How to Use Identity Data and Behavior Intelligence for Fraud Detection

The world shifted in many different ways 18 months ago because of the global pandemic, not only in how we live our lives, but also in how consumers do business with their favorite companies and brands. New business models have emerged, new customer service and purchasing channels have been implemented, and consumers and businesses are interacting in different ways. 

With so many new unknowns, fraud detection remains a top priority for your business. The key imperative is to detect fraud in a way that protects your bottom line while ensureing an optimal experience for your legitimate customers.

In a recent webinar, DataVisor experts discussed ways to approach fraud prevention in a new landscape by using identity data and behavior intelligence to make accurate fraud predictions. Here are some of the highlights from that discussion:

The Need to Go Beyond Transactional Data

In the past, companies used identity data to verify legitimate transactions. For example, if a user purchased something online, the company might look to see if the customer’s name, phone number, address, and billing address matched up. If so, then the transaction would likely be considered legitimate. 

But as Ekata’s Travis Anderson pointed out, that’s no longer sufficient. “With so many data breaches in recent years, it’s not uncommon for identities to be compromised to the point that they can trick established fraud detection systems. These synthetic identities can pass through as ‘good’ transactions. It’s time to go beyond this, which is why we’re now including behavioral patterns alongside identity data.”

Anderson defined behavior data as “data that’s been consistently used alongside each other over time.” This could be an email address/physical address or phone number/address combination that has been used together for the past several years. 

It’s this history of data that can help to improve fraud prediction — determining the probability of fraud with greater accuracy.

Increasing Fraud Prediction Accuracy with Behavior Intelligence

While fraud teams want to examine identity data and come up with a yes-or-no answer regarding fraud, identifying fraud is rarely black and white.

“Fraud detection and prevention is starting to follow more of a probable model rather than a deterministic model,” explained Anderson. “Deterministic data would be like using a credit score to decide how much money to lend someone. But the current landscape is moving more toward probability signals. Someone might have a really great credit score, but if I can’t be sure I’m really talking to that person in an online transaction, then the credit score alone isn’t effective enough for fraud prevention.”

Anderson notes that using identity data intelligence and behavior intelligence serves a dual purpose: to detect fraud and to create a lower amount of friction for good customers. To find both of these, though, requires a keener focus on fraud prediction.

“The world of ‘confirmed good’ and ‘confirmed bad’ is going away,” shared Anderson. “There is no single silver bullet. We need to think of things more in terms of likelihood and the respective actions.”

Want to know more? Watch our latest webinar: Identity Data and Behavior Intelligence for Fraud Detection.