Ad fraud falls 11%, but continued vigilance is required: Bot Baseline report

Ad fraud is expected to cost marketers US$5.8 billion globally this year—down from US$6.5 billion in the previous study in 2017—according to the fourth Bot Baseline report from White Ops and the Association of National Advertisers (ANA).

The 11% decline is particularly notable given the 25.4% increase in digital spending during the same time period. White Ops co-founder and CEO Tamer Hassan attributed the success to “unprecedented” industry collaboration that is allowing fraud to be tackled at a global scale.

Screen Shot 2019-05-23 at 2.43.57 PM.png

Bot Baseline said that fraud attempts amounted to 20-35% of all ad impressions in the past year. However, for the first time ever, more fraud will be prevented than will succeed this year—invalidated by demand-side platforms (DSPs) or supply-side platforms (SSPs), filtered out as sophisticated invalid traffic before being paid for, or invalidated via claw backs. Minus these measures, fraud would have grown to US$14 billion said the report.

The report attributed the decline in ad fraud to several key factors:

  • Advertisers are directing more spending through buying channels with dedicated independent fraud prevention measures, an option that wasn’t available at the time of the first study.
  • It has become harder to purchase bot traffic, which used to be sold on the open web. Today, many bot traffic sellers have been driven out of business or underground.
  • Buying bot traffic has become more expensive. An increase in the market price for sophisticated bot traffic has limited the arbitrage opportunity for buying visitors and showing them enough ad units to turn a profit.
  • Initiatives such as ads.txt have helped reduce desktop spoofing, with 78% of the top volume domains in the study using ads.txt files to prevent their inventory from being successfully spoofed.
  • The cost of getting caught is much higher. A number of arrests in the past year have signalled to would-be criminals that committing ad fraud is a significantly riskier proposition.

The Bot Baseline report measured 27 billion ad impressions (13.5 billion across both of desktop and mobile), excluding social and search, from 50 brands. Their campaigns spanned 2,400 campaigns and 606,000 domains.

Video was the most commonly targeted format for fraudsters, with 14% of ads determined to be fake, down from 22% in the previous report. While 8% of desktop display ads were fraudulent, 12% of other desktop formats, such as rich media and takeovers, were fake. For mobile, 3% of display ads and 8% of in-app video ads were determined to be fake.

While ANA CEO Bob Liodice said that the findings suggests that the war on fraud is “winnable,” the report also identified fraud as an “evolving threat” that requires continued vigilance.

Screen Shot 2019-05-23 at 3.27.41 PM.png

The report pointed to the IAB’s two-year-old ads.txt initiative as having a “particularly positive” impact on desktop fraud.

Ads.text was introduced as a way of curbing spoofing, which the report described as the “RTB equivalent of selling knockoff designer handbags.” Ads.txt is a text file hosted by a publisher’s webserver that lists all companies authorized to sell its ad inventory, while programmatic platforms include a file that confirms the publisher inventory they are permitted to sell.

The report says that 78% of the overall study volume was covered by ads.txt, although just 14% of all domains had published a valid ads.txt file. “This shows that there is a long way to go to bring the long tail of domains into the ads.txt program,” the report stated. “But even at the current state of adoption, a majority of ad impressions on the web are now being covered.”

The report said that the most immediate—albeit “fixable”—issue faced by the industry is the “incredibly uneven” audit-ability of ad campaigns across formats and environments. It said that a “shockingly large” percentage of video advertising continues to be sold in the outdated video ad serving template (VAST) 2 format, which offers a very low level of third party validatability, while other publishers are “bending over backwards” to provide full, dynamic audited JavaScript execution from accredited third parties.

“In effect, different publishers are being held to different standards of validatability, with very little marketer awareness,” the report concludes.

It also noted that fraudsters are moving to other formats in response to reduced profit in desktop fraud. In addition to increased mobile fraud, the report also noted “rising threats” in both OTT and server-side ad insertion, and “murky waters” surrounding incentivized traffic. “We believe that all of these issues will become more prevalent in the coming year,” it said.

Chris Powell