Process mining refers to the act of "excavating", analysing and subsequently evaluating data from the log files of business processes. This enables companies to map entire business processes as well as the associated procedures, patterns and trends. In addition to the process analysis and inventory, including all sub-steps and variations, process mining is primarily aimed at optimising processes in order to achieve cost savings.
• Processes in ERP systems, e.g. orders
• Ticket processing in help desk systems
• Workflows in business process management/automation systems
Due to a number of similarities, process mining is often mentioned in the same breath as data mining, but there are also important differences: Both methods involve the analysis of large amounts of data with a view to gaining insights and making data-driven decisions.
Data mining focuses on analysing large amounts of data in order to uncover hidden relationships. Process mining, on the other hand, combines data analytics with the modelling, control and improvement of business processes. To this end, it uses algorithms specifically adapted to log files and software system databases in order to gain new insights into processes and their variations. In this regard, three important sub-fields of process mining can be distinguished:
Discovery involves the analysis of existing process log files in order to reconstruct sequences without any detailed knowledge about the underlying processes. As such, it automatically identifies processes and their sub-steps so that they can be analysed, without the need for pre-defined process models.
If a model of the process to be already exists, process mining can be used to compare the actual situation with the target situation by means of data analysis of the process as is. This makes it possible to determine if the situation conforms to the predefined "optimal process" and to identify any deviations and their effects.
Apart from checking for conformance, the most important part of process mining is process optimisation. After the actual state has been evaluated, the process (model) is adjusted in a second step in order to achieve a new, better process.
The main prerequisite for the application of process mining is that the data on the sequence of individual steps in a business process are stored digitally. The stored data must contain information about the time of each event and its relationship to the process or process step as well as to the data set in question.
Usually, source systems create some kind of process log or log file based on which the chronological sequence of a process, including all its sub-steps, can be reconstructed. To this end, the entire process must be digitally mapped from start to finish. If parts of a process are executed manually or analogously, process mining cannot extract meaningful data about the entire process. In general, the more data are available, the more precise the evaluation and the greater the potential for optimisation.
Most of the data for process analysis come from the core systems that companies already use today, such as ERP systems, ticket systems, business process management (BPM) systems, workflows, CRM systems, and potentially any other software that uses process logging.
After process digitalisation, the next important step in the transformation towards a completely digital company is the optimisation of its newly digital and therefore evaluable processes. Various studies have shown that an average of 20-30% of productivity is lost due to poor or ineffective processes. Process mining offers the opportunity to analyse existing workflows and gain insights into the actual functioning of a company in order identify any problems and optimise its processes. As a rule, this should not only be done once, but take the form of a consistent and continuous optimisation process.
The possible weaknesses, areas for improvement and interesting process aspects include the following:
• Lead times
• The duration of individual process steps
• The reasons for long lead times
• The detection of bottle necks
• Deviations from the process “to be”
Process mining enables companies to uncover their processes as they actually operate as well as any process variations. Through process discovery, they can record and update their process models, thereby greatly reducing the time and effort required for manual process visualisation or modelling. Once the target process to be has been defined, it can be compared with the as-is state at any time. Companies can thus detect deviations, identify weaknesses and inefficiencies and use the resulting findings to optimise their processes. Another important aspect that process mining can help to ensure is compliance.
Thanks to core software such as ERP, CRM or workflow systems, companies already have access to the necessary process data. In many cases, the "ideal" process has been modelled and is already being put into practice. But is this really the case? Is the target process that has been specified really ideal? Process mining enables companies to evaluate the available data and to analyse their actual processes with a view to determining future action. Ultimately, this makes it possible to achieve productivity and efficiency gains and thereby to reduce costs. To sum up, process mining can reveal relationships, insights and opportunities that would otherwise be inaccessible.
Process mining can help any company that maps its processes digitally, in whatever form. Process mining helps companies to uncover their actual processes and provides important insights for successful process optimisation. Sometimes, this reveals problems that were previously not known at all.
Often there is a gut feeling that processes are not working well and take too long in some places. Process mining creates facts and evaluation options that make it possible to verify if this is really the case. In the context of the further digitalisation of processes, process mining software can provide better quality, transparency and insights into a company's processes and save costs through optimisation.
Process mining cannot be used for manual or digitalised processes that are still partly executed analogously (e.g. where documents still have to be printed out and signed in accordance with signature rules). In such cases, a lot of the relevant data for analysis and evaluation are lacking, making it impossible to obtain a meaningful picture of the actual process.
In other words, process mining requires that the entire process be digital. Similarly, it is not possible to magically obtain perfect results from bad data. For process mining to be effective, the data need to be appropriately prepared and then made available to business users and executives for meaningful analysis. Standard process mining software is available for this purpose, with ready-made templates, data connections and KPIs for many types of source systems and the most important standard processes.
While the digital workplace and the digitalisation and automation of business processes are the first step, process mining is the logical next step in order to gain the necessary transparency for improving, optimising or fundamentally transforming a company's processes. In a company's quest for cost savings and ways to optimise its efficiency and strengthen its market position, process mining is the ideal tool – and one that will only become more important as digitalisation advances.