While most agree that the FMEDA method provides very accurate diagnostic measurements, it is understood that overall results depend on a good component failure database that matches the application environment. A database calibrated and compared to well analyzed field failure data offers high accuracy, perhaps the best. The web seminar describes the FMEDA technique. This web seminar also then describes Predictive Analytics, a field failure analysis technique that utilizes known design information to verify quality of the data. The results from over ten billion unit operating hours of field failure studies are compared with FMEDA results. The resulting improvements to both the component database and the field failure data collection systems are covered.