Dealing with fraudulent orders costs time and money, and given the fact that the holiday season brings with it a 20% increase in orders and a 10% increase in fraudulent attempts, home improvement products retailer Build.com Inc. wanted to make sure its fraud prevention program was in tip-top shape.

Build.com, No. 68 in the Internet Retailer 2016 Top 500 Guide, had been using fraud prevention tool Signifyd since 2012 as a secondary source to supplement its fraud prevention program. Build.com risk specialist Jamie Ceccato says the company’s old risk-evaluation tools weren’t doing a good enough job of scrutinizing orders and weeding out fraudulent ones.

“The rule evaluations (used to determine whether an order was fraudulent or valid) were set when the company was built and hadn’t evolved for any sort of change or trend,” she says. Rules-based fraud tools are built around customized rules, such as flagging orders from particular countries or over a certain dollar amount, and they can result in many orders requiring manual reviews or false positives, which are legitimate transactions that are declined.

Ceccato leads a team that includes two risk specialists. The three of them handle all the risk review for orders that don’t pass the rule checks as well as the chargebacks.

About two years ago, Build.com decided to move Signifyd from its secondary fraud prevention tool to its primary one. Signifyd is an application programming interface (API) that plugs into a retailer’s existing e-commerce platform. It isn’t a rules-based tool, rather Signifyd uses real-time intelligence to determine the validity of a transaction, analyzing data from its network of more than 5,000 merchants to gauge whether or not each individual transaction is legitimate. Ceccato worked with an in-house developer to implement the new fraud prevention solution.

“We take [shopper] information that’s presented to us in the course of a transaction, get some contextual information [about the consumer] and pull it back into the transaction, which allows us to make very intelligent decisions,” adds Skye Spear, vice president of partnerships and business development at Signifyd. Spear declines to specify the contextual information it combines with Build.com’s consumer data, but often such information includes data to identify individual PCs, phones or tablets, as well as geolocation. Merchants who use the tool are charged a percentage of each approved transaction.

The new risk management system went live in mid-September, and Ceccato says the difference is clear. The new system has saved her team an estimated 10 man hours per week by automatically rejecting orders that would otherwise go out and then come back in the form of a chargeback, she says.

Orders are being rejected because they fail to meet the retailer’s standards, she says. “It has helped us in terms of time management because there was a lot of manual work before. It’s saved us time, it’s saved us money and it’s saved us on the turnaround time for order processing.”