Overcoming Cross-Channel Challenges With Next-Gen Fraud Prevention

Today’s credit union members enjoy the freedom of accessing their accounts through a wide variety of connected devices, including mobile phones, computers, tablets and smart watches. Yet, with all the conveniences they bring, these devices also provide many points of access to members’ financial data, putting them and their credit unions at risk of falling victim to cross-channel fraud. For a long time, fraud solutions have leveraged machine learning and pre-programmed rules to catch criminals in the act. However, today’s interconnectivity has allowed fraudsters to innovate their tactics, and static rules do not prevent against new attacks.

More than a decade ago, adaptive behavioral analytics was invented in the U.K., creating the next generation of fraud prevention. By understanding individual behaviors, these analytics learn what activities are genuine and what looks suspicious.

The Cross-Channel Challenge

In a recent Featurespace fraud survey, we found that 62% of U.S. consumers feel they are at higher risk of fraud today than they were two years ago. Overall, this sentiment should come as no surprise because there are simply more targets to attack; however, it emphasizes a considerable threat that is only becoming harder to stop. More channels provide criminals more opportunities, plain and simple.

EMV secured card-present transactions quite effectively, but the rise of e-commerce made card-not-present transactions more prevalent, revealing a major flaw in payment security: Authentication. Accurately identifying who a person is when they’re not right in front of you isn’t easy, and criminals exploit that as much as possible. Personal information can be collected from an unassuming victim by phone through social engineering and used to access an online account or open an entirely new one. This is an extremely common scheme used to target older individuals. The CFPB reported that the number of Suspicious Activity Reports involving exploitation of the elderly jumped from 1,300 per month in 2013 to 5,300 per month in 2017 (an increase of more than 307%). Further, some members don’t question the legitimacy of a fraudulent message or call and fall right into criminals’ hands by providing personal credentials.

Although it’s a very basic example, social engineering demonstrates how difficult cross-channel fraud is to detect, and when you consider just how much more complex criminals’ tactics can become, the task can appear impossible. Data breaches that expose critical information allow fraudsters to develop synthetic identities in large quantities and then test credit unions’ systems for weaknesses by attempting to open accounts. Further, members interact with their accounts in many different ways, and they collectively indicate typical and atypical behavior. For example, if someone always logs into their account with a desktop computer at 3 p.m. every day, then a mobile log in to that account at 1:30 a.m. from a different state would be a major red flag.

Fraud isn’t always this obvious though, and the focus must now shift to understanding individual members’ behaviors versus having rigid parameters in place.

Next Generation Fraud Prevention With Adaptive Behavioral Analytics

Adaptive behavioral analytics are the result of analyzing where, when and how the member interacts with an account and makes purchases, and then risk scoring those behaviors based on the historical profile of the individual in real time. When combined with machine learning, this is the most powerful weapon in the fight against fraud. Self-learning algorithms evolve with the member, detecting the slightest deviation from the norm across all channels.

What’s more, adaptive behavioral analytics reduce false positives, which can be as damaging as actual fraud, because it means a legitimate transaction is blocked. For the credit union, this is lost business and for the member, it’s an incredibly poor experience. Think of the embarrassment or inconvenience when your card is incorrectly declined – how did you feel? Our survey found that more than 51% of respondents had a genuine transaction declined because of suspicious activity and 44% said they felt frustrated. False positives won’t immediately drive a member away, but they certainly influence the dependence on and use of a credit union’s products, and over time, this can cause irreparable damage to the relationship.

Leveraging adaptive behavioral analytics requires advanced ability to ingest data at a large scale, develop models and analyze data in real time; however, most credit unions lack the resources to develop a system that can achieve this. As a result, single-channel fraud strategies are relied upon, and in today’s interconnected world, that leaves many areas exposed.

The industry is never going to completely solve the fraud problem, but we can consistently mitigate its impact. Members want to know their credit unions are doing everything they can to keep their accounts secure, while also receiving seamless access to their accounts. Combined with machine learning, adaptive behavioral analytics provides a powerful way to do both.

Dave Excell is Co-Founder and Chief Technology Officer for Featurespace. He can be reached at dave@featurespace.com.