2015 Was a Lost Year in the Fight Against Online Ad Fraud
The advertising industry was on pace to lose $6.3 billion last year by showing digital ads to robots, according to a December 2014 study by White Ops, a fraud detection firm. The study applied some concrete numbers to illustrate a problem that most people in the industry understood was pervasive. It was partly undertaken in the hope that the sheer scale of the issue would shock advertising agencies out of complacency.
Nope. On Tuesday, White Ops published a follow-up study showing that the percentage of traffic lost to fraud hasn’t changed at all. Given that spending on digital ads is set to grow about 15 percent this year, the firm predicted that advertisers will lose $7.2 billion to bot fraud this year.
Things aren't exactly the same as they were a year ago. Advertisers that were doing OK last year let up their guard because they assumed they had already won, said Michael Tiffany, the chief executive officer for White Ops. "They just happened to not be getting victimized when we were watching,” he said. Almost all of them saw higher levels of non-human traffic this year. The advertisers for whom bots made up over 10 percent of their traffic, meanwhile, were generally able to cut down on fraud.
White Ops worked on both studies with the Association of National Advertisers, a trade group. This year, 49 large advertisers put special tags on their ads so White Ops could check how many times they were viewed by computers, rather than humans. In a study conducted during August and September, White Ops found that, for a quarter of the advertisers, at least 9 percent of those viewing their ads weren't people at all.
That means that the average advertiser spent $10 million on ads no one ever saw, White Ops estimated. While most advertisers lost little to fraud, a few lost immense sums. The low end of the spectrum was $250,000. The worst losses totaled $42 million. Large advertisers, such as Dell, Ford Motor, and MasterCard, participated in the survey, but the report doesn’t address results for any particular company, nor does it name the publishers or ad-tech companies that were sources of fraudulent traffic.
It is well known within the world of online advertising that a large industry was set up to fleece advertisers. It uses networks of infected computers, called botnets, to mimic human activity in “viewing” advertisements. (Bloomberg Businessweek explored how this works in a cover story in September.) Ads placed through automated auctions, known as programmatic advertising, are much more susceptible to fraud. High-priced advertising, especially video, generally suffers from higher levels of bot activity.
Ad networks and publishers that end up with fraudulent traffic in their systems often claim ignorance. In an effort to close the door on questionable traffic sources, the Trustworthy Accountability Group, an industry organization, said in October that it was going to start certifying advertisers, publishers, and industry middlemen to make them responsible for bot traffic. “The programs we’ve deployed in recent months will help the industry choke off the money flowing to criminals and create an evergreen market where marketers can be sure they are working with trusted partners,” said Mike Zaneis, the organization’s CEO. While this is a good step toward reducing fraud, it's inadequate, said Tiffany of White Ops.
Bob Liodice, president and CEO of the Association of National Advertisers, said he's not surprised at how little progress has been made, given the lack of concern he hears from many of his colleagues. "A year from now, if we're in the same place again, a big, big shame on us," he said. "It's time to put up or shut up."
While the advertisers are the real victims, they may have unrealistic expectations to blame. The promise of online advertising is that marketers can find precisely the audience they are looking for. But paying to reach 100,000 women aged 18 to 29 who live in Reno, make over $250,000 a year, and want to buy a car doesn’t mean that such a population actually exists.
The more targeted the ad, the more likely it is to be viewed by a bot purporting to hail from one coveted demographic or another. For instance, marketing campaigns targeting Hispanics were almost twice as likely to be viewed by bots than the average campaign, White Ops said. “As markets change their targeting goals, bot traffic fills in the gaps between what marketers want to reach and the real online audience,” reported the study.