Fraud Detection: a Means to Remedy Losses
- 71% of the banks say lack of resources keep them from effectively detecting and preventing fraud. A study released this month by Novarica shows that 76% only deploy simple identification of account holders' devices (cookies, verification of the user's IP address) and 76% rely on knowledge-based authentication (secret questions). Properly speaking fraud detection tools are only being used by 59% of those interrogated (e mail or SMS alerts). As of OTPs, only 53% propose them.
- Neural network-based fraud detection have been deployed by 24% of the banks and almost half of the others intend to opt for this kind of technology within the next 12 months. Also, 24% propose out of band authentication (to avoid eavesdropping). The most complex tools (for instance, those implying cardholder’s geolocation) are being proposed by 12% of the banks, yet, there again one third would like to use them.
- Cross-channel fraud detection is still poorly addressed. The cost of these solutions seems to be accounting for these delays.
- Fraud detection tools are highly popular with customers (see box dedicated to BillGuard). This study highlights banks’ interest in these solutions despite obvious implementation backwardness.
- Hardly quantifiable but unquestionable benefits in terms of image have to be considered just as well as long term cost savings (see September 2011 Insight).