Featured Article:
by Henk Kox, Bas Straathof, and Gijsbert Zwart
Internet firms (online newspapers, video platforms, online games) earn revenues not only by charging subscription fees to their consumers: a significant part of their income derives from selling advertising space. Internet advertising turnover has been growing fast, and is now comparable in size to total advertising revenues from broadcast television.
An important feature of internet advertising is that it can be targeted towards specific consumers. Whereas for traditional newspaper advertising, all readers are shown the same ads, on the internet the ads you see will often be related to the websites you visited previously, the e-books you browsed or the flights you searched. This makes those ads potentially more relevant to the viewer, and leads advertisers to pay higher fees for showing them. However, it may also lead to unease for consumers who are concerned about invasions of their privacy.
In this paper, we explore internet platforms’ strategies in setting both ad targeting intensities – how much user information is collected to tailor the ad to the user – and user prices. We model a market with competing platforms, e.g., multiple online newspapers vying for subscribers. Showing more focused advertisements increases a newspaper’s ad revenues and helps the platforms to keep consumers’ subscription fees low, in a bid to attract more consumers, but it also reduces welfare for those consumers who value their privacy.
A new feature of our analysis is that, in an imperfectly competitive market, allowing more targeted ads also heats up competition among platforms. In traditional newspapers, ads will be tailored to the average consumer’s tastes. Any additional consumers won by lowering subscription prices will likely have a taste that is less aligned with those of core readers, and showing those same ads to these new consumers will be less effective. Advertisers will anticipate that a larger audience means a more heterogeneous audience, and they will be less willing to pay for showing their ads. In contrast, with targeting, advertisements to those new consumers are better matched to their tastes, so that ads shown to new consumers can be as valuable as those shown to the core audience. This means that the online platforms have a greater incentive to compete for new consumers and to win the increased advertising revenues these consumers bring.
With more targeting, fiercer competition reduces both consumer prices and firm profits. In spite of those lower profits, we argue that the industry will choose to increase targeting intensities, as long as they have difficulty convincing consumers that their personal data will not be used for targeting. Consumers gain from lower prices, but may lose overall if they value their privacy highly. In that situation, policies that help platforms commit to privacy standards increase welfare.
The following section is a Q&A with Drs. Kox, Straathof, and Zwart
- Please briefly describe the main results of your paper. For what types of websites or situations are these findings most applicable?
Targeted advertising can benefit consumers through lower prices for access to websites. The reason why websites will lower prices is that they compete more strongly for consumers if they can target advertisements to consumers that do not belong to the core audience of the website.
Yet, if consumers dislike that websites collect their personal information, their welfare may go down. Websites could refrain from showing targeted advertisements to those consumers willing to pay a higher price in return for privacy, but, as long as websites cannot convince consumers that they will not collect personal data, consumers will opt for accepting targeted advertising and paying the lower price. A policy that gives consumers control over the use of their personal data will then raise welfare.
- Are there any particular aspects of the data that should be noted?
Our article is not an empirical paper, so we have not ‘struggled with the data’. However, it is clear that, as soon as public standards for privacy are set, public authorities will have a challenging job in surveilling whether those privacy standards are actually complied with. In this business, the actual information about targeting activities is of a commercial and strategic nature, and mechanisms for reliably verifying that information can be a useful direction for research.
- How did you become interested in the topic of targeted advertising and privacy in relation to competition among internet firms?
Targeted advertisements are the main source of revenue for some of the most valuable firms in the world, like Alphabet and Facebook. At the same time, people are increasingly concerned about how firms treat their personal data. This raised our interest in how government privacy policies might affect welfare.
- How does your analysis build upon and differ from previous explorations of this topic?
The analysis builds on earlier work on competition between media, mainly newspapers and television stations, showing different advertisements in different geographical regions. Our analysis allows for advertisements tailored to specific consumers and studies the impact on competition among firms, and the potential welfare effects of privacy policy in that setting.
- In this paper, you suggest that “since consumers have heterogeneous costs of privacy loss, one may do better by allowing websites to differentiate the levels of targeting.” What might a “low” vs. “high” level of targeting look like in practice?
Low-level targeting is essentially what traditional printed media do: the New York Times prints advertisements that match well with the tastes of that newspaper’s typical reader. Similarly, online, low-level targeting could be based on the content of the webpage being viewed, the country from which the consumer visits the website, or voluntarily supplied user data. High-level targeting includes the current practice of maintaining user profiles based on past internet usage.
6. What directions might you suggest for further research on the topic of privacy?
Our paper studies the interaction of competition and privacy policies and uncovers one mechanism by which privacy choices impact on competition intensities. This is a direction that, we think, can be taken further. How do internet firms compete for consumers that are heterogeneous not only in how they value the products on offer, but also in how they value their privacy? In online markets where market power is important, how do firms distort privacy choices to increase profits? Also, what mechanisms can firms use to commit on their privacy policies? Do we need laws to enforce privacy protection, or can the market provide reliable commitment mechanisms?