This month’s topic of my weekly blogs is around the wonderful capabilities of #process mining, one of the hottest topics around in the world of #BPM. After the gigantic popularity of the more traditional Business Intelligence during the 2000’s and early 2010’s, it became apparent that one of the major drawbacks of #BI is that is focuses on concentrated and mostly isolated parts of a process. It does not take into account the journey a process travels from start to finish, but rather focuses on one or more of the stops along the way to find out what happened there. If I can make an analogy to taking a trip, than process mining looks at the trip it self (where did you start, where did you go to, in what sequence, how long did I take you to get there and, more importantly, if you took that trip multiple times (for example your daily work community) can you identify differences based on other factors than yourself why the journey duration variates. BI on the other hand, looks much more to the stops in your journey and analyzes (in more depth that process mining typically does) what happened during that stop.
Process mining has seen a renewed surge in attention over the last 5 years, but it has been around for already 20 years when the first commercial process mining solution (ARIS PPM) was brought to the market by @SoftwareAG. It could, however, not yet compete with the very popular BI at that moment and process mining evolved further in the shadow, driven by the academic community. It wasn’t until around 2015 that two students (doing their thesis at @SAP) developed a new generation, cloud-based, process mining solution (@Celonis) that focused much more on usability for end users, instead of being overly engineered and focused on data analysist. This event sparked a lot of developments all over the world and at the moment, there are three global leaders when it comes to process mining (Celonis, ProcessGold (acquired by UIPath) and SoftwareAG). However, this weeks blog is not about a market analysis (you can google on Everest Group Peak report for process mining for that), it is about the question: what is process mining?
Let me start by directly stating that process mining is not a solution in and by itself, it is a supporting technology that facilitates business process management, in the widest sense possible. Process mining reads and analyzes event logs from transactional systems. Many companies have documented their processes somehow (either in BPM platforms, or in PPTX slides on SharePoint (not recommendable by the way) and these business processes are implemented via transactional systems (the most famous examples, of course being SAP, Salesforce, ServiceNow and many more). These systems transform the input data into output data and they create event logs in the process. Based on these event logs process mining can reconstruct the way that every single process execution has happened in reality. This also defines the most basic of use cases of process mining, being process discovery. The pictures of spaghetti like structures that impress and scare the living daylights out of process owners are all too familiar.
However, this is by far not the only use case for process mining. Based on the reconstruction, process mining can also perform some powerful analyses on the transactional data and because it looks at the data from an end to end process perspective, it can also start looking for root causes for specified symptoms (you can think about long process cycle times, or high rework rates). Besides this, process mining can also perform conformance checking and this functionality compares each and every single process execution with a reference model (or baseline model) to determine if the execution of that single process was compliant to the reference model or not (and if not, it can determine the impact that this conformance issue has on certain main measurements).
Coming back to my earlier statement that process mining is not a solution, but rather a supporting toolset. Companies that take their process management seriously, also want to be able to measure and check the way the business processes are executed, in order to be able to verify if the documented processes are still in line with reality (and vice versa) and to see if improvements are needed (and where). At the moment that you implement process mining in isolation, you potentially miss out on the synergy you can get from linking it to your process management practice and you’ll be re-inventing the wheel with every process mining project you run in your organization.
I can talk (and write) for hours on this subject and I know I will have a lot of fun with this months topic of process mining. Next week, I will go into more details on the some of the use cases described above.
Happy Easter holiday (if you celebrate that), otherwise have a great weekend.
Ciao, Caspar
