Meta Ad Integrity: Why Enforcement Matters and How to Measure It

You are a product manager at Meta. Explain why Meta needs to act on ads that break its community guidelines or the law, and how you would measure whether that work succeeds. ### Clarifying Questions - Should I cover every place Meta serves ads (Facebook, Instagram, Messenger and ads shown in partner apps), or one surface? - Do policy violations, such as scams or hate speech, and legal violations, such as country-specific rules on political or financial ads, belong in one measurement framework or two? - Is the measurement about the whole enforcement system (review before an ad runs, removal after, penalties for advertisers, appeals) or only about detecting bad ads? ### Part 1 — Why enforcement matters Argue why acting on violating ads is critical rather than optional, for the people who see ads, for the businesses that buy them and for Meta. ```hint Follow one bad ad Trace what a single scam or illegal ad does to the person who sees it, to the honest advertisers competing for the same impression, and to the platform that served it. ``` #### What This Part Should Cover - Harm to people who see the ads and the effect on their trust in ads generally - Effects on legitimate advertisers and on the long-term value of the ads business - Legal, regulatory and reputational exposure that differs by market - The tension between revenue from violating ads and long-term platform health ### Part 2 — Success metrics Propose how you would track success: a primary metric, the metrics that drive it, and guardrails. ```hint Stress-test each number For every candidate metric, ask what it would do if abuse grew while enforcement stayed the same, and what it would do if enforcement improved while abuse stayed the same. ``` #### What This Part Should Cover - A primary metric that reflects what users are actually exposed to, with the reason it beats simple counts - Supporting metrics for detection, speed of action, advertiser penalties and regulatory compliance - Guardrails against over-enforcement: wrongly rejected ads, appeals and the experience of legitimate advertisers - How each metric is measured when most violations are never reported by users ### What a Strong Answer Covers - A rationale that ties enforcement to user trust, advertiser value and legal exposure, not to moderation for its own sake - A metric set that balances catching bad ads against blocking good ones - Recognition that user reports and takedown counts are biased signals - Segmentation by violation type and severity, country and ad surface ### Follow-up Questions - Your primary metric improves sharply in one quarter. How would you check whether ads really got cleaner rather than detection getting weaker? - Advertisers complain that pre-launch review delays their campaigns. How would you decide how many ads to hold for review before they run? - A new law in one country requires certain ads to be removed within a fixed time after notice. What would you add to your metrics and process?