The root cause analysis enables to find reasons for certain anomalous behavior of single process instances. This MPM module is optional, when used it unlocks the analytics in the sheet Root Cause Analysis.

 

MPM root cause analysis provides a set of 9 anomalous behaviors that are by default derived due to the MPM algorithms. These are:

•Happy Path Fitness Below 80%

•Case Lead Time Exceeds 75 Percentile

•Case Lead Time Below 25 Percentile

•Total Case Event Duration Exceeds 75 Percentile

•Total Case Event Duration Below 25 Percentile

•Total Case Idle Time Exceeds 75 Percentile

•Total Case Idle Time Below 25 Percentile

•Case Lead Time Exceeds Median

•Case Lead Time Exceeds Average

 

Furthermore, the Root Cause Analysis offers to analyze:

•critical process behavior detected by process query (see section QueryBuilder),

•anomalous custom lead times (exceeds 75 percentile or stays below 25 percentile),

•customized anomalous behavior (see section How to Perform Root Cause Analysis on Custom Process Anomalies).

 

 

Performing root cause analysis in the MPM Template App requires only five simple steps:

 

1.Go to the root cause analysis tab.

 

2.Define dimensions and attributes for the root cause analysis by adapting the inline table RootCauseAnalysisDimensions.

The first column takes the name of the table to which the dimension, aka the field in the data model, belongs. The second column takes the dimensions name aka the field name.

The MPM Template App will look like the first code example, once filled in the inline table might look as the example below taken from the helpdesk process. The MPM root cause analysis does only account for dimensions, that are encountered in the central CaseInformation table (name can be changed) but it is important that this table holds one row of context information for each case and that all dimensions needed for root cause analysis are centralized here.

 

 

RootCauseAnalysisDimensions:

load * inline

 

 

Helpdesk example:

 

RootCauseAnalysisDimensions:

load * inline

 

 

3.Select the default anomalies that should be considered.

The default anomalies are listed in the following inline table. To incorporate them into the root cause analysis, set the first colums "CalculateInRCA" to "y" for yes or "n" for no.

 

 

RootCauseAnalysisFlags:

load * inline 

 

4.Optional: Define critical process paths by queries to explain their occurrence.

By adding lines to this inline table, cases which follow the critical process variant will be marked and the reasons for this critical deviation will be extracted from the data. To do so, perform the following steps:

a)Open the MPM Extension QueryBuilder and define a critical process query.

b)Open the QueryBuilder in edit mode and copy paste the resulting process query to the first column "CriticalPathQuery".

c)Give it a unique and speaking name, as the name will be displayed as the problem description in the root cause analysis' front end.

d)If you do not want to employ this type of analysis, just delete the rows from the inline table.

 

[CriticalQueries_$(mvProcessDescription)]:

load * inline 

 

5.Optional: Explain rare custom lead times by the root cause analysis by setting the variable mvCustomLeadTimeInRCA to "1". When set to "0", the root cause analysis will not derive reasons that have caused cases to have an extremely high or low custom lead time.

 

LET mvCustomLeadTimeInRCA = 1;

 

6.Lastly, the MPM root cause analysis will be called. You do not need to change anything here, as long as the table containing the dimensions is called "CaseInformation". If this is not so, rename the third parameter to the name of the table holding the context information on cases.

 

call  mw_perform_root_cause_analysis('$(mvProcessDescription)','$(mvCustomLeadTimeInRCA)','RootCauseAnalysisDimensions','RootCauseAnalysisFlags','CaseInformation','CustomLeadTimes_$(mvProcessDescription)');

 

 

Tutorial

 

Start the video below to see a short tutorial about how to use the sheet.