As digital advertising evolves, the biggest question may not be how platforms capture consumers’ attention, but how that attention translates into advertising value—and at what cost to privacy. In the latest Insights@Questrom Q&A, Shunto J. Kobayashi, Assistant Professor of Marketing at Boston University Questrom School of Business, examines Meta’s settlement over engagement-maximizing design and what it reveals about the economics of digital advertising, the value of targeted advertising, and whether new technologies can preserve advertising effectiveness while protecting consumer privacy.
Meta’s settlement over allegations that its platforms were designed to maximize engagement among young users raises a broader question for digital advertising: when does engagement become a liability rather than an asset?
Meta’s settlement with a group of 52 attorneys general over allegations concerning engagement-maximizing design and consumer privacy for children is not an isolated case. In the United States alone, states like California have already enacted laws restricting certain platform features for minors. Beyond the United States, Australia now requires major social media platforms to prevent children under 16 from holding accounts, and the European Commission has preliminarily found Meta in breach of the Digital Services Act over features of Instagram and Facebook that it considers addictive. Policymakers’ concerns also go beyond children’s use of social media to include the effects of such features on adults. These developments highlight increasing pressure to hold platforms accountable for the potential consequences of platform designs optimized to encourage engagement.
Because platforms like Meta monetize users’ screen time by running ads, it’s natural that competition drives both engagement-maximizing design and data collection. They compete against each other for consumers’ limited attention, but also against other channels like the open web and linear/connected TV. This competition encourages features like infinite scrolling and personalized recommendations, which policymakers argue can be addictive. Personalization can also rely on extensive collection of behavioral data, raising privacy concerns. The rise of AI platforms like ChatGPT is intensifying this competition as consumers spend increasing amounts of time interacting with AI assistants. Indeed, AI companies like OpenAI are beginning to monetize attention through advertising, while incumbent platforms like Meta are responding by introducing AI assistants themselves, such as Meta’s Muse. The rise of generative AI has also brought “AI slop” into social media, raising additional concerns about misinformation produced to attract engagement.
What I find interesting about Meta’s settlement is that, of the approximately $18 billion in payments, 30% (about $5.3 billion) is contingent on YouTube and TikTok both adopting specified youth protections and making matching payments. This illustrates the difficulty one platform faces in restricting engagement on its own: users may spend less time on that platform or simply migrate to competitors that do not face the same restrictions. It will be interesting to see how YouTube and TikTok respond, and whether they adopt or are required to implement constraints similar to those Meta agreed to, such as daily time limits and nighttime restrictions.
For years, platforms have treated engagement as a key driver of advertising value. Does more engagement necessarily mean more effective advertising?
Social media has been remarkably successful in attracting advertising spending. According to EMARKETER, social media platforms capture nearly 30% of U.S. advertising spending despite accounting for only 12.5% of adults’ daily media consumption. This suggests that advertisers place a higher value on advertising on social media per minute of consumer attention than on other channels, such as television.
There are several reasons why advertisers may find social media particularly effective. Infinite scrolling, frictionless swiping, and personalized recommendation algorithms keep users engaged. Furthermore, targeting algorithms and ad auctions help match advertisers with users. Native advertising blends advertisements into organic content, while influencer marketing further integrates commercial messages into users’ experiences. The combination of engagement, targeting, and advertising formats helps explain social media’s success in attracting advertising spending.
How valuable is targeted advertising, and how much does more precise targeting actually improve advertising effectiveness?
My research illustrates how more precise targeting helps digital publishers increase their advertising revenue. In a paper published in the Proceedings of the National Academy of Sciences (PNAS), I study how removing third-party cookies affects the advertising revenue publishers receive. Even without cookies, advertisers can target ads based on the content of the webpage, known as contextual targeting. However, cookies allow advertisers to target consumers more precisely based on their browsing behavior across websites. The paper, co-authored with Zhengrong Gu, who was then a graduate student at Questrom, and Garrett Johnson, an associate professor of marketing at Questrom, finds that removing third-party cookies reduces publisher advertising revenue by roughly a third. This finding comes from a large-scale field experiment jointly overseen by Google and the UK’s Competition and Markets Authority. Another paper of mine, forthcoming in the Review of Economic Studies (co-authored with Miguel Alcobendas, Matthew Shum, and Ke Shi), reaches a similar conclusion using a different dataset and a different methodology.
