Surprising Good News: Integral Ad Science Says Digital Ad Fraud Is Falling, and Ad Viewability Rising
Here's something not seen often in reports about digital ad fraud and viewability: good news.
That's in contrast with predictions for a worsening problem with ad fraud, in which bad actors charge marketers for ads intentionally served out of sight or to software routines only simulating real people. The Association of National Advertisers, for example, said in January that fake traffic would cost marketers $7.2 billion globally this year, after earlier predicting $6.3 billion in ad fraud for 2015.
Integral Ad Science also found that viewability of programmatic display ads rose by 20.7% last quarter vs. the prior quarter, meanwhile, with slightly more than half meeting the Media Rating Council definition of viewability. And video viewability rose even more sharply, by more than a third, albeit to only 40.4%.
While the improvements in ad fraud last quarter were fairly sharp, driven somewhat by seasonality (the fourth quarter tends to have higher levels of fraud because demand for inventory is higher) the improvements weren't a one-off, according to Integral Ad Science. It was the fifth quarter in a row that fraud decreased both for inventory bought programmatically or direct from publishers, though the latter still has much lower rates of ad fraud, the company said. Integral found programmatic buys included 8.7% fraudulent impressions vs. 2.4% for publisher-direct buys.
"Brand risk," or ads placed in content deemed risky for brand messages, such as porn, hate speech or extreme violence, also declined sharply last quarter to 9.1% of U.S. placements from 12.1% in the fourth quarter.
Generally, the U.S. did better than the rest of the world in the data, in which Integral found 13% of ads placed in risky environments and 40.4% of ads viewable globally, according to MRC standards last quarter.
But the U.S. still had the worst rate of ad fraud among five countries called out in the report, including Australia, France, Germany and the U.K. France had the highest rate of viewability at nearly 61% and lowest rates of brand risk at under 7% among the five.
Viewability in the study covers mobile and digital ads according to the MRC desktop display standards, which call for at least half of an ad to be viewable for one second for display and two seconds for video. The MRC last month proposed applying to same standards to mobile ads.
The Integral report does also look at viewability over longer periods, finding it declined from 52% overall last quarter at the 1-second mark to 36.7% at 5 seconds and 29% at 15 seconds, with publisher direct inventory performing better than programmatic.
Integral, which gets paid by publishers, ad networks, exchanges, demand- and supply-side platforms, marketers and agencies for analytics to improve digital ad performance, based the findings on more than 4 billion ad impressions, said Chief Data Officer Kiril Tsemekhman.
"I think over the past 18 to 24 months there's been an increased awareness that the better the quality of the media bought, the better the advertising typically does," said David Hahn senior VP-product management of Integral.
"There is an element of economics here," Mr. Tsemekhman said. "More entities now transact on viewability, or don't pay for fraudulent impressions. So both sides of the table are incentivized."
In short, publishers and other players in the industry increasingly "aren't getting paid for inventory that's fraudulent," he said. "So they're eliminating inventory known to be highly fraudulent. This is a trend, hopefully a persistent trend."