Tech Giants as Ad Brokers

Abstract

In 2017, The Australian published a confidential document leaked from Facebook. It revealed how the company had offered advertisers the opportunity to target ads to 6.4 million Facebook users as young as 14 who were going through psychologically vulnerable phases of their lives. In these phases, the teenagers felt “worthless”, “insecure”, “stressed” or “anxious”, among other things. To track these emotional downturns, Facebook had monitored the messages, images, interactions and activity of their users on the Internet.

In 2017, The Australian published a confidential document leaked from Facebook. It revealed how the company had offered advertisers the opportunity to target ads to 6.4 million Facebook users as young as 14 who were going through psychologically vulnerable phases of their lives. In these phases, the teenagers felt “worthless”, “insecure”, “stressed” or “anxious”, among other things. To track these emotional downturns, Facebook had monitored the messages, images, interactions and activity of their users on the Internet.¹

Using big data harvested from user behavior and analyzed by sophisticated algorithms, Facebook, Twitter, Google and other tech companies can predict large parts of people’s preferences and behaviors. This is tremendously valuable for advertisers who want to control and manipulate consumer behavior. Internet guru Jaron Lanier, who in his recent book recommends for people to shut down their accounts on social media, calls the real objective of the attention economy “behavioral modification”. The tech giants quite simply strive to change people’s behavior.² Generally speaking, advertising is not necessarily manipulative. But when very specific parts of the personal information of the consumers are used against them to lure them to buy products that they would not otherwise buy, then it is. When an ad for slimming tea targets teens right at the moment they feel the most down, or when ads for liquor brands can be directed at easy-to-tempt alcoholics, then it is. The concerns of today not only revolve around how much information the tech giants have on us. The question is also to what extent this information can be used as a weapon against the users without us having—or getting—a chance to even find out.

Facebook, Twitter and Google have different and, to various degrees, idealistic points of departure. Google’s mission is to organize the world’s information and make it universally available and useful. Twitter wants to give all people the possibility to create and share ideas and information—at any time and without any hindrance. The third giant, Facebook, wants to give people the possibility to engage in friendships, create communities and make the world more open and connected. But the bottom line for all three tech giants depends on the same elementary business model: targeted, relevant advertising.

Google’s main source of income is the ads placed at the top of search results, next to the search results and on the searched websites. At the congressional hearing in April 2018, in front of a rather slow-witted Senator, Mark Zuckerberg made it clear: “Yes, Congresswoman, we run ads. That’s the business model […].”³ That was not how things were to begin with, but in 2008, Facebook recruited Sheryl Sandberg from Google as second in command. In hiring her, Facebook got a hard-working ad broker who finally managed to crack the code and radically expand the company’s advertisement foundation. According to international media agency GroupM, Facebook and Google together now control 84% of the digital ad revenue in the US, giving them a de facto duopoly on the advertising market.⁴ And at Twitter, ad revenue makes up roughly 86% of the total income of the company.⁵

The core of the three companies’ business models makes them ad brokers rather than the IT companies of their origin. Their main task is to monetize their services by the use of very detailed user profiles to capitalize on the opportunity to accurately target advertising to highly specific and selected user groups—characterized by anything from age, race, gender, housing, income, looks, religion, health, location, psychology, problems, hobbies, detailed music preferences, literature, movies, food, make-up, anything. The advertisers are the real clients of the tech giants, and the user is but the product. With this massive monitoring of user behavior, both online and offline, companies can make accurate and automated predictions about what ads, news and political messages users are most susceptible to, and about what content will make them click, like, share and scroll down the endless feed. Whether or not tech giants in fact sell data is a matter of semantics. If you claim that they do, they furiously refer to the fact that they do not hand over user data to advertisers and they might file a court case against you for libel. But they do sell data in the sense that they give advertisers data-based access to hyper-accurately target their advertising.

In an academic article written for Stanford University back in 1998, Google founders Larry Page and Sergey Brin explain the problems involved when combining search engines and advertising: “The goals of the advertising business model do not always correspond to providing quality search to users. […] we expect that advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers.”⁶ It was meant as a criticism of the advertising setup of the day, which—lacking targeted focus—would bombard users with as many ads as possible. Instead, Page and Brin wanted to target their efforts at specific users with fewer and increasingly customized ads.⁷ No doubt, it was a significant improvement for users, not to mention advertisers. Combined with new algorithms that quickly sorted through what was and was not useful online, the search engine became a success. But the two partners did not pay attention to their own criticism as they suddenly faced the choice between what was good for the users versus what was good for the company. The popular and highly admired PageRank algorithm was modified and instead, from 2009, personalized searches became the core strategy, resulting in more opportunities and skyrocketing interest from advertisers. Today, Google together with Apple, Microsoft, Amazon and Facebook are the top five biggest corporations in the world.⁸

When Facebook, Twitter, Google and others offer free services for users, there are two sides to this. From a user point of view, it is obviously attractive to have a good and well-organized service for information and friend contacts, and not to mention free. Speechifying in flowery terms, the CEOs of these companies boast that their real purpose is to assist users and improve the world. However, from an economic point of view, the primary motive is to entice potential users into a digital panopticon, one which is incredibly skilled at harvesting data.

