Unlock the Value of Operational Data

Many forward-thinking industry executives now embrace the operational excellence mindset and approach and hold the belief that the best run organizations must operate not only productively, but safely and reliably.  With more and more companies adopting process-based enterprise data platforms, predictive analytics offer a very promising approach to help organizations prioritize their investments and focus their efforts in risk mitigation. It may also allow the discovery of new solutions that are not likely to emerge from conventional observational programs or safety perceptions surveys.

This study aims to answer the research question that whether a combination of metrics that represents an asset’s incident management process behavior can be found to effectively reflect the asset’s risk reduction capability and therefore predict its safety performance.  It stresses the importance of data quality and integrity for predictive analytics to avoid misleading analytical exercises.  One of its key objectives is to identify proven benchmark indicators that are actionable, objective, routinely measurable from day-to-day activities, and are believable and predictive.

The presentation will highlight the rigorous process used in the predictive analytics as well as the systematic steps and change management approach to operationalize the benchmark indicators at local operational units. It features a case study of an energy company’s effort in applying predictive analytics to gain insights from the operational data that they collected from an enterprise management system. The case study covers more than 900 million work hours with over 30,000 incidents and identifies benchmark indicators such as team competency, risk-driven actions, root-cause based mitigation, and learning-mindedness for estimating the corresponding injury rates (adjusted R² > 0.8). The predictive model was further validated using a new set of data.

Findings from this study showcase how organizations may leverage a process-based information system to drive performance by capturing day-to-day reporting behavior, measuring responsiveness and diligence in process execution, capturing the lessons from successes and errors, and allowing “learning” to take place systemically. While this study provides statistical evidence to support the strong correlation between process behaviors and safety performance, it also emphasizes the organizational factors, such as engagement, that are necessary to enable and sustain effective risk mitigation efforts.