Since most fraud prevention platforms analyze consumer behavior to detect fraud, a substantial change in consumer behavior can throw off the models. Therefore, with COVID-19 forcing people to change the way they work, shop, and interact with one another, fraud models need to adapt in order to stay effective.
To understand how COVID-19 has changed consumer behavior and what that means for fraud models, PaymentJournal’s editor-in-chief Ryan McEndarfer sat down with Robert Capps, VP of Market Innovation at NuData, a Mastercard company.
People are staying at home more, creating a stable signal for fraud models
As the pandemic worsened and infections spread across the world, many governments imposed stay at home orders that greatly curtailed people’s movement. By April in America, at least 316 million people were urged to stay at home, resulting in a significant decrease in people’s movement.
As this happened, many physical stores were shuttered and people’s commercial activity largely migrated into digital channels. Although some states have since relaxed restrictions, consumers are still wary of frequenting physical stores and large numbers remain working from home or temporarily out of work.
While all this disruption may seem to make detecting fraud more difficult, there are actually a number of promising implications for some fraud prevention platforms.
“How consumers access [their accounts and online services], where they’re accessing from, and the devices they’re using—that is all very stable,” explained Capps. In other words, people are reliably logging into their accounts on the same device, from the same location, and over the same network for months now. “That creates some remarkably trustworthy behavior,” he continued.
In normal circumstances, the average person is conducting their commercial activity through a myriad of different channels and across a variety of locations. For example, a person may use their credit card to pay for a coffee at a café in the morning, then later browse for new clothes on online while using their work computer, and finally come home in the evening and order a pizza through their personal laptop.
Now, the majority of those types of behavior are occurring in one location and through one device, making it easier for fraud detection platforms to establish normal behavior.
Transaction types and frequency are changing
What has changed due to COVID-19 is what people are buying, when they’re buying it, and who they’re buying it from. Capps explained that with many materials in limited supply, consumers are increasingly turning to different merchants to find the goods and services they need.
In addition, people are logging into their online accounts and conducting activity at different times of day than they normally would. “Now during the workday, we’re seeing more transactions, more interactions, more logins,” said Capps.
This means that fraud detection models that place an emphasis on transaction types, times, and frequencies will become less effective. “If you look at purely transactional data, your models are probably going to have a heart attack, because those transactions are changing rapidly,” said Capps. It is no longer abnormal for someone to log into their bank and move money during the workday, whereas that behavior would be suspicious pre-pandemic.
The pandemic has created an opening for fraudsters
Calibrating fraud prevention models in light of these developments is crucial because fraudsters are looking to capitalize on all the confusion. Since the pandemic began, malware attacks, phishing attempts, and all sorts of scams have proliferated, a worrisome trend that Capps discussed in a previous PaymentsJournal podcast.
Making matters worse is the fact that digital transaction volumes are skyrocketing. To keep up with the surge in traffic, some merchants are tempted to relax their fraud platforms because they may otherwise get overwhelmed by the increased activity. Aware that this is the case, fraudsters will then probe for weaknesses and exploit vulnerable merchants.
When the holiday season begins, these problems may be exacerbated further.
Merchants need a layered approach to fraud prevention
Given the serious fraud threats facing merchants, they need to utilize platforms that can keep up. Capps identified three areas where merchants should focus on:
- New account openings: It’s common for fraudsters to make an account and then let it age for a few months before using it. Therefore, merchants should be sure to screen new account creations and look for automated activity. “Getting a handle now on new account creation is going to help organizations with the onslaught of fraud that’s going to be coming in the next few months,” said Capps.
- How transactions occur: Similar to account openings, transactions that occur in an automated fashion should be identified. “Having strong consumer or human identification in a transaction helps to mitigate a lot of these attacks,” noted Capps.
- Fraudsters aren’t price conscious: Many merchants falsely believe that sales attract fraudsters. Capps explained this simply is not the case; fraudsters are going to attempt to steal what they want, no matter the price or time. If anything, sales attract more legitimate customers, and that attracts fraudsters.
Overall, the most effective solutions entail layering on risk mitigation techniques, including passive biometrics and algorithms to screen consumer behavior, along with the device intelligence information. By layering all those together to provide a blended risk assessment, merchants can better cope with the shifting fraud landscape.