Adjust Leverages AI to Neutralize Ad Fraud
Automation in mobile advertising has opened up the door to better ad targeting, optimized ad creatives, and incredibly fast programmatic exchanges. But there’s a downside to automation in mobile advertising, too, and it has to do with fraud.
Mobile ad fraud is surging, as bots replace the click farms that plagued the industry in years past. According to research by AppsFlyer, mobile app marketers were exposed to 30% more fraud during the first quarter of 2018 than the year prior. The problem is only getting worse, as automation makes it more difficult to differentiate between bots and actual humans.
Advertisers and brands are expected to lose an estimated $50 billion as a result of ad fraud by 2025, with one of the most problematic types of ad fraud involving bots designed to mimic human behaviors. Using bots, fraudsters can imitate clicks and engagement KPIs on ad campaigns, wreaking havoc for mobile ad vendors and the advertisers that work with them.
So what’s the solution? Firms like Unbotify are pioneering a new approach to bot detection and digital fraud prevention using artificial intelligence and machine learning. Unbotify’s solution analyzes human behavior patterns within websites’ and mobile apps’ user flows in order to differentiate between bots and humans.
The Tel Aviv-based company, which was founded by Yaron Oliker and Alon Dayan back in 2015, was recently acquired by Adjust, a firm that specializes in mobile measurement and fraud prevention, in a move designed to solidify Adjust’s position as a leader in online fraud prevention.
“By joining forces with Unbotify, we’re going to strengthen our fraud prevention solution with one goal in mind: to end fraud for good,” says Adjust Co-Founder and CTO Paul Müller.
As Müller explains it, solving fraud is an ongoing cat-and-mouse game between fraudsters and those developing anti-fraud tools trying to outpace each other. As the industry develops to prevent the use of current ad fraud techniques, the methods used by fraudsters also adapt.
To date, one of the most problematic methods in ad fraud is bot-driven fraud campaigns. Bots are able to fake human-like views, clicks, and engagements on ad campaigns—and they often stay under the radar. The issue demands a highly complex solution in order to fight back.
Müller says Unbotify’s technology takes on the hardest action to spoof—human behavior. The company’s bot detection technology uses machine learning to analyze hundreds of data points, such as device orientation, ‘touch events,’ and pressure sensitivity, in order to detect bots.
“In the past, we’ve evaluated the potential of machine learning and deemed that the data available didn’t provide new insights for us. The exciting thing about Unbotify is that they are looking at a far more complex data set,” Müller says. “The sensor data from a single smartphone creates a complex concert of interactions and cross-references between the data that can actually be understood by a machine. Here, the high volume of data is better handled by computers, as opposed to people.”
In order for there to be more trust and transparency in the digital ad market, Müller says industry leaders need to agree on a status quo and decide on the right steps to take in order to stop fraud for good. As a first step toward achieving that goal, Adjust launched the Coalition Against Ad Fraud (CAAF), which is a group of companies committed to ending fraud in the mobile ecosystem. Current CAAF members include Applift, InMobi, and Fyber, among dozens of other firms.
One of the coalition’s first orders of business is to combat the misinformation being spread about detection, prevention, and rejection, which results in a lot of uncertainty and causes the market to stay in a state of confusion.
“That’s something we’re tackling head-on with CAAF,” Müller says. “We have a number of highly knowledgeable experts in CAAF that are committed to acting upon the information we share with each other.”
Stephanie Miles is a senior editor at Street Fight.
