As we saw previously Process Mining differs from Data Mining, even if both are clearly 2 pure data driven approaches. Why are they different ? maybe just because a process cannot (or that would be difficult) be seen in a tabular view like we used to manage and analyze data. Indeed when we analyze a dataset, this one is most of the time converted first in a tabular format to be exploitable by analytics tools. On the other hand, a process cannot be displayed easily in a flat format. Its nature is by default a flow (if we adopt a design view) or a tree (if we look at the process after execution for example).
Note: Of course, we can have some overlap in these two kinds of analysis. For example when drilling down to attributes and/or dimension value analysis. These investigations are really common in Process Mining (once we are focused on a specific Process Step or behavior for example) and are also really common in Data Analysis.
So instead of analyzing columns and rows in a table when we analyze a process we want to understand for example:
- How it is executed (its execution flow)
- Its global duration (we may have to take in account different types of period calculations like: the takt time, the cycle time or the lead time)
- The Process steps and how long takes each of those for example or between the main milestones
- When and which step starts/ends the process
- If there are some repetitives events in the process
- Can we identify the main rework loops ?
- Its different behaviors (we also name as path or variations)
- If there are some bottlenecks in the flow, and to be able to qualify them
- If there’s some violation in the flow (Protocols)
- By business dimensions (for the global process or just one or several steps) drilling down into some dimensions to highlight new business outcomes
And obviously many other information on the process itself, but these are the most common. But the very good news is analyzing a process should almost follow the same path as the kind of information linked to a process are almost the same. This is also the purpose of this document, providing some guidance around how to manage a process analysis.