Wednesday , September 19, 2018

Eye on E-Commerce Fraud: CyberSource And Worldpay Harness Machine Learning And Big Data

Machine learning and huge swaths of data are being enlisted in the battle to counter online fraud, with CyberSource announcing an enhancement to its fraud-prevention service and processor Worldpay Inc. launching a new fraud-prevention service Wednesday at MRC 2018, the annual conference held by the Merchant Risk Council.

CyberSource, a unit of Visa Inc. since 2010, announced machine learning is coming to its rule-creation service. Dubbed the Rules Suggestion Engine, the service uses a merchant’s unique transaction data to generate recommended rules for e-commerce transactions. The goal is to authorize as many legitimate transactions as possible and do so in an automated fashion.

Mattei: “What we’re hearing from our merchants is that there is value in the card-present data.”

The service works by analyzing the broader base of transaction data that runs through CyberSource and present new transaction rules that might catch fraudulent transactions.

“The big advantage is, it’s made for merchants that have an educated risk analyst,” Scott Boding, CyberSource vice president of product management, tells Digital Transactions News. “It’s not just about preventing chargebacks. It’s about looking at the overall process and making sure good customers get through.”

The Rules Suggestion Engine is available now in conjunction with Decision Manager Replay, a service CyberSource introduced in 2015 that enables e-commerce merchants to test various anti-fraud measures for online transactions.

Worldpay, the name of the newly combined Vantiv Inc. and Worldpay Group plc, said it is working with Featurespace Ltd., a United Kingdom-based behavioral-analytics technology company, on a new fraud-prevention service that uses machine learning to thwart criminals.

“Data is king,” David Mattei, Worldpay vice president and fraud-product manager, tells Digital Transactions News. Available in the second half of the year, the service will tap into the combined 40 billion transactions processed by Worldpay around the globe to better distinguish legitimate consumers from those with ill intent.

Unlike some other services, the Worldpay service will include card-present as well as card-not-present data, Mattei says. “What we’re hearing from our merchants is that there is value in the card-present data,” Mattei says. ‘There’s a lot of detail in the card-present transaction.”

It can help a merchant risk manager if he knows which card a consumer used in a store and if the same card appears on the merchant’s e-commerce site. Or, it may aid the validity check of a consumer who used one card in a store, and chose to use a different card online, Mattei says. The service can help merchants identify that the transaction is coming from the same individual and avoid viewing the consumer as one not seen before.

Worldpay is able to do this from information culled at the card-authorization step, Mattei says. It takes that information and adds data from the Worldpay transaction file that can be associated with the card. Then, the data is passed to the Worldpay fraud system, which is where Featurespace’s analytical prowess, and that of others, is applied. “Rather than just the authorization, we’re adding a lot more value with the fraud system and then using it for the decisioning process,” Mattei says.

The new service will be available to merchants of all sizes internationally and through multiple sales channels, he says. Pricing is still in development. A self-managed option is available, as is a managed service, which Mattei predicts will be predominant, that will see Worldpay professionals acting on behalf of merchants.

Coincidentally, Worldpay and Vantiv, prior to the merger and independent of each other, had selected Featurespace for a role in what would have been their own services.

Check Also

With Stripe Terminal, the Online Payments Specialist Moves Into Face-to-Face Sales

Stripe, the San Francisco-based online payments company, is taking a step to get inside stores. …