Attribution is tricky at the best of times and worse when you're dealing with many, small campaigns rather than a few discrete large campaigns. Let's say you m

Attribution is tricky at the best of times and worse when you're dealing with many, small campaigns rather than a few discrete large campaigns. Let's say you make surfboards that are primarily bought offline at surf shops. You run a campaign with 100 instagram influencers in the San Diego area and you see a 15% greater uplift in sales compared to nationwide. Which influencers contributed meaningfully and which ones were fake accounts where you wasted your money? How much of that 15% uplift is because of the influencer campaign vs a longer summer in San Diego this year vs some totally organic consumer spreading it via WOM? Even if you want to design an experiment, summer's already over and you'll need to wait until next summer to test out your hypothesis. It's not hard to see how a scammer could cobble together a dozen or so of these different advertisers and make a substantial sum out of fraud consistently without ever getting caught.