Enhancing Process Management with Process Mining

Process Mining has come a long way since its early days of inception by Professor Wil van der Aalst. There are growing numbers of vendors rolling out process mining platforms and its adoption is growing steadily, transitioning from early adopters to becoming one of the latest trendy tools in the industries. It has come out of age and reaching maturity in term of the methods, specialised software functions and applications, and soon having its first Gartner Magic Quadrant Report in 2021.

Being the link between data science and process science - the broader discipline that combines knowledge from information technology and knowledge from management science to improve and run operational processes (Prof. Wil van der Aalst), process mining could also play a critical role in enhancing the business process management (BPM) practice itself. This article aims to explore the opportunities of applying process mining to improve the different aspects of BPM and DMAIC activities from process redesign to process change implementation to the day-to-day management of operational processes.

Overview of BPM Stages

Figure 1 above is an illustration of the three major stages in a business process oriented management, and alignment with DMAIC. Starting from the far right, in the day-to-day operations stage, the ideal scenario would be where business processes are running around the clock, with control measures in place guiding the process executions and also the on-going monitoring of the in-flight process instances. Also keep in mind that all these activities are running concurrently, not one after the other. On the far left, the business process redesign stage, where common BPM / DMAIC activities take place in order to redesign and improve the business processes. These activities include process discovery, process modelling, process design, process analysis and also process simulation in organisations with higher process management maturity.

And the transition stage of process change implementation where it involves both the technology implementation, including process based application, RPA or other type of enterprise information systems; and staff readiness preparation - the change of jobs and roles to the very fine work instructions and mentality of adapting the change.

Opportunities of Process Mining Applications

With the maturing of Process Mining methods and techniques, it opens up many opportunities to adopt them across the BPM / DMAIC activities (see Figure 2), not only improving the effectiveness and efficiency of these activities. but taking them to the next level - semi automation of these activities.

Process Mining in Business Process Redesign Stage

Process Discovery / Modelling

Assuming that there is no existing and up-to-date process architecture, business process redesign starts with process discovery, the attempt to understand the end-to-end as-is business process in scope. This often requires huge investment of time and efforts in process modelling. The process discovery and modelling are not only costly and in most cases time consuming and inaccurate due to disagreeing views and interpretations from the many roles and players across the end-to-end business process.

Applying process mining in process discovery using information systems logs not only reduce the time and effort of process discovery on the as-is process, it also provides a full picture of the variations of business process. Hence resulting in a more precise and agreed representation of the as-is business process models. Most of the process mining tools also support the auto-generation of BPMN process models from the process maps, reducing the efforts and time significantly in the process modelling.

Based on experience, the time saving could be up to 80% comparing to a manual process discovery with iterative workshops. The advancement in process mining methods also allow extraction of business rules applied during the execution of the process. In short the possibilities to elicit operational rules and decision models, which may be implicit to the subject matter experts.

Process Analysis

The application of process mining during process analysis is a no brainer with the multitude of possibilities of slice and dice on the data. These techniques and methods offers a huge variety of data-driven and objective insights on process conformance, and process performance. In process conformance analysis, the checking of compliance of user behaviours to prescribed process, the checks are transparent to the users which avoid the potential of users behaving differently such as when under a manual check via observation.

In process performance analysis, advanced techniques enable issues identification, and improved impacts and root cause analysis, using combination of both process data (event logs) and business data (attributes specific to transactions). These insights could lead to opportunities for further improvement, optimisation and/or enhancement for process redesign.

Process Simulation & Playback

Events playback is another technique of process mining that could be applied hand-in-hand with process simulation during the redesigning stage to understand the potential impacts of the redesigned to-be process models and comparing the results based on the same given business transactions captured in the event log. This could further be extended into Activity Based Costing (ABC) on both the as-is model and the to-be model, providing financial perspective of the as-is and to-be process models.

Process Mining in Process Change Implementation Stage

In the process implementation stage, process mining on combined event logs from multiple information systems could help in targeted process automation and task automation.

Where the event logs have included multiple information systems' activities, it is possible to identify not only user tasks but also service tasks and interaction between systems or manual swivel-chair tasks.

As an example, the identified swivel-chair tasks can be targeted for RPA bots.

The next level of task mining could be used to discover steps and procedures within tasks and build bots in RPA systems, which will in turn further reducing the cost of the change implementation.

On the user front, process mining could be used as an impact assessment to identify impacted business units, groups, roles or even to the very granular level of individual users. This will empower the change team to have a better grip on the change process and increase staff readiness for the roll out of the redesigned process.

In addition process mining can be applied as a post implementation tool to assess the benefit realisation of the process change and also to identify potential process shift in post implementation which is very common as business users tends to slip back into the old way of doing things over time.

Process Mining in Day-to-Day Operations Stage

The operational stage is often the forgotten child in BPM or DMAIC practices which are mainly focusing on the process redesign stage. Although there are many business intelligence and BAM tools, most are not process oriented and definitely lack of insights on process conformance and process performance. In short, there is little oversight on the operational process health. Where process health are monitored, the works of generating these reports manually are time consuming at best, and therefore lapsing, missing data at worst.

Where process mining tool is integrated with the information systems or data lake, it is possible to aid in the process control providing an on-going insights on process conformance and identifying of any non conformance behaviour. With the process monitoring, process mining can also provides on-going process performance insights to allow continuous insights on the process health with minimal effort.

Finally the advancement in predictive process analytics will enable an early warning system - predictive process health - to allow management intervention on potential process instances that may breach process performance metrics or the conformance requirements.

Similar to your personal health care, early detection / alerts with preventive measures will make significant differences to the health and well being of your operational processes.

Moving to Data Driven BPM

With the emerging and maturing of process mining techniques and tools, it is inevitable that the data driven methods of process mining will lead BPM practice into its next page into a data driven frontier. Similarly the adoption of process mining in DMAIC will finally provide a more holistic view in the process redesign and extend the control and monitoring into the on-going operational processes.

If process mining capability is not on your organisation radar now, it's a time to start your journey. Not taking the initiative probably mean your organisation will be left far behind or even outside of the new data driven process management playing field.