Integral Ad Science, a technology company with a focus on safe advertising, on Tuesday announced the release of a real-time fraud prevention service for both advertisers and publishers. Scott Knoll, the company's CEO and president, told RTM Daily he believes the technology is the first to proactively prevent fraud rather than simply detect it.
Knoll also shared that the new solution is going to be integrated into ad exchanges. He could not name the exchanges, but said more than one partnership is in the works with "major" exchanges.
For the past two years, Integral has been working on solutions for ad fraud. Knoll said there are numerous solutions out there for fraud detection — and that detection happens in real-time — but there was not anything for fraud prevention.
"The challenge with [fraud detection methods] has been two-fold. One is that it's reactive, and you are assuming that the fraud is going to happen in the same place again," Knoll said. "So you're always chasing fraud with the assumption they will come back and…rob the same house again. The second challenge is that it eliminates a lot of good inventory.
"Here's the change now," he continued. "What we're doing now is
detecting [fraud] in real-time like we were before, [but] we're now telling the advertisers, in real-time, 'Don't buy this inventory.'"
Knoll said the real-time detection and prevention is in stark contrast to what the industry is currently doing — creating black lists, which are essentially "do not buy from" lists. He believes one of the main benefits of the new technology will be less black listing.
"When you black list, there's a lot of legitimate sites that have fraud on them," he said. "This is bad for buyers because it could be a scale issue. It's bad for publishers…because if you are aggregating, it totally eliminates one source of inventory. And even if 95% of the impressions on a site aren't fraudulent, [a black list] still eliminates all of it."
For buyers, the technology will stop an ad from being delivered if the inventory is fraudulent. For sellers, the inventory is pre-screened before it hits the market.
RTM Daily asked Knoll what the percent likelihood of an impression being fraudulent had to be in order for Integral's technology to stop the buy/sell from happening. He answered that because the technology can detect fraud at the point of sale, there is no guesswork.
"At that time there is literally not a percent," Knoll shared. "Is it a bot, or is it not? We can determine that in real-time."
RTM Daily pressed Knoll to share more on how the technology can know — beyond a doubt — the legitimacy of an impression. He didn't share the exact method behind the technology for fear of the "bad guys" learning too much, but did say that, essentially, Web browsers send signals which Integral's technology can use to determine if the user is human or not — in real-time. Knoll is confident Integral's technology won't block a good sale, but did note it won't catch every instance of fraud.