Why Understanding Commerce’s New Normal Is Key To Fraud-Fighting Strategies

The last eight weeks has scrambled financial institutions’ tried and tested fraud-fighting approaches as transactions once seen as “abnormal” have become the new “normal.” In the FI Fraud Decisioning Playbook, Colleen Taylor, head of merchant services for Wells Fargo, discusses how artificial intelligence (AI)-powered data analysis is helping financial institutions (FIs) adjust to these shifts in buying behaviors and avoid false positives.

The coronavirus pandemic has created financial hardships and made consumers more dependent on eCommerce as they adhere to stay-at-home orders and social distancing requirements, which are making it even more important for merchants to stop fraudsters without blocking legitimate customers from buying needed supplies.

Criminals often use stolen credit card information to make purchases, and failing to thwart these transactions can heap stress on those struggling financially and hit businesses with painful chargebacks, ultimately putting customer relationships at risk.

Financial institutions (FIs) can take steps to keep their merchant customers safe, however, by analyzing how fraud is evolving and helping them determine when a transaction is amiss so they can respond accordingly. Fraudsters have been taking advantage of consumers’ confusion during the health crisis, bombarding banks’ and credit unions’ defensive capabilities with new tactics to steal account holders’ money and personal data. Such schemes include selling fake cures and masquerading as government officials to fine consumers for leaving their homes too many times during stay-at-home orders.

Fraudsters who succeed may make off more than just funds, however. Paying illegitimate fines or purchasing phony cures involves submitting personal details like home addresses and payment information. Acquiring such data often allows bad actors to take over customers’ accounts and charge purchases to their cards, meaning FIs must be ready to detect and stop such crimes. Having precise and detailed data is key to accurately assessing whether purchasing behaviors indicate fraud or reflect legitimate customers responding to new needs and circumstances, Colleen Taylor, head of merchant services at Wells Fargo, told PYMNTS in a recent interview.

“A lot of data analytics will be important in the future as eCommerce continues to grow at an accelerated rate,” Taylor said. “Merchants need to be proactive, sign up for capabilities that will help them with insights into the transaction activities and be aware of what normal looks like, so when not normal comes along, you can easily identify and very quickly connect with the [card] processor to shut things down.”

Several types of data and analytics tools can help merchants and their FIs understand fraud risks, she added — even at a time when typical purchasing behaviors are in flux. These robust approaches enable FIs to quickly and accurately make decisions to keep customers safe without impeding their abilities to buy what they need.

Defining Normal Purchasing

Accurately assessing whether transactions are fraudulent requires FIs to gather and parse detailed information on customers’ purchasing situations. They must understand how typical behaviors look before they can draw conclusions about risk levels or distinguish outliers that warrant further investigation.

“You need insights into the data to see what looks like a normal transaction,” Taylor explained. “Then, if anomalies start to stick out, you can anticipate that maybe something is wrong there.”

Risk evaluation thus draws on characteristics such as merchants’ average order sizes, purchase frequencies or the geographic areas from which most orders are placed. Grasping these details enables FIs to detect when particular payments fall outside those norms, Taylor noted, such as when a small retailer with a history of selling locally suddenly receives what appears to be an international order.

“If you’re a local retailer in Nashville, Tennessee, you [might not] frequently have [online] transactions that originate overseas,” she explained. “We’ve seen where the card has been used before, and we see the URL location where the transaction originated, and it doesn’t look like it’s something that’s happened for you [or] your transaction volume in the past.”

Precisely identifying such red-flag situations is not easy, however, as many legitimate consumers’ purchasing behaviors are complex. They sometimes make purchases while traveling or decide to patronize new retailers when they cannot purchase what they need from physical stores during unprecedented events. Some atypical transactions thus turn out to be legitimate, and FIs must develop sophisticated approaches that can recognize this and avoid excessive false positives that create friction for loyal buyers.

Security During COVID-19

Honest customers’ purchasing behaviors are changing quickly amid the COVID-19 pandemic and resultant social distancing directives. Consumers who cannot physically visit stores are now buying online in greater numbers, and merchants and banks are working to ensure misaligned fraud-fighting systems do not accidentally turn away those in need.

“We’ve seen ourselves, as well as many of the different players in the payment ecosystem, having had to increase limits on the size of transactions because businesses and consumers are becoming more accustomed to buying things in an internet environment,” Taylor explained.

Consumers under stay-at-home orders are not only buying more online, but they are also making different types of digital transactions than before, she added. Shoppers are altering how they acquire items they previously purchased in person, with those who typically visited clothing shops or bookstores now finding those items online, for example.

Some merchants have gone digital to stay in business, too. Consumers used mobile to purchase physical goods before the pandemic, Taylor explained, but they are now turning to the channel for intangible services such as online gym and yoga classes — and are likely to keep doing so in the future. FIs must be careful to ensure they do not mislabel these new purchasing behaviors as fraudulent as they continue to check for bad actors.

Reading the Data

Powering these assessments and filtering out fraudsters requires careful data analysis. FIs typically have access to large amounts of transactional information, and artificial intelligence (AI)-powered tools can play critical roles in analyzing those data pools to generate insights. Wells Fargo creates its own such data pools by drawing on a wide array of information sources, including payment details from its consumer and commercial customer channels as well as the various payment flows — such as ACH, check and wire transactions — with which it works. This analysis enables it to create more comprehensive behavioral pictures.

“As we pool all that together, we can have better fraud solutions because we see the activity across a number of different places,” Taylor said. “By having a broad breadth of payment capabilities, we are better [at fraud detection].”

The FI can use AI-powered systems to identify consumers’ transaction patterns across payment types, sifting through that information in a fraction of the time it would have taken human analysts and finding patterns that would likely have remained hidden. Such systems could allow FIs to determine how payers behave when making card payments as well as when sending checks or wire transfers, thus providing a more holistic picture of their normal habits.

Fighting fraud is always key in merchants’ and FIs’ business strategies, and it is increasingly important as financially strained consumers seek secure and convenient commerce during the pandemic. Robust risk assessment approaches that analyze broad data pools to make precise determinations about transactions’ legitimacy can put consumers’ minds at ease, however, reducing fears that their financial details could be abused or that they could be mistaken for fraudsters and blocked from making the purchases they need.