Measure — Measure — Measure
A common scenario is that organizations seeking to gain the benefits of process-based management first invest a lot of time and effort into business process modelling, then develop and use process analysis methods, and perhaps even realize good intentions in some form of launch of the 'new process approach'. So far, so good. After a short while, however, the process emphasis is waning, the initial enthusiasm is dwindling, and the new process approach is starting to look a bit old. What happened?
Maybe lots of things, but a common catalyst is that process performance measurement was not sufficiently emphasized. Bottom line was that there was no bottom line. There was no way to showcase the benefits delivered as a result of the additional effort required to do process-based management.
The most important question is "What's the problem we are trying to fix?" and we must have a detailed, specific answer.
A common assessment: "Yeah, we got lots of diagrams and models, and we fixed some stuff, but it was a lot of work and we don't have time to keep doing that. We've got plenty of problems and we're keen on anything that helps in solving them, but this process approach seemed more like a new problem, than a solution. Where's the benefit?"
Surround ourselves with evidence
What do we need to do differently? Measure. Measure. Measure. Surround ourselves with objective, measured evidence, firstly of the impact of the process problem or opportunity, and then with the benefits of success.
Surely, we can all agree that having the right performance data at the right time in the right format is a prerequisite for effective management. Of course, 'having data' is just one part of good management — it is necessary, but not sufficient. We need to use that data to determine where, when, and how we should actively intervene to secure a well-understood organizational performance outcome.
It is also obvious that we need good process performance measurement if we are to prove that process improvement has worked, that it has delivered real benefits. "Seems faster" or "probably cheaper" or "customers seem happier" are not the evidence we need for a solid business case. In developing support for process-based management there is no substitute for credible evidence of proven positive results.
A measured response?
Why is it then that process performance measurement is both an obvious requirement and commonly absent management tool? I think there are four main reasons — all of which can, and should, be overcome with appropriate responses.
- Lack of motivation. Why bother? We already have lots of KPIs. Some people immediately see the benefits of process-based management, some never get it, and the majority don't care much either way. Benefits need to be defined and continually sold. Organizations always have functional KPIs (think KPIs related to boxes on the organization chart), but not so many that measure cross-functional processes, perhaps none. For process measurement to gain traction, the benefits of process-based management must be clear and accepted.
- Wary of measurement. In this organization, measurement is a precursor to punishment. A manager in an IT services company resisted process improvements that would radically improve a serious customer satisfaction problem. Why would he do that? Because he knew that he would be sacked for not improving the problem earlier. That sort of organizational culture will never be conducive to active process performance measurement — to believing that continuous problem finding is an obvious prerequisite for continuous improvement.
- Swamped by measurement. You've got to be joking — we've got hundreds of KPIs already. This is true. Organizations inevitably have lots of informal and formal performance measures. They are given different names, they are derived in different ways, and they have different impacts on management. Mainly, though, these existing measures are functionally bounded; i.e., they relate to something that is within the control span of one organization chart box. This says nothing about end-to-end, or customer-to-customer, performance. Measuring the critical few KPIs of the high-impact processes will, over time, reduce the aggregate number of active KPIs as cross-functional process performance becomes more important than isolated points of performance. Process performance measurement rollout can be done one process at a time.
- Daunted by perceived complexity. How can we have KPIs for our thousands of processes? The idea of 'measuring all of our processes' is obviously daunting. However, not every process needs to be measured. A hierarchy of processes is defined by the process architecture. The measurement of the performance of any process also reflects the performance of its subprocesses. Starting at the top of the process architecture, and identifying the subset of high-impact processes, allows the capture of practical process performance without the need to continuously measure every process. Which processes need to be actively measured will also change over time. Process performance measurement need not be complex if approached in a structured, methodical way.
These four reasons for a lack of effective and sustained process performance measurement can be summarized as: a lack of knowledge of the benefits, an organizational culture that is not measurement friendly, mistaken concerns about the enormity of having to actively measure a very large number of processes, and unnecessary concerns about the complexity of building a new measurement system.
A practical approach to process performance management is needed that will deal with all these potential barriers to success.
Taking the measure of process
The first step is to document a process architecture from the top down, to visualize the processes that execute the organization's strategy. I've written about this in detail before here, there, and everywhere. In short, start with the strategy and the promises it makes to customers and other stakeholders. From those promises (value propositions) determine the highest-level core processes. Decompose those processes down a couple of levels to give a comprehensive and coherent high-level process architecture.
A common mistake is to then try to measure and manage every process named so far in the architecture. Even with just three levels of the hierarchy defined this can be at least 50–100 processes, depending on the number of highest-level core processes. It doesn't need to be like that. Start with just a few processes — with just one if you prefer.
For the selected process(es), determine the measures or KPIs (or whatever terminology you prefer). If this process were performing as well as its key stakeholders would want it to, what would it be doing and how would we know? What are the critical few measures? That is, if we could only measure three things about this process what would they be?
It is critically important to determine the measurement method for each KPI; i.e., where is the performance data to come from, how will it be gathered, and by whom? If it's not reasonably easy to gather credible data in a timely way, the measurement of process performance will lose its appeal.
We have the KPIs; e.g., customer waiting time. Now we need to set specific targets; e.g., 98% of customers wait no longer than three minutes. It might be more complicated than that if, to continue the same example, there are platinum, gold, and silver status customers, with each getting a different service level. Targets might also vary by time of day, day of month, etc. The target for a KPI might be a matrix rather than a single number.
Targets change over time. What was once important to measure is now not so important. Perhaps some measure we had originally chosen is now stable and not changing — we don't need to measure a constant. A new process problem or ecosystem change might require different measures. This is not set-and-forget; it requires a continuous review cycle.
Collect and analyze the performance data and report the outcomes so that evidence-based decisions can be made about where to best invest in process improvement. Get the circles turning!
Taking the measure of management
Some processes are harder to measure than others, but none are impossible to measure. Process performance can never be simultaneously both important and unmeasurable.
If we are not measuring process performance then we are not doing process management, and we cannot know if we are doing process improvement.
# # #