The ANA estimated that U.S. advertisers will lose $7.2 billion to fraud in 2016
Advertisers are more aware than ever of the dangers of online ad fraud, but the most common prevention measure advertisers are using isn’t working, according to the Association of National Advertisers‘ annual review of bot traffic among top advertisers
The ANA’s Bot Baseline study, conducted by security firm White Ops and widely regarded as the definitive benchmark for ad fraud in the U.S., found that despite growing attention to the problem, no significant inroads have been made into reducing the amount of fake impressions lurking in publishers’ inventory. The ANA found that in fall 2015, fraud levels had declined just 0.2 percentage points from its 2014 assessment of 11%.
Fraud levels have remained intractable in part because fraudsters have become more sophisticated in the techniques they use to evade detection by security vendors.
But it also has to do with advertisers adopting prevention measures that don’t work, the study said — specifically, services that identify and exclude fraudulent domains and IP addresses, or so-called blacklists.
Three quarters of the 49 advertisers involved in the study made use of blacklist-based detection technology, but the study found that their campaigns weren’t any safer than those that were totally unprotected. Security firms like White Ops and Integral Ad Science frequently criticize blacklists as a reactive measure, which isn’t able to keep up with fraudsters who are constantly creating new domains to front their operations. More sophisticated alternatives can identify fraud from suspicious traffic patterns and avoid buying it in the first place.
The ANA estimated that U.S. advertisers will lose $7.2 billion to fraud in 2016, up 13% from what it projected advertisers would lose at the beginning of last year. (The study attributed the widening loss to increasing advertiser spend, since fraud levels have remained stable.)
No participant in the study was completely immune to fraud, but some were hit much harder than others. Fraud rates for each advertiser ranged from 3% to 37%, and 30% of the advertisers ended up paying for 80% of of losses. The highest impacted advertiser was estimated to have lost $42 million on an annualized basis.
“Much of the media purchased by the typical advertiser is clean, but when fraud does affect an advertiser, it tends to hit hard and in very concentrated areas,” the study authors wrote.
Advertisers participating in the study for the second year in a row didn’t fare much better than newcomers, either. Of the 28 returning advertisers, only nine saw a reduction in fraud.
Higher value media (such as video) tended to be more heavily afflicted with fraud. Programmatic campaigns that made heavy use of data targeting to reach a niche audience were also more vulnerable. For example, campaigns targeting Hispanics had a 70% higher fraud rate than campaigns that weren’t targeted at that demo.
The study also confirmed that direct buying tends to be safer than programmatic buying. This was especially true in video, where fraud levels were 73% higher in programmatic campaigns than in direct bought campaigns. In display, by contrast, there was 14% more fraud in programmatic than in direct bought inventory.
