A lawsuit claims that Facebook inflated video metrics to lure advertisers. Facebook denies, but it’s worth noting that it would be much easier for Facebook to defraud advertisers than for any of its predecessors in traditional media.
John Wanamaker (1838-1922), one of the first department store entrepreneurs in the US, is often credited with the phrase: “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” Marketers always needed to rely on advertising, but they knew that a big part of the audience they reached was irrelevant for their products and services.
Wanamaker would probably be thrilled to learn that, a century after his death, marketers don’t need to waste money on the wrong audiences; internet advertising enables marketers to pay for performance and target relevant audiences. The sophisticated targeting and performance measures tools available on the internet look like the utopian world marketers have waited centuries for.
But did we eliminate the waste? Probably not. The targeting revolution has given marketers efficient marketing tools and created a whole category of new products and services, but also brought with it some phenomena that might turn out to be more wasteful and harmful to society than the waste of Wanamaker’s era.
Last week, a group of advertisers filed a class action lawsuit against Facebook, claiming that the social media giant committed fraud by knowingly inflating its reports regarding the time users spent watching videos for over a year, causing modern Wanamakers to spend more on ads. To be sure, inflating numbers was not invented by the giant digital platforms—newspapers have been inflating their circulation numbers for years—but what is striking about the Facebook lawsuit is the degree to which the company is accused of inflating the numbers: up to 900 percent.
While Facebook denies the allegations and will fight them in court, it’s worth noting that, unlike traditional media outlets, it is much easier for Facebook to defraud advertisers—without any retaliation—than for any of its predecessors: Facebook and Google are gradually becoming a de facto duopoly in the targeted online advertising market, and their advertising and targeting algorithms are totally opaque—both to the advertisers and to users.
In the US, Facebook’s largest market, fast user growth has all but stopped—and yet the company is still showing significant growth in profits. The reason is probably Facebook’s market power. According to data from AdStage, an ad tech startup that is one of Facebook’s official marketing partners, in January 2018 Facebook CPMs—the cost of a thousand ad impressions—were up 122 percent year over year. In February, CPMs were up 77 percent.
The network effect at the heart of their business model has given digital platforms an unprecedented amount of data on the profiles and behaviors of billions of users. Targeting has never been more sophisticated and impactful. One obvious question is who is going to capture the benefits of these ever more efficient targeting tools: the advertisers and users, or monopolistic platforms? The second question may turn out to be even more important: Does targeting have negative externalities to society that dwarf Wanamaker-style waste?
Up until last year, one could go to Facebook’s corporate website and find “case studies” and PR boasts about how efficient its targeting algorithm is as a tool for marketers and political parties. In 2014, Facebook published a psychological study that showed it can manipulate users’ emotions by changing the composition of their newsfeed. In 2015, Facebook bragged about its ability to influence elections, publishing a case study that showed it helped the Scottish National Party “achieve an overwhelming victory” in the UK general election.
Facebook is now publicly underplaying its algorithmic ability to modify the feelings and behaviors of its users. The “case study” on the UK election was removed from Facebook’s corporate site following the Cambridge Analytics scandal. Even the tone-deaf Mark Zuckerberg who announced, 10 years ago, that users don’t care about privacy anymore, now understands that the world has changed.
But the company is only starting to address the potential behaviors and harms that its business model encourages:
- Unlike traditional media outlets, Facebook has not had enough incentives to limit fake, sensational, polarizing content. While traditional media need to balance their quest for more attention and revenues with their reputation and legal liabilities, Facebook enjoys immunity from both. The fake or polarizing content on your Facebook news feed is perceived by users as “neutral”—after all, it’s what your friends shared—although the way Facebook’s algorithms are designed encourages a specific kind of content. Section 230 of the Communications Decency Act immunes Facebook from legal liability for content that is published on its network.
- Traditional media usually had one edition, seen by all its readers/viewers. The diversity of the audience limited the incentives of the most trusted publications to push content that is polarizing and sensational. Facebook, on the other hand, has an unlimited number of “editions” and can offer each reader a worldview that is in concert with her views, biases, tastes and even weaknesses. Targeting is changing the public discourse in ways that we never envisioned and that we still might not fully comprehend.
- One of the most important roles of the news media is to hold the powerful to account by exposing their activities and shaming them when they are involved in corruption, or when they promote policies that cater to their allies. This is done by informing the public at large and by “signaling” to elites and decision makers in the public and private sector. News media informs them that the public is informed, which can help shape norms and limit bad behavior. But if we gradually move toward a world where most readers are treated by algorithms as consumers only, each one getting their own “edition” of reality, then reaching common ground and common fact will become increasingly difficult.
- In the past, digital platforms were hailed as “democratizing” technologies that contribute to a more pluralistic discourse. This is partially true, but it is still the traditional media that does most of the heavy lifting—investigative and accountability journalism—that is needed to inform the public. The costs of this work have remained in traditional media, while the revenues have moved to the digital platforms.
- Destroying the business model of the traditional news media is not the only way the platforms influenced the information ecosystem. Newsrooms that saw their distribution gradually move to the platforms now increasingly produce news, headlines and agenda that have the potential to be more viral. We need more research and evidence on this, but it would be surprising if this shift has created more incentives for expensive, independent and accurate investigative work.
While advertising efficiency is easy to measure and compare—the social costs of a world where most people are constantly targeted by identified and unidentified interested parties—politicians, foreign countries and businesses are much more complicated. The negative externalities are spread over hundreds of millions of people.
The string of scandals around the world involving the platforms has created demand for regulation. But before we go that path, first we have to demand that platforms open their black box algorithms and data hoards to scrutiny, so we will have a better understanding of what these machines are doing or are capable of doing.
Disclaimer: The ProMarket blog is dedicated to discussing how competition tends to be subverted by special interests. The posts represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty. For more information, please visit ProMarket Blog Policy.