For the third consecutive year, the Association of National Advertisers (ANA) has partnered with analytics firm White Ops for its Bot Baseline study, covering digital advertising fraud in the North American market.
The study, which monitored the digital advertising activity of advertisers between October 2016 and January 2017, found that bots and fraud will contribute to a projected $6.5 billion in global losses — a drop from $7.2 billion last year. The study also reported that with digital ad expenditures expected to increase by 10% this year, the drop in ad fraud/bot losses is an even bigger feat.
White Ops monitored bot activity broadly, while also looking specifically at 49 participating advertisers from the ANA. Of those 49 members, 27 had participated in a previous study, and 18 had participated in the full three years.
Chris Williams, VP of digital of the Association of Canadian Advertisers (ACA), said with the number of returning advertisers, the ANA’s member participants no longer reflect the overall market, because “they’ve learned from the analysis what the best practices and strategies are to fight fraud.”
The study determined that among ANA member participants, the overall loss projected due to fraud would be $3.3 billion globally for 2017, about half of the projected total loss. The ANA members found to be in the top 20% of performance were projected to lose $700 million to fraud and bots for the coming year. For mobile, fraud losses accounted for 2% of spending.
Mobile’s strong performance was driven largely by lower CPMs and a lower number of ad units, which decreases the profit margin for publishers buying traffic. In-app fraud is also limited by the install base for fraudulent apps. Counterfeit inventory on mobile also does not often achieve a very high price for programmatic bid requests.
Fraud is still most common in desktop video, with losses counting for 22% of spending, as opposed to desktop display advertising, where it accounted for 9%. Participants also found fraud rates comparable between programmatic and direct buys, which led the study to determine that programmatic was “no longer universally risky.”
But bots are getting smarter, and the study reminded members that fraud will likely continue to exist. Of fraudulent impressions studied, 76% were found to be from machines mixed with humans and bots, meaning the added “human” intelligent bots allows them to get past detection methods.
But Williams said as the study continues, it’s helping to start valuable conversations, especially on the publisher side.
“Really, it comes down to the publishers — the people who write the content and manage the site,” he said. “It’s garbage in, garbage out, and [if] you let garbage into your system, it’s going to work its way around.”
Williams said he has observed an increasing number of publishers are pursuing certification with the Alliance for Audited Media, which verifies the quality of digital publishers for advertisers.
Image: Shutterstock
