There is no certification body for search and matching tools.

Testing these tools requires that, not only are you confident that all the data found is relevant, you must also be confident that very seldom is relevant data missed.

In systems that screen transactions against files of alert lists or other such negative data, the very normal and common expectation is that extremely few matches will be found. Designing testing strategies to prove that nothing relevant was missed, when the normal result is to not find anything, requires a lot of experience and skill in the testing of this class of system.

Many products that find "duplicates" in files have been purchased because of the high volume of duplicates discovered, when the critical criteria may still be "how many duplicates remain undiscovered!"

Failing to find anything is clearly a desirable and acceptable result of the process, but only if it’s true that nothing was missed.

By using software that has been used for long periods of time by organizations that have more critical needs and higher risk that your own, it is possible to be more confident that nothing is being missed.

The only way to be sure is by using software that allows controllable "overmatching". If the software can be controlled such that overmatched results can be made visible and matches can be intelligently ranked in relative order of relevance of match, it is now possible to audit the quality of work. Such "overmatching" is the only possible way to expose undiscovered "undermatching".