Maureen Fleming of IDC spoke in the Process Intelligence and Performance Management track on process measurement, and how it’s used to support decisions about a process as well as having an application context. She defines strategic measurement as guiding decisions about where to focus across processes, providing information on where to improve a process, and supporting fact-based dispute arbitration.
She showed a chart of timeliness of measurement versus complexity:
- Simple and timely: measure and spot-check performance within a process
- Simple and time critical: need for continuous measurement and problem identification within homogeneous processes
- Complex and timely: regular reporting to check performance across heterogeneous process islands
- Complex and time-critical: need for continuous measurement and problem identification across heterogeneous process islands
Leading enterprises are moving towards more complex measurement. I’m not sure I agree with her definition of “timely”, which seems to be used to mean “historical” in this context.
She breaks down measurement tools by the intention of the measurement system: what happened (process intelligence and reporting)/what will happen(analytics, complex event processing)/what is happening (BAM)/why it is happening (root cause analysis))/how we should respond (intelligent process automation).
She went through IDC’s categorization of BPMS — decision-centric automation (human-centric), sensing automation (integration-centric and complex event processing), and transaction-centric automation (integration-centric) — and discussed the problem of each BPMS vendors’ individual BAM creating islands of process measurement. Process metrics from all process automation systems need to feed into a consolidated process measurement infrastructure: likely an enterprise process warehouse with analytics/BAM tied to that more comprehensive view, such as ARIS PPM.
She discussed KPIs and how the goals for those KPIs need to consider both business objectives and past performance: you can’t understand performance variations that might occur in the present without looking at when and why they occurred in the past.
Although her presentation mostly focussed on process measurement, the Q&A was much more about sense and respond: how to have specific measurements/events trigger something back in the process side in order to respond to an event.