By: Issa Kerremans
As part of your process mining exercise, data has been pulled from multiple sources, it’s been cleaned-up, organized and saved in a format that is usable by the process analyst team and their stakeholders. So, now you’ve got all this data, what do you do with it? How do you gain value from your mined data?
Start by analyzing the process data, then select opportunities for improvement, create ‘what if’ scenarios and run simulations. Read on to find out more…
Analyze the process and process data
Start by looking at the data to see what it tells you about your processes. Things to look out for include:
- Exceptions – process flows that break away from or follow a different path to the defined standard. For example, scenarios where steps are skipped or an alternative flow is being used – this becomes even more important for processes that include regulatory requirements.
- Bottlenecks – are there particular stages in the process that become blocked and slow down the overall process due to a lack of capacity, broken sub-processes, or a lack of relevant skills and knowledge?
- Resource utilization – be that people or systems, you want to see how your important resources are being used and if they are being used efficiently. For example, are seemingly simple tasks taking prolonged periods of time to complete or are you only using a small percentage of an expensive systems capacity.
- High costs – are whole or parts of a process costing significant sums relative to the value they deliver e.g. with an average order value of £10 and 30% margin is it costing £4 to process the order.
- Death spirals – process scenarios that loop round and round without ever coming to a conclusion. For example, a process that unintentionally bats a customer back and forth between two departments.
- Wait times – prolonged waiting because the step in the process is unnecessarily long or there’s a queue generated by a lack of resources further down the line.
- Non-compliance – are there processes that breach regulatory or internal requirements, such as response times to customers or KPI thresholds?
Tip: process mining can only capture process data from systems, if there are gaps use manual process capture techniques to fill the gaps
Looking at a process with the naked eye will discover all of the above scenarios and more. This discovery and analysis exercise should be supported by heat maps that highlight where issues may exist and running simulations that bring the process to life through showing the process flows and hotspots that require particular attention.
Select opportunities to improve
With a visualization of the process and the areas that require attention highlighted, it’s time to create a view of the opportunities for improvement (see opportunity dashboard).
From here prioritize the opportunities that will deliver the highest return. Sometimes the opportunity to improve will be glaringly obvious and sometimes you will need to dig a bit deeper to work out where to start. For example, on the surface multiple points in the process may seem like they have the same level of issue, but by looking closer you may identify that some of the issues are impacting customers while other, although problematic, are having no material impact on the over process flow.
Continuously evolving, the opportunities dashboard should be the place from which initiatives are derived and driven.
Tip: when starting out, start small and grow.
Model ‘what if’ scenarios
Once you’ve identified the areas for potential improvement, it’s time to start modelling different scenarios to see which will have the biggest impact. Using ‘what if’ statements you can start to create and simulate different process configurations to see which has the most potential. For example, ‘what if we added more resources’, ‘what if we automated parts or whole processes’, ‘what if we moved resources from one part to another part of the process’, ‘what if we removed a part of the process’. This list could go on and on to cover any one of a multitude of scenarios that are thought of by you and your team or based on industry best practices (see example below).
Industry Best Practice Comparison
Important in this, is to run simulations, as opposed to working directly on the live process. Not least because making experimental changes to live processes may have a wider impact on the organization.
Instead by running simulations that include the changes you could make to the process, you should be able to run more simulations at a faster pace. Take a 5-step process that has been identified to have a bottleneck at step 2 as example. You could simulate a scenario that sees you fix the bottleneck, maybe by adding additional resources to the step. On producing this, you may see that fixing step 2 only moves the issue to step 3 and 4. However, by remodeling and sharing the additional resources across steps 2 and 3, you could see that overall performance improves.
This potentially oversimplifies the problems to solve, however by using simulations, more complex scenarios can be quickly generated and tested without impacting the live process until a suitable alternative flow is found. Compare the situations, Team A identifies a potential process improvement and starts work on making it live, this takes 6 weeks to deliver and once live shows a negative impact. While Team B runs multiple simulations and is able to not only find an initial improvement, but hone that to deliver more improvements to the process.
Create your Digital Twin
Go a step further and you see that you are starting to create a Digital Twin of your processes. Namely, a version that is live and being used every day and a version that replicates this and becomes the space for creating and testing potential improvements. What’s more, in this low-risk environment, you are able to think big and test dramatic changes knowing that you will only implement them if the results are positive.
Most organizations will start this exercise within a specific business unit or by looking at a specific process. But it does not need to stop here, it is possible to run simulations across functions and processes so you can start to see how changes to one process impact other processes within the same area or across the organization.
Coming right back to the start of this article, process mining is a really valuable exercise that allows you to capture a detailed view of your processes. However, without sufficient plans and activities to analyze, understand and create alternative views, the data becomes interesting but not actionable. Looking beyond process mining and using analysis, scenario building and simulations, you will be able to identify and implement process improvements in a low-risk environment that is faster to value.
Find out more today!