Signifyd has launched a e-commerce fraud protection service that tracks 120 indicators, including social media profiles, device IDs, location and purchase history to determine whether an online purchase is valid.

The startup applies machine learning to its dataset to score transactions based on the likelihood of fraud. The idea is that if your social media profiles indicate that you’re on vacation in another country, Signifyd could pick up on that fact and approve a purchase that would have otherwise been declined.

analysis Signifyd wants to use your social media profiles to detect e commerce fraud

Signifyd claims that its platform will not only reduce fraudulent purchases, but also provide retailers with additional revenue by finding legitimate transactions in high-risk areas. Customers will also benefit by receiving fewer calls from their credit card providers.

CEO Rajesh Ramanand said in an interview that merchants on its platform saw a 60 percent reduction in chargebacks on average and a 15-20 percent increase in revenue from approving transactions.

The company, which is backed by Andreessen Horowitz, counts a number of Fortune 1000 and Internet Retailer Top 500 members among its customers.