Operational data collected from process automation systems can be analyzed to accomplish many goals including improved safety, and minimized down time. However, data collection methods can vary substantially. This can impact the results by an order of magnitude. This web seminar describes the exida “Predictive Analytics” analysis method which uses benchmark limits to detect and correct problems in data collection methods. The exida Predictive Analytic benchmark limits are based on nearly 100 billion unit operating hours of field data from the process industries. Examples are shown where potentially misleading data analysis results were detected and corrected preventing potentially dangerous changes to safety equipment design.