Too often, the theory, practice and aspirations of process-based management are reduced just to process modeling. A stated intent in some organizations is that “we will model all of our processes”, and this is said as though achievement of this goal would solve some serious business problem. In reality, there is never a business problem called “we don’t have enough process models”, and an inappropriate focus on modeling severely handicaps the achievement of genuine process management.
Why model processes?
There are four general purposes for process modeling: document, understand, improve, automate.
- Document: describing how a process is meant to work enables consistency, quality and standardisation
- Understand: understanding the detail of how a process works gives insight into what works well and what might be able to work better
- Improve: discovering ways in which the process might work better is enabled by modelling and testing the possible changes
- Automate: using the process models as input to a process execution system is increasingly the way that IT applications will be built.
Enterprise Process Architecture (EPA)
Every organization should have an Enterprise Process Architecture (EPA), a hierarchical model of the first three levels of its business processes. Individuals processes in the EPA, and there may be hundreds of them, need only be identified as a single object (a box of some form depending on the graphic notation being used). The EPA model is about identifying relationships between processes, rather than the details of individual processes. Modeling the process detail is then undertaken on demand; it’s done to address a particular problem or opportunity. Therefore, detailed process modeling should be done as part of a process improvement project that has a very clear organizational performance goal.
Should we model all of our processes?
Some processes will never be modeled if there is never a good reason to do so. Looking through an EPA, we should find that some processes have been modelled in lots of detail and to many levels, others will have been modeled only at higher levels, and others will have never gone lower than the original single box. Addressing process issues will always be a matter of prioritizing scarce analysis resources. Modelling ‘everything’ is waste.
Next time someone says “We should model all of our processes”, ask the important question “Why?”