The Child-Privacy Law That Made Kids' YouTube Worse

Based on the research ofJohnson, Lin, Zhong, and Cooper, "COPPAcalypse? The YouTube Settlement's Impact on Kids' Content," Management Science, 2026

A $170 Million Settlement, and a Compliance Deadline

In September 2019, the Federal Trade Commission and the New York Attorney General announced a record settlement with Google and YouTube over alleged violations of the Children's Online Privacy Protection Act (COPPA): $136 million to the FTC and $34 million to New York, totaling $170 million, at the time the largest penalty the FTC had ever obtained in a COPPA case. The complaint centered on YouTube's practice of collecting persistent identifiers and serving behaviorally targeted ads on channels that were, by any reasonable read, aimed at children, without the verifiable parental consent COPPA requires.

The settlement's conduct relief, not just the fine, is what reshaped the platform. Starting in January 2020, YouTube required every channel owner to designate each video as "made for kids" (MFK) or not, and it began treating anyone watching an MFK video as a child by default, regardless of their actual age. For content marked MFK, YouTube disabled personalized advertising, comments, notifications, end screens, info cards, live chat, and the ability for viewers to save videos to playlists, the standard toolkit platforms use to keep an audience engaged and coming back. This is documented in YouTube's own creator communications and Help Center guidance issued around the January 2020 rollout, which laid out exactly which features would be switched off for kids' content and why.

That is the regulatory event Garrett Johnson, Tesary Lin, Liang Zhong, and James Cooper use as a natural experiment in "COPPAcalypse? The YouTube Settlement's Impact on Kids' Content," forthcoming in Management Science. Comparing 5,066 top American channels, MFK creators against otherwise similar non-MFK creators, in a difference-in-differences design around the January 2020 cutover, they isolate what happens to a content ecosystem when a platform is legally forced to strip personalization from one segment of it.

What Actually Got Turned Off

It is worth being precise about the mechanism, because the intuitive story ("ads got worse, so revenue got worse") undersells what changed. Personalized advertising was the headline casualty, and it mattered financially: behaviorally targeted ads typically command a premium over contextual ones, so MFK creators took an effective pay cut on every impression. But the settlement also switched off the discovery and retention machinery built into the platform, subscriber notifications that pull viewers back for a new upload, "save to playlist" that keeps a channel in a viewer's rotation, comments that build community and signal relevance to the recommendation system. Losing personalized ads reduced the per-view payout; losing notifications and playlists reduced the number of views in the first place. Both channels of harm point the same direction, and the paper's demand-side result, a 20% drop in views for MFK channels, is consistent with both operating at once.

Creators Saw It Coming, and Said So

The reaction was not quiet. Coverage from late 2019, including reporting in Variety, described creators as "worried and confused" about how the FTC's new rules would be enforced and whether their existing catalogs would be swept into the MFK designation regardless of intent. Trade press covered the FTC's rulemaking process drawing more than 175,000 public comments, an unusually large response reflecting how directly creators felt threatened. Outlets tracking the creator economy, Tubefilter among them, used the term "COPPAcalypse" months before the academic paper adopted it, warning that channels could be effectively demonetized in a single policy change many had not budgeted for. None of that reporting could tell you, at the time, whether the fear was proportionate to the eventual damage. The paper is the first rigorous answer: yes, roughly proportionate, and worse on quality and equity dimensions nobody was tracking in real time.

The Numbers Behind the Backlash

The paper's core estimates give the backlash a number. MFK creators produced 18% less content after the policy took effect, and what they did produce tilted away from kids' programming entirely, creators pivoted toward non-MFK content, where the old monetization tools still worked. Among the content that stayed MFK, quality slipped on two measurable fronts: the share of original (non-repurposed) content fell 9%, and manual closed captioning, a feature that costs creator time or money to produce and pays off mainly in accessibility for deaf and hard-of-hearing viewers, not in ad revenue, dropped 28%. Viewer ratings on MFK content fell 9%, and, as noted, MFK channel views fell 20%. Layered on top of the supply and demand contraction, the paper documents rising concentration: both content creation and viewership shifted toward the largest incumbent channels, the ones with brand-deal revenue, cross-platform distribution, or scale enough to absorb a personalized-ad shortfall that would sink a smaller creator.

Pulling ad personalization from a content class doesn't just dent revenue, it shrinks the class and hands what's left to whoever can survive without that revenue.

Put those pieces together and the paradox comes into focus. COPPA's conduct relief was aimed squarely at protecting children from covert data collection, and on that narrow metric it plainly worked, MFK content is no longer behaviorally targeted. But the same intervention removed the commercial and product incentives that had been drawing creators into making kids' content, cataloging it well, and captioning it for accessibility. The children the rule was meant to protect ended up with a smaller, less original, less accessible, more homogenous supply of programming, produced by a narrower set of larger creators.

The Pattern Is Bigger Than YouTube

COPPA's YouTube settlement is not a one-off. It sits inside a broader regulatory wave aimed at how platforms treat young users: California's Age-Appropriate Design Code Act (AB-2273) was signed into law in September 2022 and set a compliance horizon of July 2024, applying design and default-privacy obligations to any online service "likely to be accessed" by minors, a scope far broader than COPPA's child-directed-content trigger. The direction of travel across US states and other jurisdictions is the same: restrict data use and personalization for young users by default. The YouTube case is the first setting where researchers can actually measure what that restriction does to the supply side of a content market once ad personalization is switched off for an entire content class, rather than assume the answer.

What This Means for Anyone Designing Platform Rules for Kids

The uncomfortable lesson is not that child-privacy regulation is a mistake, the underlying behavior COPPA targeted, tracking children for ad targeting without consent, is hard to defend on its own terms. The lesson is that removing personalization from a market segment is not a narrow, costless fix; it is an intervention in that segment's entire economics, and content markets respond to economics the way any other market does. When you make made-for-kids content less profitable to produce, fewer people produce it, the people who keep producing it invest less in the parts that don't drive revenue (like captioning), and the segment consolidates toward whoever can operate at a loss on personalization the longest. A rule optimized purely for the privacy harm it targets, with no accounting for the supply-side response, can leave the population it was designed to protect with a worse product. That is the finding worth sitting with the next time "protect the children online" legislation gets drafted without a model of how creators, not just platforms, will respond.