What is process mining?

Digitalisation of key business processes has created an explosion of opportunities to measure and extract data from processes. Broadly speaking, whenever a process happens in IT systems, low-cost and low-effort collection of data is possible.

In all its simplicity, process mining is about collecting, analysing, visualising, and acting on data that various processes create. Compared to more generic business intelligence, process mining is tailored to explain the processes and especially answers WHY some process inefficiencies occur. Process mining therefore enables organisations to have a more objective view of their operations and makes data-based decision making easier.

Process mining in practice

Process mining requires at minimum three types of data points from each process step:

1. Case ID: Unique identifier of the item flowing through the process steps

2. Activity: Unique identifier of the process step itself

3. Timestamp: When did the case go through the activity

(+ some additional case or activity attributes)

In other words, each process step creates its own set of three data points. By aggregating these to a table, we get “event logs”.

Example of an event log

This table format however does not tell us much. The true power of process mining comes from its visualization and explanatory capabilities. Combining the various cases that flow through the IT systems can be merged into a visual format. The charts below showcase how the most common process variant gives a highly different picture compared to the spaghetti diagram of the messy reality.

How processes run in reality is much different from the ideal path.

Works like magic. To this interactive real-time process map, you can combine your ideal process map, use this to analyse how your process conforms, use the tool’s intelligence to get closer to root causes of the processes, receive impactful improvement proposals, and depending on the tool, do much more.

Limitations of process mining

It goes without saying that no software or tool, including process mining, should be viewed as a silver bullet to all organisational problems. However, the role of the right tools at the right places should not be downplayed either.

• Data limitations: Even though the modern process mining tools are suitable to practically all systems, some legacy systems don’t simply have the necessary data points, e.g. time stamps or unique case identifiers. Even though process mining is a powerful tool to bring clarity to fragmented IT, data capabilities are important to analyse early during the process mining projects. With new system development initiatives, we often design the system data models to process mining requirements.

• Non-IT work: Process mining relies on data created from actions in IT systems. Obviously, processes not working on top of any IT systems, or work happening by people between the actions in IT systems are not visible. Task mining (we will come back to this) in one sense tries to tackle a part of this limitation, however, this is a limitation that will persist in some processes more than others. This just highlights that process mining can increase the speed and impact of operational excellence teams, but it does not replace them.

• Case orientedness: This part goes a step more technical, but currently process mining tools are hindered by so-called “divergence” and “convergence” issues. This is solved by a new process mining technology called Object Centric Process Mining (OCPM).

Opportunities of process mining

Process mining tools are quickly developing to overcome some of the limitations and expanding to new feature areas. The three most exciting opportunities to continue process mining efforts are:

• Execution management: Traditional process mining focuses on visualising processes and finding the root causes for problems. Next logical step is to act on removing these root causes, where execution management functionalities step in. This functionality enables you to trigger actions directly from the process mining tool (e.g. divide shipment into two in SAP, initiate and close vendor bidding,...), making it a true process management tool instead of merely an analysis tool.

• Task mining: Process mining works on “centralised” IT systems, including ERPs, CRMs, etc. Task mining has very much the same idea as process mining, but is built to extract data from users’ desktops. Read more about task mining.

• Object Centric Process Mining solves the convergence and divergence issues of traditional process mining. It is a big leap in process mining tools’ capabilities to analyse whole operations of your organisation. To dive deeper, take a look at this extensive article.