Meta loses $16B to scams, Reuters

Meta’s scam-ad problem matters because this is not just about a few bad posts slipping through. The core allegation is that Facebook and Instagram kept making huge amounts of money while fraudulent ads kept flowing. Reuters got hold of internal Meta documents that framed the scale in blunt financial terms — about 10% of 2024 revenue, or roughly $16 billion, tied to scam ads and banned goods. That is the part that changes the story. It stops looking like a moderation miss and starts looking like an incentives problem. (msn.com) ### What actually was in the documents? The headline number was the big one — Meta internally projected that scam ads and ads for banned goods would generate about $16 billion in 2024. Separate documents broke out a narrower bucket called “higher-risk” scam ads and put that at about $7 billion in annualized revenue. One December 2024 document also said users were seeing an estimated 15 billion of these higher-risk scam ads every day. (msn.com) ### Why is 15 billion a day so alarming? Because that number suggests scale, not leakage. “Higher-risk” did not mean proven fraud in every case, but ads that showed strong signs of being fraudulent. If users are seeing that many suspicious ads every day, the issue is not that the filter occasionally misses. Basically, the filter is letti(msn.com)ting against them for attention and inventory. (abc.net.au) ### Why didn’t Meta just ban them? The reporting says Meta often required its systems to be at least 95% certain an advertiser was committing fraud before banning that advertiser outright. For advertisers below that threshold but still considered suspicious, Meta could charge higher ad rates — a “penalty bid” approach that was supposed to deter bad actors. But t(abc.net.au) fast. (marketingbrew.com) ### Why does the ad system make this worse? Because ad targeting can reinforce bad outcomes. The documents described users who clicked scam ads as more likely to be shown more of them later. That is the ugly flywheel here — the system reads engagement as a signal, even when the engagement came from fraud bait. So the same machinery that makes ads efficient for normal businesses can also help scammers find the most vulnerable people. (marketingbrew.com) ### What is Meta saying back? Meta has pushed back hard. The company says the documents give a selective and distorted picture of how it handles scams, and it says that audits found many ads in the estimate were not actually violating its rules. Meta has also said user reports of scams fell 58% over the prior 18 months. So the company’s defense is not “there is no problem.” It is “the Reuters numbers overstate it, and enforcement is improving.” (wtsp.com) ### Why is this becoming a legal and political problem? Because once an internal estimate puts a dollar figure on scam revenue, regulators and plaintiffs have something concrete to point at. Senators Josh Hawley and Richard Blumenthal asked the FTC and SEC to investigate. New lawsuits have also started t(wtsp.com)form.” (blumenthal.senate.gov) ### Why should advertisers care? Because if scam ads stay in the auction, they can distort prices, placements, and brand safety. A clean ad marketplace is part of what advertisers think they are buying. If suspicious advertisers can keep bidding — even at premium rates — then the platform is not just failing to protect users. It may also be degrading the product sold to legitimate brands. (searchenginejournal.com) ### Bottom line? The real story is not whether every dollar in that $16 billion estimate was perfectly counted. It is that Meta’s own paperwork appears to have treated scam exposure as a measurable, material revenue stream. Once that idea is out in public, the company has a trust problem on three fronts at once — users, advertisers, and regulators. (msn.com)