As part of the digital transformation, many companies are currently preparing to optimize their business processes. Process mining is often used to identify the difference between the ideal state of a process and its actual execution. This article explains how to successfully anchor the analysis methodology in the company. Process mining enables companies to use transactional data to scrutinize individual processes, identify bottlenecks and compliance violations, and identify optimization potential. Such a check makes sense especially in older, established processes, since there the desire and reality are often far apart.
Ideally, processes interact seamlessly to support overarching corporate goals. In practice, however, process owners often wondered why their processes were not going as they should or how they were improving processes whole mining, and process mining offers the opportunity to answer these questions. There is no general recipe for success for the implementation of this analysis methodology in the company, but there are now some universal criteria and best practice approaches that characterize successful process mining initiatives.
|Table of Contents|
Determination of the project scope
A successful process mining initiative depends on well-thought-out planning. Only on this basis can a company implement all further project steps. Workshops with the most important project participants are usually suitable for this initial phase, in which the scope of the project is defined.
The process details must then be determined based on the project goal. This involves defining all the individual steps between the beginning and the end of a process that is to be analyzed. The data requirements must also be determined. This means that the relevant business documents must be determined concerning the defined process details.
Preparation of data
Once the company has defined the scope of the project, it can start with the technical preparations. Step two essentially involves the extraction, conversion and transfer of data to the process mining software. Two methods for data integration have proven themselves useful: the connection via a software connector or the use of ETL tools to extract, convert and load the data into the process mining application.
Before extraction, the relevant data must of course first be identified, whereby the necessary data is obtained from the processes identified in step 1. However, very few IT systems are based on processes, but rather on business documents: some data sources contain sales orders and other invoices, for example, and therefore need to be identified more closely; typically, these are data-based tables of transactional systems such as ERP or CRM, analytical data such as reports, log files and CSV files.
After extraction, the data is translated into a chain of different events and converted into so-called cases, i.e. into a sequence of different steps in the process execution. The information on these cases is stored in the event logs, which the process mining software accesses. All steps from extraction to data conversion are controlled via the connectors or ETL. This process is triggered regularly so that a company can access up-to-date process information at any time if required.
Targeted evaluation of the processed data
Once the non-technical and technical preparations have been completed, the actual process mining, the analysis of the data, can begin. We recommend starting comparatively high up in the process flow to then gradually analyze the various components of the process. When process experts compare the different process information, they can see how it affects the company. According to different BPM providers, it may well be necessary for a process expert to question managers in various departments to interpret certain information. This step makes it easier to understand the flow, the metrics, the bottlenecks and the optimization potential of a process.
Measurement of results
In the fourth and final step, possible process improvements are then evaluated, tested and documented. The planned changes are then discussed in the team and then implemented. At the same time, the team continuously measures and monitors the key performance indicators of the processes to identify bottlenecks and undesirable process behaviour. In principle, it also makes sense to extract new data a few weeks or months after the initial process improvements. On this basis, companies could recognize what has changed and which measures have led to greater efficiency. The next step was to identify further optimization potential and initiate subsequent improvements in the process life cycle.
However, the project is still not finished even if the steps are finished – it will probably never be. Rather, process mining should be understood as an iterative method that takes companies step by step on the way to success.