Last month AdExchanger reported on plans by AppNexus to roll out a “Certified Supply” stamp of approval for demonstrably valid impressions, and thereby cut off the flow of spend to impressions that are fraudulent, nonviewable or otherwise undesirable.
The certified program came amid growing industry concerns about the persistence of fraud in the AppNexus supply.
On Wednesday at AppNexus’ New York summit, CEO Brian O’Kelley detailed how the program will work and named partners and fraud vendors that will collectively create a definition of fraud that can be applied across all its supply – and then enforce that definition.
The Certified program will use AppNexus’ own fraud-detection technology in concert with two third-party ad verification vendors: DoubleVerify and Integral Ad Science. Beginning in the second quarter of 2015, when AppNexus or its buy-side partners use either vendor to ferret out fraudulent or invalid impressions, the buyer will not pay for that supply and the seller will not receive payment for it. (AppNexus left the door open to accepting other anti-fraud firms in the future.)
What’s more, AppNexus has secured buy-in from a handful of key partners on the sell side, including its two most prominent “frenemies,” PubMatic and Rubicon Project. Whenever a bid passes through AppNexus to one of its publisher customers, the two SSP platforms have agreed not to receive payment on that customer’s behalf.
How much money will AppNexus “rip out?” “Millions or tens of millions,” O’Kelley estimated. But it’s hard to say until the second quarter, when dubious ad networks and other firms representing questionable supply will be forced to contend with the new anti-fraud measures.
Additionally, the damage caused to suppliers benefitting from bad traffic such as bots and ad-injecting toolbars is likely to increase over time, as more marketers and buy-side platforms adopt anti-fraud technologies.
In addition to AppNexus’s competitors, key supply partner Microsoft has signed on to the Certified program. So have Yahoo and Xaxis, whose parent WPP Group is now a major investor in AppNexus. Xaxis says it has blocked payments for fraudulent impressions for some time.
Could AppNexus’ move create an industry standard for fraud measurement and policing? Today there isn’t one, O’Kelley argued; vendors have varying methods, and exchanges each enforce in their own way. In this environment, he said the Interactive Advertising Bureau stands little chance of forcing a solution.
But he argued AppNexus has the scale to force alignment on a fraud-fighting standard – essentially a market solution.
“If we set the standard, every major advertiser says [to other supply partners], ‘Why don’t you?’ If we do this, everybody else has to follow suit,” he told AdExchanger.
Refunding Fraud Dollars
In his interview with AdExchanger last month, O’Kelley said, “We do our best when we find [fraud] to rip out those impressions and credit everybody and do the right thing.”
Some readers of that interview pushed back against that claim, saying they had identified fraudulent impressions, but rather than give direct credit, AppNexus referred them to the publisher that floated the contested ad space on AppNexus’ platform. In these cases, the onus was on the buyers to lay claim to wasted dollars.
O’Kelley did not deny that this had happened, but he said the company is progressing toward a fix.
“Today it’s a mess,” he said. “Some of the buyers bill off ad server numbers, not our numbers, and that’s where the rebate comes in. There’s no agreed-upon currency. We’ve been working on this for nine months. The reason it has taken so long is the logistics are kind of crazy.”
As the company gets up and running with its Integral Ad Science and DoubleVerify, most payments for fraudulent ads should never happen to begin with, obviating the need for a refund, O’Kelley said. But he promised that when and if the fail-safe fails, AppNexus will issue credits.
“You will get money back,” he said.
Sarah Sluis contributed.
