7. The Legacy of the Dot-com Era

The Discipline of the Market

In 2004, U.S. marketers spent nearly $10 billion on internet advertising, roughly equivalent to the amount spent on magazines.17 Internet ad spending was once again growing at a rapid clip, but the proportion going to search had plateaued in just three years.18 Google found itself in control of a saturated domestic search advertising market that accounted for nearly 100 percent of its revenues. Now subject to the uncompromising scrutiny of Wall Street, Google needed to find a way to maintain forward momentum. It was a classic case of what Douglas Rushkoff calls the “growth trap,” where companies’ operations become rigidly oriented toward expansion in order to meet shareholder expectations for increasing profits.19 Simply maintaining a successful search advertising business was not sufficient. Google needed to move into new markets. One obvious place to look for growth was the second largest chunk of internet advertising: the adjacent market of targeted banner ads. “Google is a one-trick pony,” declared a securities analyst. “It’s a nice-looking pony. But they have to grab a bigger piece of the display advertising market.”20

Google had already created a program called AdSense that enabled any web publisher to host Google’s text ads, broadening the company’s reach considerably. Lifted directly from the ad network playbook, AdSense was a distributed service that, in Ken Auletta’s words, “turned the web into a giant Google billboard.”21 Still, Google remained distinct from surveillance advertising because these ads were targeted on the basis of the context provided by the surrounding web page rather than consumer data. Google began to move into display advertising with modest success, adding some graphical options to its text ads and striking partnerships with the likes of AOL to tap into its sales force in order to sell banners on Google’s AdSense partner sites.22 As broadband internet service diffused, online video emerged as a promising new ad format. In 2006, Google acquired YouTube, the hugely popular (and virtually advertising-free) video streaming platform. Enamored with prospects for “interactive television” since the early 1990s, marketers were eager to test out the format. Yet these efforts fell short of what was needed to maintain the company’s upward trajectory. Online video remained a poor substitute for television advertising, and dabbling in display ads was not nearly enough to move the needle. To effectively exploit networked display and video formats, Google needed to get into the business of consumer surveillance, and it needed to do so quickly.

Competition was closing in. The other major companies in the search engine space, Yahoo and Microsoft, had also declared their intentions to move forcefully into targeted ads. Compounding the matter, social networks including MySpace, Twitter, and Facebook broke onto the scene in the mid-2000s and began to attract large numbers of users. Social networks offered novel experiences, giving people the ability to create their own unique profiles, share content, and connect with others. The fast-growing social networks began to capture firsthand a range of personal information that was otherwise difficult to obtain. Soon these companies began using data about user demographics, interests, and social connections—what Facebook CEO Mark Zuckerberg called the social graph—to inform targeted ad campaigns. The central business proposition of social networks stemmed from their arguably superior capacities to collect and operationalize personal information, not from any significant technical innovations in advertising practice. The stunning growth of Facebook, which amassed over a billion worldwide users in its first decade, brought strong competitive pressure to the entire internet advertising industry, and Google in particular, to build out—or acquire—capacities for consumer surveillance.

It was perhaps Google’s good fortune that the leading purveyor of surveillance advertising was about to be put up for sale. Outflanked by search advertising and sorely missing its dot-com era market valuation, DoubleClick was downsized in its private equity takeover. Still, the company’s strong market position and technological prowess in targeted display put it in a better position than most to weather the bubble’s collapse. And there was still an appetite among marketers for data-driven advertising. As one industry observer put it, “You’re going to see online advertising is extremely targeted . . . because every cent has to count.”23 DoubleClick remained the industry’s de facto standard for surveillance advertising, delivering more targeted banners annually in 2001 and 2002 than it did in any of the years leading up the crash, and for the first time ever, the company was profitable.24

In early 2007, Google, entered into a bidding war with Microsoft and Yahoo to buy DoubleClick from its private equity owners. Google emerged the victor but paid an estimated $1 billion premium.25 The final cost was $3.1 billion (in stock), nearly double what Google had paid for YouTube the year before. Company leaders justified the expense as necessary not only to acquire DoubleClick but also to keep it out of the hands of Google’s competitors. The acquisition instantly gave Google a major market position in display ads and yielded an extensive roster of DoubleClick clients from both the supply side (web publishers) and the demand side (marketers and ad agencies) of the online advertising market. “This shores up Google as the absolute leader,” said an analyst with Forrester Research. With major positions in search, video, and targeted display, “there isn’t anything they don’t have.”26

