Like any ordinary business, fraudsters flow to where the money is. And because advertisers are putting more of their budgets into video, fraudsters are flocking to the platform, as demonstrated by a recent experiment by The Guardian.
During Q1 2018, The Guardian worked with Google and MightyHive to determine how much of the inventory on ad exchanges purporting to belong to The Guardian was indeed legit. They used ads.txt—a text file on publishers’ sites that lists the vendors that have permission to sell their inventory—to accomplish this. Many ad buying platforms give advertisers the ability to restrict their purchases to inventory that is validated by ads.txt. About eight in 10 publishers who sell inventory programmatically have adopted ads.txt, according to Adzerk.
The Guardian and its research partners found that when ads.txt filters were applied to their ad buys, there were no discrepancies between what they bought and what ended up coming back to The Guardian. But when ads.txt filters were not applied, 1% of display ad spend and 72% of video ad spend went to unauthorized programmatic platforms.
These disparate results underscore how ad fraudsters are more incentivized to target video than display. This is because video ad rates are higher than display ad rates, and video ad spend is expected to keep growing. According to eMarketer, US digital video ad spend will grow from $17.9 billion in 2018 to $29.6 billion by 2022.
Some video ads, like the 300x250 pixel units that OpenX banned from its platform this spring, are used in arbitrage plays. As more ad dollars get put into digital video, unscrupulous middlemen have found they can make a few quick bucks by buying display inventory on one exchange, repackaging it to look like video and then reselling it to a second exchange for much higher CPMs. According to Pixalate, the rate of fraud for these particular units was 31% higher than the average for all the other video units OpenX sold.
Research from DoubleVerify found that in 2017 video-based ad fraud accounted for about 10% of impressions, which was twice as bad as display ad fraud rates. The Guardian’s test shows that ad buyers who don’t protect themselves against domain spoofing are much more susceptible to video ad fraud than they are to display ad fraud.