The next question is how valuable targeting is to advertisers themselves. In a working paper with Zhengrong Gu and Garrett Johnson, I study the effect of targeting on advertiser outcomes. We specifically study retargeting, which involves showing ads to consumers for products they have previously browsed, typically using third-party cookies. We find that retargeting increases advertisers’ conversions by roughly 5% relative to a baseline without retargeting. This is an incremental effect, meaning the causal impact of retargeting compared with not using it. This finding further highlights that targeting provides value not only to publishers but also to advertisers.
As tracking becomes more limited, could advertisers be pushed to focus less on clicks and more on whether ads actually change consumer behavior? And how can they measure those changes in consumer behavior?
Consumer behavior is very noisy, and click-through rates tend to be low. This makes it difficult for advertisers to optimize campaigns based on other consumer actions, such as conversions and purchases, which are even less frequent than clicks. Furthermore, measuring these actions has traditionally relied on tracking technologies such as third-party cookies. For instance, without cookies, it becomes harder to match an ad impression to a subsequent conversion. This makes it harder for advertisers to evaluate campaign outcomes and optimize their campaigns, such as comparing different ad creatives and advertising intermediaries. However, the industry is developing alternative methods to preserve advertising capabilities while relying less on third-party cookies, including server-side conversion APIs (such as Meta’s Conversions API, or CAPI), data clean rooms, and other privacy-enhancing technologies.
If privacy protections reduce the value of some forms of digital advertising, who ultimately bears the cost— platforms, advertisers, publishers, or consumers?
My research discussed above indicates that privacy protections implemented through the removal of third-party cookies reduce advertising revenue for publishers and platforms and advertising effectiveness for advertisers. In the short run, consumers may not be directly affected, but we have to remember that many digital businesses are funded by advertising. It’s easy to take for granted that many websites and platforms are free to use, but advertising revenue is what makes this possible. If significant value is lost from digital advertising, websites may turn to subscription-based models, potentially pricing out many users, or simply go out of business. Subscription-based models may be especially problematic for social media platforms, where users benefit from having their friends on the same platform. Lower advertising revenue could also discourage new businesses from entering the market.
Can new technologies preserve consumer privacy without sacrificing advertising effectiveness?
The paper discussed above with Zhengrong Gu and Garrett Johnson on advertising effectiveness speaks to this issue. Privacy-enhancing technologies (PETs) are designed to preserve key advertising functionalities while limiting exposure of individual consumer data, using methods such as differential privacy and on-device computation. In particular, these technologies can help sustain targeting and measurement, which have traditionally relied heavily on third-party cookies.
Our research studies a privacy-preserving technology for retargeting called the Protected Audience API, developed under Google’s Privacy Sandbox. We find that, compared to cookie-based retargeting, its advertising effectiveness per dollar spent is statistically indistinguishable. However, its overall effectiveness is substantially lower, generating only about 40% of the incremental conversion lift achieved by cookie-based retargeting and about half as many click-through conversions. Our findings suggest that this gap is largely due to limited industry adoption, particularly on the publisher side, which restricts the opportunities for advertisers to show ads using the technology. This is also consistent with our publisher-side study.
Google announced in October 2025 that it would retire much of Privacy Sandbox, citing limited industry adoption as one reason for the decision. Nevertheless, I believe that privacy-enhancing technologies still have a vital role as privacy protections become more widespread. Indeed, the World Wide Web Consortium (W3C) is developing an interoperable standard for privacy-preserving advertising measurement, which could help address the industry adoption challenge.
