How to catch a fraud: Signifyd opens up tech for helping e-commerce thwart scammers

Each year, it’s estimated that e-commerce merchants lose about 1 percent of revenue to fraudsters. That may not sound too bad until you consider that the industry overall is expected to bring in $1.3 trillion in sales this year. Beyond revenue explicitly lost to fraud, startup Signifyd estimates that merchants take another 3 percent loss when it comes to transactions wrongfully declined, in an effort to protect against bad actors.

For the past couple of years, the Santa Clara, Calif.-based company has been building out its fraud detection software and testing it with a select group of clients. On Wednesday, Signifyd said it was taking its service out of beta and launching it widely for e-commerce companies.

Historically, fraud prevention and detection has involved a range of tactics, from manually reviewing transactions to blocking out entire countries (like Nigeria or Ghana) linked to fraud to flagging “consumers” initiating too many transactions at once to mapping the distance between the IP address that placed an order and the billing location.

But Rajesh Ramanand, co-founder and CEO of Signifyd and a former risk expert at PayPal (s EBAY), said his company’s approach is to combine those tactics with mounds of social and publicly available data to establish who a consumer might be and then determine the authenticity of his or her transaction.

“We help e-commerce merchants protect against these problems by connecting an online person to an offline identity,” he said. In total, he added, the company considers about 120 data points, including their social media data, public records and mapping information.

For example, the company said, if a customer makes a purchase from an IP address in New York but lists a billing address in London, current fraud systems would automatically decline the transaction or initiate a manual review (which takes time and money). But Signifyd would first examine where the card was issued, as well as the person’s social media profile. If their Twitter account shows that they live in New York but the bank issuing the card is in London, the company may accept the transaction.

By using data to establish a consumer’s context, Ramanand said Signifyd is able to save e-commerce companies of all sizes time and money.  Among its clients, to date, he said, the company has seen a 63 percent reduction in chargeback rates, a 70 percent reduction in manual reviews and a 15 -20 percent increase in revenues (particularly for companies selling internationally).

Some consumers may not love the idea of a company digging through their Twitter and Facebook (s FB) accounts to determine whether their Amazon (s AMZN) purchases are legit. But, as more people turn to the web for purchases, Signifyd’s service means that fewer authentic transactions are declined.

The startup, which has raised $2 million from Andreessen Horowitz, Data Collective and others, offers packages, starting at $19 a month, for small businesses through large e-commerce companies.