Data scarcity is one of the biggest challenges companies face when combating online fraud. This issue is compounded when each stage of the user journey is siloed into its own fraud workflow. For example, separating account opening from transaction. Additionally, if there’s a failure to include even lightweight fraud checks during account opening, the opportunity to gain valuable insights into data captured at this early stage is lost.
Account opening checks are a critical part of any fraud strategy. This is because they can identify risk at sign-up, but also because of how powerful account sign-up data can be when combined with transactional data.
Ekata’s use-case–driven products — built specifically to address fraud challenges at account sign-up and transaction — enables companies to employ Ekata at multiple points in their fraud strategy. While many companies initially implement Ekata at the time of transaction, due to the more easily and clearly defined financial risk at that stage in the customer journey, the use of Ekata data insights often expands into account opening, account modification, and beyond.
How a Marketplace Protected Its Bottom Line by Employing Multiple Fraud Checkpoints
When working with marketplace companies, it’s crucial to prevent fraudsters from ever entering the platform in the first place. In one case, Ekata had been integrated at account opening for a year before the marketplace saw a large-scale collusion fraud attack. We immediately worked to integrate Ekata at the point of transaction, building on that initial risk assessment at sign-up to further deter fraudsters and prevent future fraud attacks. After preliminary data testing at the transaction point, we found a rule that identified fraud with 94% precision and about 5% recall.
As a next step, we broke down the silos containing the account opening and transaction data sets by layering in the signals from the sign-up workflow into the transaction flow. Following this, we incorporated fraud signals from the account opening risk checks. We looked at our Identity Risk Score, Network Score, and Email First Seen Days — some of Ekata’s overall top performing signals — and achieved a staggering improvement. By layering in these additional signals, we were able to detect 5% of fraud with 99% precision. This is an almost unheard of precision for simple business logic.
By breaking down data silos and including the risk signals from the account opening stage, we were able to improve precision by 5%. For a business focused on stopping fraud without impacting good users, the precision improvement we were able to achieve by layering the sign-up signals at transaction made all the difference.
Find a new way to look at the signals that are already available in your fraud workflows to dramatically improve your fraud strategy at little to no added cost. This often can be done by layering in signals captured at different points in the customer journey.
As fraudsters become more sophisticated, so must the fraud prevention strategies companies employ. By thinking about your fraud strategy more holistically, you’re setting up your company for greater success in fighting fraud in the digital world. Data scarcity will always be a challenge. This is why it’s so important to use every piece of captured data to its fullest potential.