The Trustworthy Accountability Group (TAG), which is uniting industry players to root out criminal activity in digital advertising, has unveiled a pilot program to block fraudulent traffic coming from data centers.
Google proposed the data center project and is leading the pilot program, which will see member companies sharing intelligence about fraudulent data centers. The companies can then block the bad actors collectively, turning off their access to advertising revenue.
“Google built an army of anti-fraud engineers around the Spider.io acquisition,” TAG CEO Mike Zaneis said. “They’ve spent a tremendous amount of time identifying this data center traffic and it’s as robust of a database as anyone in the world. It’s a great starting point for TAG, and we’ll build off that from the intelligence.”
“We’re looking not just to protect our clients, but make the entire ecosystem cleaner and safer,” Vegard Johnsen, Google’s display ad traffic quality product manager, said of its involvement in TAG.
Dstillery, Facebook, MediaMath, Quantcast, Rubicon Project, The Trade Desk, TubeMogul and Yahoo are also piloting the program.
The data center block list complements TAG’s Fraud Threat List, which will graduate from its pilot program in October. That list focuses on finding fraudulent domains and bots on consumer computers.
“There’s no silver bullet to identify sources of fraudulent traffic, and you have to look at the problem from a 360-degree view,” said Zaneis.
If the data center blocking works, there will be some negative consequences for advertisers and advertising intelligence companies. A Google chart shows a steep decline in click-through rates after data center traffic is filtered.
“Even though it will present transitional challenges as you filter out high-performing bots, you’re investing in actual people and engaging actual consumers,” Zaenis said. “Ultimately, the ROI on the ad spend will go up.”
Another casualty of the data center block lists will be advertising intelligence companies that run bots to determine advertisers’ spending patterns. One such company was responsible for 65% of automated data-center clicks in May, Johnsen wrote in a blog post. He called its removal “collateral damage.”
Other advertising intelligence companies may be affected too, Johnsen told AdExchanger. Traffic from data centers determined to be bad actors are “blocked pre-bid when possible but also post-filtered depending on the situation,” because post-filtering “allows us to protect advertisers without alerting bad actors of our activity.”
During the pilot program, TAG will develop a set of operating principles to present to the anti-fraud working group and open them up for public comment. TAG plans to have the program operational by the end of the year.