This is a situation which needs to be treated carefully: when two or many Process Steps happen exactly at the same time! The consequences can be structural as the order of a Process Step can create a new variation in the Process Flow analyzed itself. But, why and how does this situation occur ? Most of the time it is just because the data collected was not precise enough (we were not able to collect the seconds or better) but it could also come from the source applications themselves.
In the example below we have 4 steps in the Process Flow. Unfortunately the Step 3 and the Step 4 (different) occurs exactly at the same time:
We then have to option for this simple situation. We can decide to use first Step 3 or Step 4 …
There’s two options to manage this situation:
- We unduplicate at the sources (before importing) the data based at the right time level (Timestamp) and during the preparation stage. This solution can be a good solution if the business is able to provide solid rules to manage that (which is mostly not the case). This option is quite dangerous as we remove some steps in the Process, so doing such action must be a wise choice validated by the business.
- We specify an order for these simultaneous Process steps. By this way when the Process Mining solution detects some groups of following events with the same timestamp, it then places them in the order in which they occur most often. The Process Mining can normally change this order himself. The Process Mining can also detect groups of Process Steps having the identical timestamp. Then it could place them in the order in which they mostly occur. By this way the business user can decides himself order which makes more sense.