Grounding Process Data Analytics in Domain Knowledge: A Mixed-Method Approach to Identifying Best Practice

The often used notion of 'best practice' can be hard to nail down, especially when a process involves multiple stakeholders with conflicting interests, as is common in healthcare, banking, and insurance domains. This exploratory paper presents a novel method that leverages both domain knowledge and historical precedence as recorded in IT systems to derive relevant dimensions, measures and behaviours representing best practice. To test our approach, we explored best practice in the area of injury compensation claims management involving multiple stakeholders. We evidence that best practice can be identified by semistructured interviews with stakeholders (a qualitative method) allowing their perspectives to guide the application of various forms of analytics on historical data (a quantitative method). This led to the identification of four best practice dimensions: process fairness, process quality, process cost, and process timeliness and their respective measures, which are then used to assess the performance of compensation claim cases (i.e., 'which claims are the best performing cases?'). By analysing the process behaviours of those cases through historical data together with additional stakeholder input, we propose to identify potential best practice behaviours.