In Bankers Tools it states: "As companies evaluate their different options, there are several key requirements to consider: data quality; flexible matching algorithms; minimizing false-positives; efficient suspect-match review; and extensibility." We are trying to establish the algorithm for % match and I am wondering what others use. We seem to think 80% is an industry standard, as that is what I have seen in other banks. What do you use, suggest?

We test the percentage match at which a hit will escalate to compliance for review; we test it for each system and transaction type that we bring on board for scanning. We take into account the overall volume of the transaction under consideration, the hit rate and quality of the matches and the data present in the transaction.

Not sure I understand your explanation. Do you scan only a percentage of the transactions? We scan all transactions but our new software will show a possible match at a percentage of the word(s), i.e. Wright Brothers' Aircraft Company is on the OFAC list, we have the software set at 75%, the wire comes through as Wright Brothers' Company, since that matches 75% of what is on OFAC, it will kick out as a possible match. I'm looking for what that XX% you use, if your software is able to do that. Thanks.

What I had hoped to explain was that knowing my FI's "% match" should mean nothing to your FI since the percentage is a function of:
the specific software we use
how it is configured
the number, kind and content of transactions that are scanned
the quality of the matches at different scores
the experience level of the people dispositioning hits at different match rates
our processes and procedures
our risk tolerance.

Instead of telling you the number I use, I'm telling you that you have a heap o'work in front of you to determine the right number for your FI.

Oh, I was hoping you didn't mean that! Thanks.

However if you want to set a % you can start at 80 - 85% and see how many hits you receive.
We are set at 85% and receive hits on common hispanic and asian names. I have trained the staff how to determine if it is a false positive when comparing dates of birth, full names, etc.