With traffic and ad fraud the bane of the digital ad industry, another ad platform is taking steps to fight back.
Today, San Francisco-based Smaato announces the addition of an automated fraud detection engine to its publisher-oriented mobile ad platform, exchange, and server.
This new automation, CEO Ragnar Kruse told me, uses machine learning and artificial intelligence to find patterns in authentic and fake traffic, so it can tell the difference. He claimed that 80 to 85 percent of fraudulent traffic can now be detected automatically on the company’s newest platform, plus another 10 to 15 percent is found manually by a team of specialists.
Only about one to three percent of fraudulent traffic now gets through, he said. The key enforcement mechanism: pruning out suspicious publishers. In the vast majority of cases, publishers are the ones paid for the ads on their space that are loaded by software bots, so weeding out the bad ones assumedly removes much of the fake traffic.
Kruse said about a quarter of publishers are now being blocked because of suspicious activity, roughly 80 percent of that automatically and the rest manually.
“Where you have a platform with good publisher inventory,” he said, “it will [lead to] good ads.” The system also blocks suspicious traffic, including up to a third of traffic intended for apps.
The company’s newest platform, which launched in January, has about 8,500 publishers. The 80,000-plus publishers on the older platform, where fraud detection is manual, are being migrated to the newer one. Publisher clients include Microsoft, Blackberry Messenger, Grinder, Ask.FM, and MeetMe. The 2005-founded company says it handles over 7.5 billion mobile ads on its platforms every day.
Smaato’s efforts to combat fraud joins that of other platforms and participants in the digital ecosystem who are trying to turn this scandalous situation around.
Some estimates indicate that half or more of all digital ad traffic comes from software bots loading the ads, not actual people. This means, of course, that advertisers who pay by impressions are paying for their ads to be seen by machines. One assessment by Solve Media estimated that advertisers lost $11.6 billion last year to click fraud.
Absent some industry-wide technology or other solution, it’s likely that each ad platform will offer its own solution, like a Wild West where only the towns with the toughest sheriffs are safe to live.
By booting out suspicious publishers, Smaato and other adtech providers are in the process of demarcating “good” and “bad” neighborhoods. In the former, respectable site and app publishers/developers will live; everyone else will live in the less-picky environments.
Sifting publishers, ads, and traffic through intelligent automation is becoming the new best practice, Kruse told me. He noted that Google, OpenX, and The Rubicon Project, among others, are undertaking similar efforts to reclaim their industry.
Speaking of Google, the tech giant today said it is supporting a new pilot blacklist at the Trustworthy Accountability Group (TAG), an industry initiative to combat ad fraud, malware, and piracy. Google will provide the group with its blacklist identifying the software bots running in data centers that manage to avoid detection by masquerading as human visitors.