Target >Measure >Respond is the essential lifecycle of business process management. Set a performance target for a process, measure what is really happening, and do something if the results are not what they should be. This is it; all process management and process improvement comes down to this. If we aren’t improving processes, sometimes with radical effect, then we are wasting time. Process-based management is both as simple and as complicated as target>measure>respond.
Which process? Be clear about the boundaries of the process. What triggers its execution, and what results from that execution? Now that we know which process we have in focus, and if it were working as well as a consensus of key stakeholders would like it to, what would it be doing? And how would we know? There’s a lot in that. We need to know who the stakeholders are, and what they need from the process – what value is the process supposed to deliver? The performance targets need to be directly related to the expected performance. Set the targets, and look to have the ‘vital few’ rather than the ‘trivial many’ (as JM Juran et al. have explained).
For each performance target, it is likely to be useful to have upper and lower limits of acceptable performance; a band within which a consensus of key stakeholders would be happy.
Also, for every performance target, identify the ways in which the performance data will be collected (i.e. the measurement method). If you can’t measure it, or can’t afford to measure it, then the target is meaningless.
Having established the measures and the measurement methods, don’t stop – make them! This requires a very clear allocation of tasks. Data collection cycles might be anything from real time to quarterly, or longer, and the data might come from IT systems or gathered manually (e.g. via customer surveys). The logistics of process performance measurement need to be well designed and the data gathered, analysed and delivered reliably.
Make the measures (i.e. collect the performance data), but also analyse and report it in a useful way. Develop a strategy that makes process performance reporting both practical and sustainable. Across an organisation, there are many processes and many more process measures; so, avoid an avalanche of performance data, report by exception. Only report actual or emergent performance anomalies, and in a way that is easy to understand with available detailed analysis as required.
This seems obvious, but it’s a common failure point. Without an appropriate response, all measurement is a waste. The purpose of measurement is to correct problems, indeed to avoid them. In an environment where the management structure is concerned with overseeing individual functional areas, who will respond to cross-functional performance issues? This is the key role of the Process Owner.
The Process Owner, or whatever term is used, is often said to be responsible for the performance of the process – that is, we ask someone to take responsibility for something over which they are unlikely to have complete, or perhaps any, control.
A better summary of the Process Owner role is to say they are accountable for responding when the process performance is outside of the accepted range or trending in that direction.
How should a Process Owner respond? That will depend on the circumstances, but the options might include one of these:
a) commission a process-improvement project
b) request more analysis of the underlying data
c) collect more (or different data) to give deeper insights
d) discuss the issues with those involved in executing key parts of the process
e) simply go and watch the process being executed.
The Office of BPM is a life support system for Process Owners. The detail can be left to the Office and others, but strategic process leadership must come from the Owner.
If you are not doing this already, how should you start? Don’t wait. Don’t over-analyse. Don’t try to do it all at once. Pick one cross-functional process and implement the simple steps of target>measure>respond.. No management support? Do it anyway; just one process will prove the value. Start now.