When trying to optimize business processes, Conformance Checking forms the base of any optimization measures. To put it simply, conformance checking is to scour the process for critical or previously unnoticed deviations from the norm. In the past, companies had to spent countless hours manually gathering process statistics and comparing them to the ideal reference process. However, with Process Mining tools such as LANA, checking process conformance becomes as easy as checking your email.
The project might seem daunting at first. Loading your data into the tool, visualizing the actual process, generating a target model and then comparing models and identifying potential for improvement – all of this can be overwhelming if you have never used a Process Mining tool before. Luckily, in LANA we can boil this down to six (and a half) easy steps which will get you optimizing in no time.
1 – Examine your discovered model
The very first thing a Process Mining tool is going to show you after you upload your data is the discovered model of your process. This is the base of your conformance check.
As a visualization of your actual processes, it serves as a window into every detail of the process in question. Including potentially problematic deviations, loops and skipped activities.
2 – Create a reference model
To identify all relevant process deviations, LANA compares the discovered model with a target or reference model. This is an ideal version of the process in the form of a BPMN model. (BPMN refers to Business Process Model and Notation, a standard for process modeling.)
There are 2 ways to create a reference model in LANA Process Mining:
1) Upload an existing model as a .bpmn file.
2) Create a new model in LANA’s integrated BPMN editor.
2.5 – Create a BPMN model
In case you do not have a ready-made reference model that you can upload to the Process Mining tool, you can simply build your own. LANA Process Mining includes an integrated BPMN editor. Here you can easily create a target model from scratch with a simple drag-and-drop system.
First, take the discovered model as a base and choose which process variants to include in your reference model. Use the variants slider in the discovered model to identify the most common variants.
Then, build a new model with LANA’s BPMN editor. Your goal should be to accurately model the process variant you want to use as a reference.
3 – Connect the models and start comparing
Whether through importing a .bpmn file or building your own, it is now time to connect the reference model to the actual discovered model. Several icons will indicate potential conformance risks.
The red “plus” icon symbolizes repeated activities that are executed at least twice. Clicking the icon shows statistical key figures about that activity.
The bue “arrow” icon marks skipped activities that are not being executed at all. Once again, you can see the relevant statistics by clicking the icon.
Use the variants slider on the right to compare different process variants with your ideal target process. Pulling the slider further right shows more variants simultaneously, while dragging the slider to the left focusses on fewer, major variants. This way, you can easily prioritize how deep your conformance checking will go.
4 – Compare variants side by side
The Variants View shows you more details about specific process variants. Use the sidebar navigation to reach this view. It highlights the identified process deviations side by side and separated along the different individual variants. Once again, blue coloring shows skipped and red inserted activities.
5 – Check the statistics
Another helpful tool is the Statistics View. Use this dashboard to check key figures and get an overview about your conformance statistics, among other more detailed insights into your analysis. The higher the percentage in the top right, the better your overall process conformance.
6 – Drill down to the root causes
Of course, any kind of conformance checking can only be the start for actual process improvement. To turn the statistical insights into actual actionable elements, you need to find the root causes for process deviations.
Luckily, all identified problems and deviations are also listed in the Action View. From here, you can start a Root Cause Analysis by clicking the “lightbulb” icon on the right. LANA’s internal machine learning algorithm will identify the possible root causes for any individual issue. Use this as an intelligent basis for your optimization strategy.
Time to put this into practice!
Reading is only half as fun. Use this opportunity and try your hands on your own conformance checking project. To get started with LANA Process Mining, all you need to do is contact us and request a trial license. To sign up for a trial license, including access to two sets of example data, go to Contact and simply fill out the contact form.
So you always have this tutorial ready, you can download this conformance checking guide as a PDF document by clicking here.