is launching a feature in its Peer39 platform Tuesday that allows ad buyers to detect and avoid clicks from botnets in advertising auctions.
The feature aims to protect brands from spending millions on ad clicks made by non-human Web traffic, a problem that the company estimates costs the online ad industry $400 million annually.
The new fraud protection supports the 60 billion bid requests Peer39 analyzes daily. The technology identifies the non-human traffic, impressions generated by bots, and stops advertisers from placing bids on that inventory. As automation continues to rise, so does the possibility of fraud.
Alex White, GM of data and trading at DG Peer39, realizes that marketers are aware of the fraud issue surrounding real-time bidding, but not all understand the consequences and the benefits from greater vigilance. "By eliminating bot traffic from the impressions that are bid on, advertisers should see improved performance and greater efficiency across the board," he notes.
DG made Peer39 Fraud Prevention available for free to clients using Peer39's Brand Safety and Custom Brand Safety data solutions available through MediaMind and other ad platforms. The company patented natural language processing and machine learning algorithms; it also developed proprietary semantic advertising technology that matches online ads to content.
Peer39's technology cross-references traffic patterns with public and private data. They're not the only ones combatting fraud. Competitors like Integral Ad Science and DoubleVerify offer similar technology. "Everyone is blaming the exchanges, but they're only operating a marketplace, and they will never invest enough to prevent fraud," White said. "Brands need to become more aware and monitor the activity."
Some estimates put fake display-ad impressions at 30% of overall online traffic. The Interactive Advertising Bureau has a task force called the Traffic of Good Intent task force. The group's mission is to identify, understand and raise awareness of the issue of non-intentional traffic, either initiated by humans, controlled by a botnet infecting personal computers or click fraud.