Global - en
Choose your preferred location
Global
English
Germany
Deutsch
Sweden
English
Norway
English

Process Mining

Elegant and sustainable process optimization

In recent weeks, I have been engaging with a highly topical issue for my first webinar: Process mining – system-supported, objective and continuous process analysis. This subject is not entirely new to me, but nor is it something I am intimately familiar with. That changes as of now!

First of all, how were processes analyzed before now?  

Interviews with employees in the various divisions. The departments describe their current ‘process blockages:’ Where do they feel they are being held up, what is currently standing in the way of an optimal process? What is said is subjectively spot on, but does it always set the right focus? How can one objectively demonstrate how subjectively perceived problems significantly influence value creation? Without an extremely experienced consultant at your side, it is possible that processes and functions may be optimized without creating any genuine added value overall.

System evaluations are a first, important step toward process mining. Too often, systems and the internal processes are considered in terms of function or only in sections. Lists about the use of transactions or the frequency of use of certain document types have an informative purpose; often, only certain fields are evaluated. Function is the focus here, but it is the process that is to be mapped. Such evaluations come to an abrupt halt when the system reaches its limits at the latest. In addition, this procedure is so complex that, once done, it is often repeated only some years later. This is not conducive to continuous optimization. 
So how can process changes, once initiated, be tested – do they continue to add value as hoped or is further fine tuning required? 


Process mining provides clarity

This is precisely where KPS and its partner Celonis come in: by providing a continuous and sustainable view of processes. Process mining is not used purely for a one-time initial optimization, but for ongoing process analysis and improvement. How does it work in practice?

There are many possible applications and almost any business process can be analyzed. Examples include:

  • Traditional procurement and fulfillment processes, such as Order Purchase to Pay, Order to Cash or Dropshipment

  • Analysis of financial data and flows

  • Processing, frequency and quality of returns and complaints

  • A system analysis for a SAP S/4HANA migration


Process mining in defined steps

Step 1: The happy path  

Data are collected from and provided by various systems – not only SAP, but all common ERP systems. All the systems involved in the process are connected to the cloud and the data are transmitted. The data are then used to determine the happy path for the process concerned: This is the workflow that the data indicates will be performed the most. This is done graphically in a user-friendly and comprehensible manner. 


Step 2: Validation

The next step involves the analyst, who validates the happy path based on the result. Is this fundamentally the correct procedure or is there already potential for optimization? The result could show, for example, that the main process step is mostly initiated manually by the user, indicating all deviations from the main process. And this is where a more accurate picture usually starts to emerge: workflows that have crept in over the years become visible – including workarounds and special processes. For the analyst, it is now a matter of understanding the causes and working out the optimizations.

 
Step 3: Leveraging potential

The third step is to implement the necessary measures to really leverage the potential of the analysis.

Examples of possible improvements include:

 Example 1: Increase process automation; in other words, by replacing the previously manual steps with automatisms. Note: There is a big lever here – according to the SAP study “SAP’s Vision of the Retail Future” from January 2020, 58% of today’s manual steps will be automated by 2025.

Example 2: Eliminate special workflows that slow down the process or make it expensive.

Example 3: Lay the basis for machine learning. The system identifies automatically, based on the historical data, whether a particular event could also occur for future documents and data, and will indicate this directly. This allows the process to be supported proactively.


Step 4: Continuity

The final step is the continuous monitoring of the processes together with the implemented measures. This means that the effect of the measures can be analyzed – and if necessary – controlled directly or a swift intervention implemented without additional effort.

Conclusion

Applications such as Celonis can lay the foundation for elegant and sustainable methods. Of course, we still have to talk to each other, but on an objective and completely system-supported basis.

Since this topic is too interesting to end here, in my next article I will present a concrete example from the fulfillment area, showing how process mining is used operationally.

 

Check out the german webinar recording on this topic: