Better Data, Better Decisions, More Revenue
By Aaron Press, Director - LexisNexis Risk Solutions
According to the LexisNexis True Cost of FraudTM study, verifying a consumer’s identity remains the most frequently cited challenge when approving or denying ecommerce transactions. But validating your customer’s identity has become more and more challenging.
There are many places in the transaction flow where adding robust identity data can improve the ability of the models to assess transaction risk. Increasingly, we are finding that adding predictive identity attributes early in the process allows a merchant to look more deeply into the consumer’s information and integrate that intelligence directly into the rules and scores. The result is more accurate decisions and a better bottom line.
For example, LexisNexis Risk Solutions recently worked with a merchant selling a digital product requiring near-instant delivery. This merchant was seeing high levels of fraud from new customers, but traditional methods of verifying identities, such as phone and address verification, were failing too often. After analysing the merchant’s data and understanding the challenges, we recommended adding additional identity information available in LexisNexis Order Score Attributes. With richer data in the rule set, including the use of deep attributes about phone numbers and more accurately matching identities to addresses, the merchant was able to build a model that captures significantly more fraud, reduces the need for manual reviews, and creates a better customer experience. And, when a manual review is required, the agent can more quickly zero in on the specific discrepancies, allowing for faster, more accurate decisions.
This approach requires some up-front work to do the analysis and build the models. Each merchant faces a unique set of challenges, fraud vectors vary across verticals, and which combinations of attributes and rules will be most predictive will differ depending on a variety of factors. If you are willing to dig into the data, the rewards will be worth it.