to transportation—the business interests that might have protested their increasing power in the marketplace remained silent, because they were benefiting from the targeted advertising model that the platforms delivered. Companies of all sorts could now spy on customers 24/7, and target them ever more precisely. It was a Faustian bargain, since they were giving up more control than they were getting, and allowing the Big Tech firms to grow fat in ways that would eventually come back to bite them, as Silicon Valley began to enter their markets.
But the promise of an economy that ran on data, rather than dollars, was simply too enticing to resist. The attention merchants had risen—and the race was on to capture as many eyeballs as possible, and keep them online for as long as possible. The filter bubbles, fake news, and challenges to liberal democracy that would come from that would be a problem for other people, at another time.
Don't Be Evil Page 8