Competent management of business processes is one of the most urgent tasks of any business. Even small improvements in a large company can significantly increase profits or reduce costs. That is why Process Mining is becoming increasingly popular in many companies.

Process Mining is a group of methods that allow for the in-depth analysis of business processes based on event logs. The creator of the concept of Process Mining is Wil van der Aalst – a professor at the Eindhoven University of Technology (Holland) and the Queensland University of Technology (Australia). He is the author of the fundamental work “Process Mining. Discovery, Conformance and Enhancement of Business Processes”, which defined the key capabilities, areas of application and prospects of business process analysis.

Process Mining is used to evaluate multi-step processes with a complex decision-making hierarchy, with a large number of typical, repetitive operations that are logged by an information system. Only the logs of mass operations are suitable for Process Mining – it makes no sense to try to analyze single processes in this way.

Technology’s main advantage is that you can get interesting and useful results having only three attributes for each event: process identifier, action name and timestamp. Using a minimum of input data, the user can perform non-trivial analysis and get quick results. Modern information systems, including RPA, generate similar data, so Process Mining can be applied in all industries that digitize their processes. Of course, the more analytical attributes are captured, the more in-depth the analytics will be.

Some information systems allow logs with more analytical attributes to be uploaded in a form suitable for Process Mining. But more often, preliminary data preparation and consolidation, enrichment and calculation of metrics are required. Below is an example of preparing the initial data for Process Mining. As you can understand from the scenario, the stage of data preparation for in-depth analysis of processes can be quite time-consuming.

Scenario of data preparation for Process Mining

In practice, process mining is actively used in manufacturing and hi-tech, energy and telecommunications, trade and pharmaceuticals, insurance and fintech, aerospace and defense. Process Mining is used by such market giants as Siemens, Dell, Royal Dutch Shell, Unilever, Adidas AG, LʼOreal Paris, The Coca-Cola Company, AstraZeneca, Citigroup Inc. The most common applications: customer service, credit pipeline, call centers, logistics, technical support, shipping, maintenance and repair.

The value of Process Mining is that it allows you to restore the actual, real model of a mass business process rather than an “expert-ideal”, regulated one that ignores many variants of events.

Process mining is diverse and depends on the specific application. However, there are standard stages, typical for the majority of cases, regardless of the industry:

Discovery – automatic process discovery – extracting process models from the event log. Calculation of process metrics, exploratory analysis, finding the “bottle neck,” etc.
Conformance checking – finding deviations from the benchmark processes.
Automated construction of simulation models – modeling and continuous improvement.
Monitoring – automated process control. Informing about deviations in the sequence of process steps or process metrics.

Process Mining solves the following business intelligence application tasks:

Real-time intelligent process analysis.
Behavioral analysis.
Benchmarking – comparison with benchmark processes.
If-else analysis – what-if analysis.
Calculation of the cost of each stage of the business process.
Calculation of time and financial losses.
Analysis of compliance with requirements and regulations of the process.
Identification of “bottlenecks” of the process.
Detection of redundant links in the process.
Detection of fraudulent transactions.
Detection of looping.
Simulation and stress testing.
Search for process anomalies.
Detection of atypical behavior.
Evaluating the extent to which each factor influences the process.