There should probably be some kind of process to prioritize different test structures, and who's got time to characterize them