Stages of implementation and configuration of Process Mining tools

[Mimaktsa10]

For more productive use of Process Mining, we recommend following these steps:

Step 1: Selecting the optimal solution
There are about fifteen business process analysis systems on the market today, mostly from American and European developers. When choosing, two key aspects must be taken into account:

Recently, there have been cases when foreign suppliers of various software stopped functioning of cloud services in Russia, stopped technical support of their products. Therefore, it is recommended to pay attention to domestic platforms, among which there are quite mature and reliable solutions.

For large corporations, the ability to jordan email list deploy a system on their own servers (which is not always possible), the availability of integrations with existing enterprise information systems, flexibility in customizing and adapting the solution to the specific tasks of the company are often critically important.

Step 2: Process Identification
It is recommended to start with the process that has the highest degree of digitalization: only in this case will the system be able to conduct its full analysis. But ideal conditions for such a study are rare.

Step 2 Process Identification

Source: shutterstock.com

A fairly typical situation is when a process is only partially digitized or is conducted in several unrelated information systems. In this case, it is necessary to analyze it step by step and gradually transfer the "analog" components to digital format. This will allow, over time, to cover the entire process and obtain a more complete picture for optimization.

Step 3: Data accumulation
The analytical base may consist of the following sources of information:

corporate platforms (such as SAP, 1C, Salesforce);

relational databases (PostgreSQL, MySQL and others);

various file depositories;

event logs generated from electronic correspondence or spreadsheet documents during the work of personnel.

There is no need to concentrate all data in a single repository. If necessary, it can be exported using special connectors, spreadsheets, or by directly connecting to databases.

Naturally, the question of the optimal volume of material arises. The principle "the more, the better" works here, but usually a sample for a year and a half of operation is enough. For mass and cyclic processes, even a month or a week of information may be enough.

Step 4: Assessing the quality of information
For effective classification, data must have certain characteristics. The most important of these are:

unique identifier of the process instance (this could be the number of a purchase, contract, application or request to the technical support service);

stage name (may include status, department, or a combination of department and status - essentially, an identifier for a specific action);

chronological marker.

Additional attributes may include performers, branches, geographic regions, various types and attributes, and in some cases even financial amounts. It is important to understand that the more characteristics you can collect, the wider the range of hypotheses for potential process improvement.

Stage 5: Project Initiation
In an optimal scenario, by this point all process segments should be digitized and have an end-to-end instance identifier.

The next step is to set up direct access to the event log data of the selected process through integration with source systems. After this, a full-fledged opportunity for its deep analysis, identification of potential optimization points and daily monitoring of the effectiveness of the implemented changes opens up.