News, Education and Events Decoding Digital Payments & Fraud

News, Education and Events Decoding Digital Payments & Fraud

Sift Targets ATO Fraud with Machine-Learning Approach

Sift Science Targets ATO Fraud with Machine-Learning Approach

Account takeover is a growing problem for online merchants and payment providers. Breaches like the one uncovered last fall at Yahoo, along with many others, have dumped hundreds of millions of data records—many which include login/password combinations—onto the open market for fraudsters to leverage. Increasingly, they are using it to access merchant and digital wallet online accounts that are linked to stored payment information. ATO is especially dangerous because after a criminal gains access to the account it is very difficult to tell them from the legitimate account holder, giving them ample time to make fraudulent orders and perpetrate other forms of fraud. To combat the rapidly growing problem, San Francisco-based Sift Science—one of the pioneers of using machine learning to predict and prevent e-commerce fraud—yesterday launched Account Takeover Prevention, adapting its approach to ATO fraud.

Jason Tan, Sift’s CEO said the new product is a response to the needs of businesses that are experiencing losses in this new way and a realization that machine learning is a natural way to attack the problem.

“It [enables us to] understand what a profile of ‘normal’ login activity looks like for a customer,” Tan told “Then, if something deviates from that profile significantly based on a set of heuristics, let’s flag that and make sure our customer can react appropriately. This is one of those types of problems where time is of the essence. The moment you detect someone’s account has been compromised, it has to be dealt with swiftly or the consequences could be huge.”

Tan noted that the rise of ATO among retailers has been an issue in the industry, but the fact that it spans so many merchant types was what convinced Sift to invest in developing a solution specifically to identify and prevent this kind of fraud.

“We hear about account takeover as a lowest common denominator amongst all kinds of businesses: online marketplaces, content Websites, community-driven Websites, dating, travel,” he said. “Hearing about the plague of account takeover firsthand from a breadth and diversity of customers gave us the confidence that this is solving a real problem.”

Learn more about this topic at these CNP Expo Education Sessions in Orlando, FL from May 22-25, 2017.

  • Identifying and Preventing ATOs and New Account Creation Fraud
  • Selecting the Right Vendor
  • Minimizing Non-Payment Account Abuse

Read the next article in today’s CNP Report.