A few factors contribute to this huge disparity in figures. There is currently no one way to detect whether an ad impression is fraudulent or not, and anti-fraud vendors rely on different methodologies and verification technologies for their work. Depending on what methodology is used, a vendor may say an ad impression is fraudulent or they may not. Firms also use technologies proprietary to them, and no identical results are being produced across the industry.
“Anyone who is detecting fraud is still only detecting the types of fraud that they’ve figured out how to detect,” said eMarketer principal analyst Nicole Perrin.
On the advertiser’s end, not all marketers are getting all their impressions verified. For companies that can afford to do so, they may only be checking a percentage of impressions and then that gets extrapolated into a larger picture. Smaller advertisers don’t have the budget to use verification services, so those numbers are not captured at all.
For fraud impressions that are captured, there’s no standardized way to translate that into a dollar figure, as impressions are all priced differently. “You have to make assumptions and estimates," Perrin said. "Is the average fraudulent impression the same price as the average real impression? Does it cost more or less? There are many reasons why the answer could be any of those things."
The takeaway for marketers is that no matter which ad fraud estimates they look at, billions of dollars are still being lost every year. If an individual marketer is using verification services, they can feel confident that they’re doing everything they can to stay safe, but the reality is that not all marketers are doing the same.
“Even with the most optimistic interpretation, marketers are throwing billions of dollars away," Perrin said. "So any marketer should be concerned about that—or at least concerned enough to really think about, ‘Am I pushing my agency to buy impossibly cheap inventory? Am I asking for a media plan that can only be fulfilled with fraud?'"