Sep 26, 2018

Victor Levi Forter, a company that uses machine learning to detect and prevent fraud in online retail transactions, announced today that it has raised $50 million in a series D funding round led by March Capital Partners, bringing its total financing to $100 million. Salesforce Ventures joined the round, along with previous investors including Sequoia Capital and New Enterprise Associates (NEA). Global online retail sales amounted to more than $2 trillion in 2017, and this number is projected to reach more than $4.5 trillion by 2021. But as online retail grows, so does online fraud. Account takeover attacks are on the rise, as are high-volume automated “bot” attacks and so-called policy attacks like coupon abuse. Innovations in mobile payments and expanded services like in-store pickup for online purchases have given fraudsters even more new leverage in the e-commerce space. Jamie Montgomery, a managing director at March Capital Partners, says that fraud prevention is a “much bigger market than most people realize,” and that the service Forter provides is indispensable. "You’ll ask a large corporate customer, 'what would you do without them?' And the answer is, 'we’d be screwed'," he quipped. Recommended For You Michael Reitblat, Forter's CEO, says that the company can reduce fraud levels for a merchant by 50%. Forter has tripled its customer base in the last twelve months and Reitblat says it's on track to grow more than 150% between 2017 and 2018, though on the whole, Forter is not yet profitable. "What we’re looking for is a growth in the business, and the company is growing," Montgomery says. Of course, Forter isn't the only player in the fraud prevention space. Riskified, founded the same year as Forter in 2013, raised a $33 million series C in 2017, and Signifyd, founded in 2011, raised $100 million in a series D round in May. All three companies offer a form of "chargeback guarantee," meaning that they reimburse merchants for chargebacks related to fraudulent transactions that were wrongly approved. Reitblat founded Forter along with Liron Damri, the company’s president, and Alon Shemesh, its chief analyst. Reitblat met Damri at a boarding school in Jerusalem before serving as an officer in Israel’s military intelligence unit. All three co-founders previously worked for Fraud Sciences, an Israeli company specializing in online fraud prevention that was acquired by PayPal in 2008 for $169 million. “We were on to something,” says Reitblat of his time at Fraud Sciences, but the technology at the time was still in its nascent stages. “Fraud Sciences was very, very manual. It was very accurate, but it required a lot of people to review transactions and make decisions.” Reitblat, Damri, and Shemesh helped develop fraud prevention technology at PayPal before founding Forter five years after the acquisition. Today, Forter processes more than $50 billion in transactions annually for its network of retailers, which includes Fortune 500 companies like Nordstrom and startups like Away Travel, using a machine learning algorithm rather than manual reviews. Reitblat says the algorithm is both substantially more accurate than older fraud detection methods and claims it is infinitely scalable. The algorithm collects thousands of data points about online transactions, from the type of device a would-be customer is using to their geographic location to their browsing behavior on the website. Data like this is collected well before a customer enters their credit card details and hits “submit” on an order. “We start looking at how you behave,” Reitblat says. “What are you doing? Are you behaving like someone who wants to order something? Or are you testing different cards?” Suspicious transactions get rejected based on thousands of indicators, and the system accounts for changes in a shopper's regular habits to prevent legitimate transactions from being flagged as fraudulent. And as more customers sign on, Forter says its performance will only improve. “The more customers we have, the more merchants we work with, the more data we see, the more patterns we recognize sooner, the better the performance is,” says Reitblat. In other words, more customers don’t strain the network; rather, they enhance it. Forter’s database already covers over 180 million U.S. consumers as well as 50 million additional shoppers in 188 countries on more than 2.1 billion devices. Better pattern recognition means fewer errors, especially when it comes to false positives, which Reitblat estimates cost the U.S. economy $20 billion per year in lost sales. Eido Gal, CEO of Riskified, says a major challenge in the space is keeping up with the changing fraud landscape without enacting imprecise fraud detection criteria that could potentially turn away legitimate customers. “That's a big challenge for the industry as a whole,” Gal says. “How do we maximize approval rates and maintain the customer experience as fraudsters get more advanced?” Ultimately, better fraud detection means more profits for all retailers. While many online transactions from developing economies in Africa, southeast Asia, and even China, for example, would have been blocked by older iterations of these solutions, today’s more advanced algorithms are more likely to approve legitimate purchases. Reitblat says he hopes that one day, retailers won't have to differentiate themselves based on their ability to prevent fraud: it should be a given. "Fraud should not be a problem they're even thinking about," he says. When merchants can focus on the other aspects of their business, “it benefits everyone.” Follow me on  Twitter . Send me a secure  tip .