What is causing your tree of inefficiencies to grow! A bit cheesy I know but its something a lot of people really struggle to answer.
In today's post we wanted to show how easy (and quick) it is to run some machine learning in Process mining to see the Root Cause Analysis of behaviors in your data. This is always a great starting point to find out what is causing records to achieve a desirable or undesirable overall process.
This was inspired by a couple of different customers who were Process Mining and when the map loaded they were really excited to get going but there was a small speed bump. They knew there had to be inefficiencies because of certain metrics they were tracking but they had no idea what could be causing them so they didn't know how to proceed. We suggested root cause analysis as a solution here and that really got the ball rolling.
Let's get to the root of it!
Here is a quick 7 minute video on how to achieve this with many examples or follow along below:
Using Process Mining for Root Cause Analysis
You can find other Process Mining use cases here
How to do the analysis
This approach can be applied to any workflow (Incident, HR Case, Customer Service Case, etc) and does not require any additional/special Process Optimization configuration beyond the out of the box content packs offered with a given workflow
Open your project to the Analyst Workbench view:
From here we will run root cause analysis on any state that we find desirable OR undesirable! Remember you can use RCA to find why good things are happening in your process so that those can be standardized as well.
For this example I want to understand what are the leading contributors for an incident moving into the "Awaiting Caller" state. I will click on that node -> hit the investigate dropdown -> click on "Key Contributors (RCA)"
THATS IT - 3 steps and you will have your RCA running. When you hit that last button it will take a couple minutes to run before giving you the results of why things are moving to this state with a higher likelihood.
Note: at least 100 records are needed for RCA to run as it is using some ML models in the background.
Now we can look at the results:
It looks like we have some "Services" that we need to take a look at in our process, great! In this case, things that come in with the SharePoint service are 22.86 times MORE likely to be moved into Awaiting Caller than any other Service (80% of all incidents with the SharePoint Service in this case move to this state). Maybe there's something we should take a look at in the intake process for these incidents 🤔
But we can run RCA on even more than just states, we can run RCA on BEHAVIORS. In this next example I want to see what is causing Incidents to move from Resolved back to In Progress. Well we would just run RCA in the exact same way but on the transition. I will click on that arrow (transition) -> hit the investigate dropdown -> click on "Key Contributors (RCA)"
The final example I wanted to share is that this can be used for even more application. What are the leading contributors for a record in the parent table (incident) affecting a record in ANOTHER table. Maybe we can see what are some leading causes of breached SLAs 🤓
We'll first need to have a mine with the Task SLA table associated to it, after that it is exactly the same. Choose a transition that is leading to the most breached SLAs -> investigate -> mine.
And just like that with another 3 quick steps I was able to find some leading contributors for breached SLAs. Our demo data may not always make sense but the process is the same every time!
Thanks for reading and as always,
If you are looking for more in-depth training you can use the Process Mining Academy library of content\
You can find other Process Mining use cases here
BelalA
