When I studied my Master’s eight years ago, academic research was mainly focused on the technical aspects of process mining with little (direct) practical relevance. Academic researchers ardently pursued topics such as theory of regions &Petri net algorithms, conformance checking topics of alignments with a*-algorithm and model fitness, reachability and soundness, and others. Of course, implemented as new ProM plug-ins.
As more and more organizations accelerate their digital transformation initiatives and adopt process mining to analyze and monitor their business processes, academic research has started to address the business application side of process mining as well.
In her bachelor thesis, Inge van Dijk from Utrecht University examined the factors contributing to a successful implementation of process mining within an organization. Her findings are based on an extensive literature review and supplemented by interviews with practitioners. I was invited to be one of these practitioners and was glad to share my personal views and provide input for this research. Below I list the process mining success factors as identified in the thesis and add my own interpretation.
As the famous saying goes “garbage in – garbage out”. Process mining, as any data analysis technique, stands or falls on the quality of the underlying data. A process mining analysis that is based on low-quality unrepresentative data has little or no value for the organization and more often than not leads to false conclusions.
Since process mining is still a rather innovative topic for most companies, take time to educate your stakeholders and subject matter experts on process mining techniques, what it can do and –even more importantly— what it cannot do. Demonstrate the possibilities of dedicated process mining tools and explain the benefits.
I have written on the importance of top management support and endorsement in my previous article. This research confirms that top management support is critical, as often general staff does not have the resources or the authority to make impactful decisions.
Process mining analysis can result in a long list of insights and, consequently, many potential improvements. Restrain yourself from working on an unmanageable amount of improvements at any given time. Rather, focus on a small selection of the most promising areas, then iterate.
Dedication is understood as the willingness of senior management to allocate resources, both human resources as financial resources, to realize process improvements.
When it comes to understanding process behavior, context is key. If you leave the context out while interpreting the data, findings can be misleading. Data typically does not take into account local legislation or customer-specific requirements. That is why it is crucial to have a process expert be part of the core team in order to interpret data in the context of business process.
The purpose of process mining is not to reveal negative performance and find someone to blame. According to the research, a frequent mindset in companies that work with process mining is that disappointing performance indicators are a bad thing. This results in managers who do not want to show their process mining results, because they fear being reprimanded.
In contrast, process mining should be seen as an aid for continuous improvement and as a tool for spotting problems that could not be identified otherwise.
Based on research of Inge and my personal experience, several conditions have to be in place to make process mining successful within an organization.
Ekaterina Sabelnikova is Innovation Consultant at Philips Innovation Services