Their data is often most closely related to the executed process, and contains the most interesting information. Look out especially for CRM, ERP, workflow and ticketing systems. You can talk to someone who works in the process and ask them which applications they use to perform their tasks. The next step to determine which IT systems are involved in the execution of the process. Process mining is an iterative activity and you can extract more data, or define better questions, the next time you go through the cycle. You can start with your process mining understanding from the Hands-on Tutorial. performance) can influence the way in which cases should be chosen.ĭefining questions before you start with your first process mining analysis can be a bit of a chicken-and-egg problem: You don’t know yet what kind of questions you can answer with process mining. Furthermore, as discussed in How Much Data Do You Need?, for some data selection methods the analysis goals (compliance vs. For example, if you want to benchmark the processes in different countries or regions then you need to include the country information as a attribute.

The questions you want to answer can pose additional requirements on your data extraction.

However, a few main questions as a starting point is critical not to lose sight of your analysis goals. You can add these questions to your analysis in the process. It’s one of the advantages of process mining that it allows for explorative analysis: You can get answers to questions that you did not even know you had just by looking at what is really happening. Once you start analyzing the data, you will get many more ideas. Otherwise it is very easy to get lost in the data.

You never know in advance which delays might occur in the IT department and which data quality problems the data will contain. If you are under time pressure to show some results from your process mining analysis within a certain time period, consider initiating the data extraction procedure for two different processes in parallel. If you still have a tie between multiple processes, go for the one with the highest volume (for example, based on the number of cases that are processed). You can either assess the improvement potential with a classical business case analysis or base the decision on the gut feeling and anecdotal evidence from the process experts.

Unless you just want to play around for learning purposes, there must be an interest from the business in the results of your process mining analysis. With data availability and champion support being equal, pick the process that is most relevant and has the most improvement potential. You also need the support of the process manager in defining the main analysis goals for the process mining project. Process mining does not happen in a vacuum and an unavailable project sponsor who wants a “surprise me” analysis is a red flag. Without the support of a good champion, you will hit a wall of questions very quickly.

You will need at least one strong supporter who knows the process well (domain expert) and can answer questions during your analysis. Equally important to the data availability is a good support from the team that is responsible for the process. In contrast, a process for which you already have access to some data via the data warehouse may be better than a process that is outsourced and requires the collaboration of an external service provider. Can you obtain good process mining data for this process relatively easily? For example, a process that is handled in an ad-hoc manner through email may be possible to analyze to some extent, but it will require a lot of pre-processing.

Web data extractor v7.1