The reason these ad types are susceptible to fraud is because they’re used in a common arbitrage play. 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.
For advertisers, the result is that they pay for video impressions, but their video ads get squeezed into small units originally designed for display ads. And oftentimes, as Pixalate and OpenX found, those scrunched video ads are shown to bots.
Advertisers can use the knowledge that certain ad units are more susceptible to fraud as a clue when trying to purge bots from their media buys. For instance, if a buyer reviews ad server log files and notices a lot of 300x250 video units being purchased, that should raise a red flag. Programmatic ad buyers should also be on the lookout for fraud, since Integral Ad Science found that among buyers using fraud prevention tools, video ad fraud is more common in programmatic than direct sales.