In addition to consolidating market power, the deal gave Google control over the surveillance advertising platform that DoubleClick had been building for over ten years.27 Google’s overarching objective was to integrate the companies’ operations to offer an expanded range of targeted advertising services to publishers and marketers alike. According to BusinessWeek, Google saw DoubleClick as a way to “lock the interactive [advertising] agencies and media buyers deeper into their clutches, and . . . offer an integrated network that [was] easy to buy, easy to measure, and easy to manage.”28 Although its AdSense program had already made Google a multisided platform, DoubleClick added a host of new relationships with marketers and publishers that were grounded in the collection and exchange of consumer data. “Our goal is to create a single and complete advertising system,” said Sergey Brin and Larry Page in a letter to shareholders.29 It was clear that Google’s founders were developing a platform monopoly strategy and that surveillance advertising would play a key part.

Before the ink was dry on the DoubleClick purchase contract, Google’s biggest competitors went on a shopping spree to avoid drowning in its wake. Within four months, Microsoft, AOL, Yahoo, and the advertising holding giant WPP each announced plans to acquire one or more internet advertising companies with core competencies in consumer monitoring and networked ad distribution. In its largest acquisition to date, Microsoft paid $6 billion for aQuantive, a large ad-services provider built in part with the remnants of DoubleClick’s former rival, CMGI.30

This cluster of mergers marked an inflection point where the differences among internet advertising formats and their respective data practices began to fall away. Propelled by market forces, the formerly distinct practices of contextual search advertising and targeted display converged on consumer surveillance. Google’s acquisition of DoubleClick was finalized on March 11, 2008.31 One year later to the day, Google announced a new service called interest-based advertising that would, for the first time, allow advertisers to target consumers on the basis of user profiles populated with behavioral data. As Susan Wojcicki, vice president of product management, explained:

To date, we have shown ads based mainly on what your interests are at a specific moment. So if you search for [digital camera] on Google, you’ll get ads related to digital cameras. . . . We think we can make online advertising even more relevant and useful by using additional information about the websites people visit. Today we are launching “interest-based” advertising as a beta test on our partner sites and on YouTube. These ads will associate categories of interest—say sports, gardening, cars, pets—with your browser, based on the types of sites you visit and the pages you view. We may then use those interest categories to show you more relevant text and display ads.32

The new service would power targeted display ads across Google’s publisher networks, thereby reincarnating the surveillance advertising campaigns pioneered by DoubleClick back in the dot-com heyday. Importantly, the new “interest-based ads” would be enabled by default for everyone.

Google’s pivot toward surveillance turned out to be a smashing success. In 2009, the company sold display advertising to ninety-four of Advertising Age’s top hundred advertisers and earned $6.5 billion in profits, up 50 percent from the year before.33 As its founders explained to shareholders, Google had “really benefited from a successful integration with DoubleClick.”34 Over the course of the next decade, Google systematically incorporated surveillance advertising into its operations, profiling billions of people worldwide using data harvested from its own services, as well as a network of millions of web publisher and mobile partners. Charlie Warzel and Ash Ngu of the New York Times highlighted this dramatic expansion of surveillance with a detailed account of how Google has modified its privacy policy over time.35 Google’s first privacy policy, created in 1999, sparsely explained that the company did not collect data on individuals. Twenty years and thirty versions later, it now takes four thousand words to outline the company’s extensive data collection practices across its array of websites, applications, and partners.

An important moment came in 2012 when Google announced that it would begin tracking users universally across its many services in order to further refine its ad-targeting capabilities. Before this, Google’s data collection was more of a patchwork operation wherein information was kept in silos that were not connected to composite profiles. Critical observers argued that the move toward universal profiles undermined the ad industry’s long-standing claim that online surveillance was anonymous. From this point forward, personally identifiable information gathered from Gmail, Maps, and the ill-fated Google Plus social network would be combined with Google’s extensive histories of search, web browsing, and YouTube viewing habits. Although the company allowed users to modify certain settings via a renovated privacy manager tool and promised to keep targeting data anonymized, there was no option to opt out of the broader centralization of data. As Gizmodo put it in a piece subtitled “The End of Don’t Be Evil,” “If you want to use Google services, you have to agree to these rules.”36

Google’s appropriation of surveillance advertising plots the historical trajectory of the entire online advertising sector, and indeed the commercial internet at large. After a brief stall in the wake of the dot-com stock market collapse, surveillance advertising and its progenitors were summarily integrated into the rising central power structure of internet advertising. By the end of the 2000s, the five largest U.S. internet advertising companies—Google, Facebook, Microsoft, AOL, and Yahoo—all served profile-based targeted advertising and collected consumer data across expansive networks that included their own properties and millions of other websites and applications. Each also offered ad exchanges, software-powered markets for buying and selling access to individual consumers in real time, a process dependent on persistent transfer of identification data among any number of buyers and sellers.

