The mobile phone has become an extension of most people’s lives. It is usually on or near every one of us, most of the day and night. In fact, the majority of us actually sleep with our mobile phones next to our beds! Now, more than ever, if you know where the mobile phone is, you likely know where the person is. Finsphere was founded in 2007 on this very premise – if a mobile phone is generally found on or near its owner roughly 24 hours a day, then that mobile device could act as a “proxy” for that individual – and could be a highly credible source for authenticating an activity, such as a financial transaction.
Banks today use a variety of scoring tools and fraud risk models to determine the likelihood that a credit or debit card transaction is valid or not valid (i.e. likely fraud). If the bank gets it right, they prevent fraud and issues with their customer. If they get it wrong, they likely deny a legitimate transaction, which means the customer has to find other means to pay or spend considerable time on the phone convincing the bank’s customer service that they really are the person making the transaction!
The premise we set out to evaluate was whether geo-location of a mobile device could enhance the authentication process. After establishing Finsphere, we quickly developed an extensive portfolio of pending and approved patents and began trials with some of the world’s leading financial institutions to validate the model in a variety of situations including:
- Card-present transactions – when a person and their credit or debit card are both present, like at an ATM or during an in-store purchase
- Card-not-present transactions – when the person and card are both remote from the merchant, like when making a transaction online from a desktop or mobile device.
Based on our ongoing bank trials, we determined that mobile phone geo-location worked very well as an added factor in reducing denials of legitimate transactions, but did not work well for fraud detection or card-not-present situations, by itself. In these cases, geo-location had to be combined with other financial and mobile factors to make it a stronger predictor for fraud. After years of evaluation, we did just that. We built an analytics engine that combines geo-location data with over 200 identified disparate factors. The results thus far? A significant decrease in legitimate transaction declines and fraud in card-present and card-not-present study groups.
Geo-location, alone, clearly is not enough. But geo-location, when used in an analytics engine that incorporates a multitude of other factors, significantly improves fraud model scoring. For the consumer, this means banks getting it right more often in the future.
I’ll be taking a deeper dive into this topic in future posts. I look forward to your feedback and hope you’ll join the conversation.