Home / News / Investigate your Customer Service Quality with Process Mining
26 October 2020
By: Natasja Franzini

Investigate your Customer Service Quality with Process Mining

Process Mining for Power BI

The Covid-19 crisis is creating challenges for businesses worldwide, but can also create opportunities for others. Not only is there a big increase in the digital transformation journeys at many companies, but there’s also a great impact on customer service departments due to the work-from-home policies.

Now seems like good time to share a Dashboard we created to track our own customers service quality. In this blog I’ll show you how easy you can investigate your Customer Service Quality with PAFnow Process Mining for Power BI.


Process Mining with PAFnow

A powerful way to improve performance and enhance process quality is through Process Mining. PAFnow is a tool for Process Mining and offers powerful ways to visualize and analyze your business processes based on event  data, meaning any data that you want to measure about an event. It is the leading Process Mining tool implemented in Power BI, recognized by Gartner. With the free visual that is available in AppSource everyone can get started. It is easy to use, powerful and innovative.


Let’s start

First ask yourself: what am I trying to achieve? In this case we want to map out what it takes to have a happy customer and how we can improve this. To make any process analysis possible, make sure you have at least three components:

  1. An activity name. The activity name should describe what has happened.
  2. A case identity. The case identity should be unique, so you can properly identify each separate case.
  3. A timestamp. The timestamp denotes when something has happened.

The classic data format for process mining is an event log. Event logs are often rich and hidden sources of information in your company IT system. Not all tables are tracked automatically, so make sure your event log is turned on. For Dynamics 365 CRM it’s not possible to use the regular data export functionality, so at Dynamic People we developed a special tool to get access to the data.

When the Visual is loaded I can start populating my fields. As described above, we need to supply a Case Identifier, an Event Label (Activity Name) and a Timestamp.

The visual is gives me output straight away. This can be a bit overwhelming at first, because it shows the complete process and all its paths. 

Using the zoom button on the bottom of the page you can start exploring this area in more detail.


With the slider you can decrease or increase the complexity of the visual. In other words, you can choose the number of unique paths to show. These unique paths are called variants. Slide all the way to the left to see which path is occurring most often. This is the variant with the highest volume of cases and is called the “happy path”.


The nodes give you the activity name and the volume of cases that go through this activity. The color of the node is based on the amount of cases going through it: the higher the volume, the redder.


The lines between the nodes are called Edges. Some of them are thicker than others, which is based on the amount of cases following this edge. Additionally, the redness of the edge is based on the average duration of the edge. We can use it se the average amount of time spent between activity A and B.

Edge Labels

In the Edge labels you see the amount of cases going from one activity to the other, including the average time it takes between these two activities. Giving you an immediate clou on where to focus on and what steps in the process to investigate.


Legend Group By

The picture above shows that we often follow create a case and send a notification very quickly. When we look further down in the process, you’ll notice it takes us on average 15 days to resolve a case once it has been assigned and a follow-up date is secured. We can already determine where to start digging into our customer service process to make our customers happier.

Let’s  look at two more things in the (free) visual. The first one is a legend ‘Group by’ option to see more variants. When you group by variants it looks like the visual underneath. By checking variants, you get a better understanding of what different paths exist next to one that is most taken.

Layout options

Then there’s also the multi graph layout option ‘Side by side’. I changed the group by option to ‘Happy Path’ and choose to see the different lanes ‘Side by side’. I also removed the mini map on the bottom. This creates a very readable graph where the visual gives you a suggested ‘Happy path’, ‘Happy Start’, ‘Happy End’ and or ‘Not Happy path’. What is interesting to see is that the happy path has some overlap with the process we follow already, but it also includes some steps we do not have in our first chart. All these things together give you a great head start to improve your customer service or any other process.

A deeper dive into Process Mining?

Even though the free visual gives you a great place to start and is enough to get you excited about process mining there are some great advantage to having any of the licensed options. Like I mentioned before I would like to be able to select the different variants. The licensed version gives you this option. Furthermore, with the premium license you get 9 PAFNow Custom Visuals, Predefined Report pages which you can all custom redesign with your own house style. On top of that you can you get access to the PAFNow Companion, which is a great tool which allows you to do transformations and thereby helping you to get the most out of your data.

If you are interested in using this tool on a bigger scale (or small scale) Dynamic People is here to help. From identifying which event logs could be of most interest to you within your company, to setting it everything up in Power BI and analyzing with you as a team.