This configuration is the realization of the platform monopoly that DoubleClick and its ilk sought to build in the dot-com era. By fusing consumer monitoring and ad distribution, the first generation of ad platforms created a technical prototype and provisional organizational model for web advertising that placed surveillance at the core. As the New York Times described in 2001, the business model advanced by DoubleClick “fundamentally altered the nature of surfing the web from being a relatively anonymous activity, like wandering the streets of a large city, to the kind of environment where records of one’s transactions, movements and even desires could be sorted, mined and sold.”37

Once surveillance advertising gained a foothold, competitive pressures made it increasingly difficult for organizations to abstain. This is the discipline of the market in action. When Susan Wojcicki announced Google’s first behavioral profiling initiative in 2009, she acknowledged that “this kind of tailored advertising does raise questions about user choice and privacy.” Pressed on the question of whether users should be able opt into targeted behavioral ads rather than have them turned on by default, a Google spokesperson simply said that “offering advertising on an opt-in basis goes against the economic model of the internet.”38 Wojcicki went on to explain that that such practices were already widespread and that advertisers had been “asking us for a long time to offer interest-based advertising.”39 Privacy concerns were signaled as legitimate, but they were “questions the whole online ad industry has a responsibility to answer.” Wojcicki seemed to suggest that, as powerful as it was, Google had little choice in the matter. Surveillance advertising was not a Google problem. It was an internet problem.

If Google, after building a wildly successful advertising business free of consumer monitoring, was compelled to embrace surveillance advertising, then how much choice did individual publishers or marketers have? As Stacy Lynn Schulman, an executive at the advertising giant Interpublic argued in 2007, adopting data-driven advertising was “less about competitive advantage and more about survival.”40 Conducting consumer surveillance was not optional for marketers and ad agencies, explained Shulman. It was “simply the price of entry.” Surveillance advertising stems from the instrumental desires of marketers and ad platforms to grow their businesses. However, it is also deeply rooted in structural factors: the competitive pressures and growth imperatives of the global capitalist economy.

Today, organizations of all kinds face such immense pressure to participate in consumer monitoring that nearly all components of the internet function as terminals of surveillance for platform hubs, sweeping data in and out of the systems of a handful of surveillance advertising’s big winners. Although it has not been the focus of this book, it is important to note, as Yasha Levine argues, that this centralization of commercial surveillance capacity abets the data collection efforts of the national security state.41 In important ways, commercial and state surveillance are two sides of the same coin. One need look no further than the 2013 revelations of whistle-blower Edward Snowden, which, among many other things, documented the National Security Agency’s close partnerships with surveillance advertising’s biggest players to conduct illegal mass surveillance on the American public.42

From banners to search to video, surveillance has been embedded in the internet to a greater extent than any other communications medium in history. Virtually all websites and digital applications now serve as gateways to cascading layers of data collection. Numerous studies show that the web’s most popular sites and services not only overwhelmingly monitor their users but also share that user data with third-party ad platforms by giving them direct access to collect user information.43 In recent years, it has become difficult to draw the lines demarcating where internet surveillance stops and off-line data collection begins. Aggregating online and off-line data, widely considered a bridge too far when DoubleClick attempted it in the late 1990s, is now standard industry practice. Mobile devices transmit precise location data to any number of companies thousands of times per day.44 Data brokers and ad platforms combine this information with names, addresses, web browsing histories, brick-and-mortar retail purchases, and much more for use in targeted advertising campaigns. Contextual spillover is the norm, as data gathered for marketing purposes are used in myriad other applications, factoring into determinations about things like health insurance premiums, loan interest rates, college admissions, and housing applications.45 As John Cheney-Lippold argues, these data are systematically constructed into identities that stand in and speak for us, “whether we know about, like it, or not.”46

Internet users are now faced with the very situation that government analysts warned about in 1994. As I described in chapter 1, policy planners at the dawn of the internet’s commercialization understood that a market-based system for internet data collection would create powerful pressures for companies to indiscriminately surveil internet users. They cautioned that if the internet were to become dominated by a handful of big companies, consumers would have few options to escape commercial surveillance. Whether this future would come to pass or not was less a question about markets than it was about politics.

