There are a lot ow ways to handle this. For the ultimate in flexibility and power, I would suggest you look at
TestStand. This is an excellent test manager. If you want to go your own route, one thing you can do is break your test into subtests. Instead of one large test that does several setups and diffreent measurements, make each setup and measurement separate. Then you can load a file based on serial number/product id that has a list of tests to perform, i.e. test1, test3, test4, test7, etc. Either of the two shipping examples on test managers (Test Sequencer and State Machine Test Sequencer) can be modified to run only those tests that are contained in the file. Another way could be to load a file with all tests listed but additional columns for subtests to be performed or not performed. In your program, put the optional tests inside case statements that are wired to the Boolean values read from the file.
