Root cause analysis basics

Look for probable causes option

You can begin an RCA investigation from all of the places where you find problems to solve. The origination point might be from a slow transition, an undesired activity, a model violation, or even a particular metric column that you are investigating, such as Case Duration.

As you work in Pega Process Mining, you see options throughout the application called Look for probable causes. This option is how you access and begin an RCA investigation.

From these examples, you can see that you typically reach for RCA when you are in the middle of analysis and find an interesting question to answer or problem to solve. This is the reason why Look for probable causes is available in so many areas of the application.

Root Cause Analysis page

You might also approach RCA in a more direct way. For example, you might already know that you need to look at specific columns to find correlations to known issues. Pega Process Mining provides an RCA page where you can start the analysis in this way.

To access this page, in the header of the application, click the Analyses icon. In the navigation pane, select Advanced > Root Cause, as shown in the following figure:

In the upper area of the Root Cause Analysis page, you can select a Column name and value and then click Analyze to begin the RCA.

The bottom section of the page provides shortcuts to the analysis pages where you are most likely to find issues that you want to analyze further. From these pages, you can initiate an RCA using the Look for probable causes option.

RCA basic use case

After completing initial process discovery and analysis, you have noticed that rejected Cases spiked last month. You want to use RCA to see if you can determine possible root causes of the increase.

This spike occurred during a supply chain disruption, and you suspect that these events are correlated. However, you want the review the actual data so that you can better address the problem.

To do this, you decide to look at the Rejection Reason column name and select Lack of spare parts as the the column value. You could look at multiple column values, but you want to focus only on this particular rejection reason. The following figure shows your selections:

After clicking Analyze, the system displays the results, as shown in the following figure:

By default, the RCA includes up to the first ten Case columns in the RCA evaluation, as shown in the left pane. Based on your specific analysis, some of these columns are not applicable, and can remove them to clean up any Null values in the results. In this example, Cancellation Reason is not applicable because you are looking at rejections, not cancellations. You clear the Cancellation Reason checkbox and click Perform RCA to run the RCA again:

You want to ensure that you take into account any filters applied to the data. You can check the filters by reviewing the Applied filters section in the left pane. In this example, the only filter is the Case Interval, which includes the month of data that you want to analyze. The results list orders the possible root causes by the number of Cases affected. The results show both the number of affected Cases and the percentage of affected Cases, as shown in the following image:

From these results, you see that Case rejections caused by a lack of spare parts correlate with battery installations. As you scroll down the list of results, you also find correlations to New York City, Company C, and Texas, as shown in the following figure:

Each result also displays a frequency ratio. In the following figure, the Type value of Battery is 37.48 times more likely to be correlated with the rejection reason of lack of spare parts than average:

Based on the RCA results, you now have a much more focused direction to continue your analysis. You are going to look more closely at battery parts and what happened in particular in New York City and Texas.

To continue the analysis, you can use graphs, Cases, and dashboards to make even more connections. Each possible root cause provides links to these additional tools:

For example, for the possible root cause of Battery type, clicking See in graph displays the Process Map with the Battery type filter applied, as shown in the following figure: