There are two key best practices for optimizing your card payment fraud strategy: profiling your customers, and then using the information for transaction monitoring.
During our second session of Fraud Office Hours, an attendee asked, "What are the best fraud prevention tools?" Watch this video clip and read below to see how Unit21's Head of Fraud Risk, Alex Faivusovich, responded.
Best Practices to Deter Card Payment Fraud
"With card payment fraud, there are two elements to have a successful strategy.”
Element 1: Accurately Profiling Your Customers
This means understanding the true behavior of each and every customer that you have. Not only what the average transaction type is, but you want to drill down as much as you can to get really good insights into how they operate on a daily basis. This is the first step to combating card payment fraud.
We are creatures of habit, so chances are that if your customers are using ATMs, they would draw cash regularly.
They pretty much have two or three favorite locations where they will usually go to the ATM and take money out. One of the places will probably be next to an office. Another one will be in approximate relation to their home address. And a final ATM will be located either around where they do grocery shopping or some other location of interest to that customer.
When doing customer profiling, try to profile the locations that the customer goes to over and over and over again, and try to understand what time of the day they usually perform those types of transactions.
For example, a customer who always takes out cash on their way to the office will usually visit an ATM in the morning. On the other hand, you will have some customers who will try to take out cash on their way home, or maybe they'll do that only on the weekends.
Try to drill down as much as possible and create those profiles within your database.
Element 2: Use Profiling Data During Transaction Monitoring
Once you have accurate customer profiles, you'll want to expose this information to the vendor you use for transaction monitoring. This allows you to prevent card payment fraud by writing rules and looking for a deviation or anomalies in the transaction that you try to flag.
Good customer profiling can make your strategy much more sophisticated, but you should also take into consideration having some type of hybrid approach between machine learning and some type of risk scores that are associated with transactions and the traditional rule-based approach.
If you have in-house capabilities to build your own models for machine learning and try to predict your customers' behavior and rescore the transactions or the entity, that's great. If not, you have many vendors today who will do that for you.
One of the latest features that we have at Unit21 is to provide scoring for the alerts that you generate in Unit21 that is based on how the entity interacts within your ecosystem, and takes into consideration how your operation teams are working, how they are distributing different alerts, and how fraud is trending actually on your platform.
We provide that to our customers today. The hybrid approach is the smart way to go about it where you can use your best practices of rules and transactions and thresholds and scenarios that you're familiar with, but also take into consideration machine learning and some type of risk-scored approach.
Looking for more insights? Check out our second session of Fraud Office Hours on-demand for a deeper dive into current fraud trends and which preventative measures to consider.