Back to Politics

The history of surveillance advertising is not a story of disruption but rather one of continuity. This history reveals just how long the modern surveillance advertising apparatus has been under construction and how deeply it is entrenched in the capitalist political economy. The world’s dominant internet companies pivoted toward consumer surveillance not with new technological innovations but by incorporating and building on the sociotechnical systems created in the dot-com era. Among the many interlocking components of surveillance advertising’s construction addressed in this book, public policy played a foundational role.

In an op-ed entitled “What If We All Just Sold Non-creepy Advertising?,” Gabriel Weinberg argues that marketers and publishers could do just fine without consumer surveillance.47 Weinberg is the CEO of Duck Duck Go, a search engine that has managed to build a business doing purely contextual advertising and collecting no consumer data. Although Duck Duck Go is proof that the discipline of the market is not totalizing, the company nevertheless exists at the margins of the advertising industry. Duck Duck Go is a tiny fraction of the size of Google, operating entirely in the tech giant’s shadow. Weinberg understands the difficulty of swimming against the current of surveillance as well as anyone but remains an advocate for internet privacy. Arguing that “people have a fundamental right to avoid being put under surveillance,” Weinberg sails past the standard list of industry talking points about how to address privacy concerns. He does not suggest enhanced transparency in data collection, greater market competition, more responsible executives, or industry self-regulation. Instead, Weinberg points immediately to politics, advocating for “strong privacy laws” to “force the digital advertising industry to return to . . . contextual advertising” without consumer surveillance.48

As I have argued in this book, the history of surveillance advertising provides strong support for Weinberg’s political instinct. Public policy is among the only levers of power capable of tempering capitalist digital enclosure and the relentless drive to commodify information. It is, after all, the exact set of tools that private and public actors used to set up the legal foundations of surveillance advertising in the first place. When Google bought DoubleClick back in 2007, the merger had to be cleared by the Federal Trade Commission. After a yearlong investigation into potential competitive ramifications, the deal was permitted by a 4–1 vote. Although privacy advocates filed objections with the FTC, concerns about consumer surveillance were not formally factored into the deliberations. Explaining the exclusion, the FTC noted that privacy concerns were “not unique to Google and DoubleClick,” but rather “extend to the entire online advertising marketplace.”49 In other words, the FTC argued that surveillance advertising was already so well established that it did not make much sense to question the institutional buildup of surveillance capacity that would result from the merger. Equally significant, the commissioners admitted that even if they had wanted to consider data collection and privacy issues as part of the merger review, they simply had little jurisdiction over such matters. Consumer surveillance on the internet is industry’s domain. The private sector is in charge. This is the political legacy of the dot-com era.

The surveillance advertising industry remains acutely aware of the importance of public policy. Google and Facebook have built empires on the proposition that the founding political principle of the internet—private sector leadership—can endure any techlash with enough lobbying and public relations maneuvering. Testifying before Congress in 2018, Facebook’s Mark Zuckerberg told lawmakers he was “not opposed to regulation,” as long as it was the “right regulation.”50 As Weinberg notes, surveillance advertising executives “may individually make public statements welcoming federal regulation, but in practice they are doing everything they can to weaken existing laws and shape new ones in their own interests.”51

It is crystal clear to anyone paying attention that industry self-regulation and the “notice and choice” privacy paradigm are utter failures.52 That these policies remain intact reveals the antidemocratic nature of policy making in the United States, a feature that is unfortunately hardly limited to the domain of internet governance.53 When pressed, surveillance advertising platforms will continue to roll out transparency tweaks, privacy dashboards, and other changes that fiddle at the margins of their enterprises. They will curb some of the more egregious uses of their systems, while their public relations teams applaud a job well done. What these companies will not do, however, is anything that might undermine their core business model of unaccountable surveillance—that is, unless democratic society gives them no choice.

The only solution to a problem of this magnitude is a political program that confronts the surveillance advertising business model head on. There is no easy path forward, but the United States might look to the European Union as one source of inspiration. Over the howls of the digital advertising industry, the E.U.’s 2018 data protection directive introduces a set of limitations on the collection and use of consumer data. Heeding ample evidence, regulators in Brussels have decided that free markets are simply not up to the task of protecting privacy.54 It is not a perfect solution, but it points to a future in which people have a say over the communications technologies that structure so much of their lives.

In the United States, we now know what twenty-five years of neoliberal internet governance looks like. We are living with the outcome of letting the private sector lead. It is past time for an alternative political vision for the internet, one that includes greater democratic accountability, more equitable distribution of power, and far less subservience to the demands of the market.