First of all, this system is not a sharpshooter that can always identify a single person in a crowd, but it will direct you to the right group so that you can  find the one person you need to spot. There are two types of error rates.

False rejection: Blacklist Identification marks a caller as fraudulent, even though it is legitimate.

Miss: Blacklist Identification fails to identify a fraudster.

These error rates are 10% in a 70-person fraud list. The system marks calls that are scored higher than the threshold score that we determine. By looking at only 10% of potentially fraudulent calls, you will successfully find 90% of fraud.

For example: Think of 70 people who you know are real fraudsters and are enrolled in your system as such. On a given day, 10,000 calls come into your contact center, but you know not all the calls are risks, such as calls from customers trying to obtain information, so you narrow the number of potentially risky calls to 700. Additional information indicates that there are 10 actual fraudulent calls in this group of 700. Let’s find out what Blacklist Identification can do with the calls.

The system misses 10% of fraudsters, so we assume that it will find 9 out of 10. The system also creates false alarms, so 69 of the remaining calls (690) are marked as false fraud. As a result, you will we listening to only 78 calls instead of 700 risky calls to find those 9 frauds out of 10.

Remember that success rates depend on both your fraud list size and the threshold. A larger fraud list will lower the system’s success rate.

With a high threshold, the system will mark fewer calls as scoring highly and create fewer false alarms, but it will also miss more real fraudulent calls. Listening to fewer call will mean finding less fraud. With a low threshold, the opposite is true.

Related Products: VB- Blacklist Identification