haven’t they claimed that their “AI” catches 99% of spam before it gets seen or something? seems like there’s even lower hanging fruit or they’re just lying.
I would bet money Facebook does very little fraudulent advertising filtering, since they are explicitly incentivized not to.
But "We block 99% of fraud/spam" is an outright useless and purposefully non-informative datapoint. It doesn't tell you anything. As a statement, it is equally true both for the SuperGoodMerchant who lets through a single fraudulent advertisement a year, and it is also true for UltraScamAds who let through a million scam ads a day. It depends entirely on the rate of fraud attempts compared to the rate of legitimate requests.
An actual useful metric if you want to avoid fraud or scams would be, what percentage of ads do they run that turn out to be fraudulent?
reply