Users do not really sense they are being watched. They may be surprised once in a while, when they find a remarkably relevant ad on their screen. The platforms present themselves as networks of unlimited freedom and communication, offering nothing more than posts, search results, tweets and comments. But let us for a moment zoom in on the example of Facebook. The company registers every single click, every “like” and every comment. However, the collection of data is not limited to content the users explicitly choose to upload (and which they recently, after numerous protests, have the opportunity to remove). It also collects browser history, what the user has been doing on Facebook and outside Facebook, and achieves external personal data such as geographic, financial, medical and other information about the user by purchasing large amounts of data from lesser-known but huge data broker companies, e.g. Acxiom, DataLogix, Epsilon and Experian (European countries have laws to partly block data brokers, especially since May 25, 2018, when GDPR came into force). These are large but not very publicly-known companies, which, ever since the 1960s, have provided detailed personal data to an already extensive industry of targeted traditional paper advertisements by mail. These players now gained new digital customers. After the Cambridge Analytica scandal of early 2018, Facebook declared that the company would stop brokering such personal data, but not that the company would eliminate already brokered data.⁹ The company even creates so-called shadow profiles of people not yet on Facebook, if all they have done is contact someone who is indeed on Facebook, or if they visit a website featuring a like and share button.¹⁰ At the congressional hearing, Zuckerberg was forced to admit that the company follows user movements around the Internet, even when they are logged out of their Facebook accounts.¹¹ With this massive surveillance, tech giants have the opportunity to gather large amounts of raw data on users and translate the data into personal information. At the congressional hearing, Zuckerberg was unable to give an estimated average amount of data per user, but it is estimated that Facebook has developed more than 52,000 categories for organizing these data.¹² The data broker Acxiom, one of Facebook’s biggest business partners, is said to have 1500 data points on each and every one of more than 500 million individuals across the globe, most of them in the US.

These data are categorized and converted into information and thus become a valuable asset for advertisers trying to predict, direct and manipulate consumer behavior. In the digital panopticon, the user becomes a commodity—in the form of a data package—to be processed and negotiated for financial use. Big Brother and Big Business have formed an unholy alliance. The detailed personal data are not forwarded to the advertiser, but it is the access to users by way of these data that is being sold. In that way tech giants can claim that they do not sell data—and even sue anyone who claims that they do.

An interesting aspect to the digital panopticon is that it comes off as a friendly kind of power. No explicit threat, command, prohibition, or physical punishment is involved. On the contrary, the panopticon is based on positive benefits and convenience: free communication, information, community, knowledge sharing and searching. The worst sanction a user can meet is exclusion: without accepting the terms of service (that is, without surrendering data or complying with the community standards), access is denied. This is a very subtle form of coercion, based on the user’s fear of missing out. This seeming kindness has turned out to be an incredibly effective strategy for harvesting data. In the eyes of the tech giants, it is a question of connecting with people’s psyches. This kind power motivates the user to constantly disclose, share and participate, to communicate points of view, wishes and preferences. According to German cultural philosopher Byung-Chul Han, this “friendly” power is to blame for the user’s crisis of freedom: this is not freedom repressed, but freedom exploited.¹³ In order to gain convenience, users subject themselves to a form of self-surveillance, because they willingly hand over their information to the platforms and thus also indirectly to the hungry eyes of the advertisers. In 2019, professor Shoshanna Zuboff developed an analytical conception of the business model – “surveillance capitalism”.¹⁴

Usually, involuntary surveillance is seen as a problematic or even criminal invasion of privacy. But in the digital panopticon, the surveillance takes place voluntarily because it is considered a mutual advantage – a partnership. This is what makes this form of surveillance so effective and relatively uncontroversial. But in a sense, this partnership is an illusion. The users take part without real informed consent. They may have accepted the terms of service, but they never know how they are categorized and measured when profiled and resold to advertisers. The users participate without knowing exactly what they are participating in. Once consent is given, the tech giants are free to continuously modify the terms of service—for instance they can change the algorithms without having to ask the user to acknowledge such modified conditions.