I recently returned from giving the keynote address at the 2012 Financial Cryptography and Data Security conference. The event draws crypto/data security/privacy experts, researchers, and practitioners across the world to focus on the latest advances in this space. During this week-long deep technical conference, a central theme was how to balance the need to implement security technology with minimal impact on customers in terms of privacy and inconvenience.
Conference attendees were keenly interested in how FICO leverages analytics to combat fraud across the banking sector. One question that arose was why banks don’t adopt more secure authentication techniques—for instance, one-time key fobs and facial/voice recognition. We discussed weighing the value of these techniques with the resulting customer hassles and privacy concerns. In the payment card space, strong analytics allow banks to find that balance between keeping fraud in check, while not unnecessarily complicating the customer experience at point of sale and impacting interchange revenue.
Another topic we discussed was how to determine which data elements are captured and utilized in the development of our fraud models. This is an area that may get quite exciting as we look to mobile payments and new alternate payment channels, where new data elements and context-based feeds may be captured. Analytics will be critical to best fusing these new feeds/detectors with current data feeds and technologies. The additional data and insights will continue to drive analytic innovation in these new payment channels and allow better fraud detection with limited customer impact.
As is typical in this venue, we discussed recent compromises of data, vulnerabilities of payment channels, Man-In-The-Middle mobile intercept, and recent stories of compromising NFC (Near Field Communication) cards from a distance. These trends reinforce the need for advances in analytics to keep fraud losses in check, while minimizing customer impact and inconvenience.
In this regard, FICO continues to drive new analytic innovations. We are working on the use of text analytics to derive an even deeper understanding of customers in order to improve fraud detection and reduce false positives. In addition, we recently filed a patent application for a major improvement to our self-calibrating fraud analytics, which improve detection under dynamic conditions. I will discuss this topic further in a future blog post.