Throughput Time for subprocesses versus throughput time in Process Explorer

#1

Hi, we have build up a deep-dive analysis on cycle times where viewers can select two activities to measure the throughput time for this subprocess. However, the time never matches the time on the Process Explorer when switching to Throughput Time (MEDIAN). Here is the formula we are using to calculate the throughput time between the selected activities from the drop down:

MEDIAN(CALC_THROUGHPUT(FIRST_OCCURRENCE[<%=From%>] TO FIRST_OCCURRENCE[<%=To%>], REMAP_TIMESTAMPS("_CEL_P2P_ACTIVITIES".“EVENTTIME”, <%=unit%>)))

The variables [<%=From%>], [<%=To%>] and <%=unit%> work fine and are correctly connected to the dropdown menu for the activities and the unit (I’ve ensured by hardcoding two activities and the respective unit as a test in the formula --> same result than with the variables and selecting the before hardcoded activities from the dropdown).

I guess the difference might come from different variants between two activities. How does the Process Explorer work for throughput times (taking all variants into account in the happy path for the throughput time between two activities or only taking the time for the currently shown variants)?

E.g. analysing “Release Purchase Requisition” to “Create Purchase Order Item”:
In the Process Explorer I’ve 146 hours between those two activities whereas in the subprocess calculated with the formula from above I end up with 166 hours. When selecting activities which are usually not followed after one another, the differences becomes significantly higher.

#2

Hi Oli,

can it be the case that some activities happen multiple times (then it might lie on “first_occurence”)? Can it be a rounding Problem where you should use Floor or Ceiling to get the same results or just Median*(1.0*Calc_throughput()))?
It also might be what u mentioned: direct and indirect following. As your Formula calculates also indirect following (i.e. also the case when “Release Purchase Requisition”->some other activity->“Create Purchase Order Item”) and the process Explorer shows the throughput for the direct connection. It is then quite obvious why u get higher numbers with your formulas…as the activities between “Release Purchase Requisition” to “Create Purchase Order Item” usually increase the throughput between them.

Best,
Maria

2 Likes
#3

Hi Oli,

the reason for different values is well explained by @DrWindy .

The Process Explorer only shows the throughput time for the direct edges. It uses the SOURCE and TARGET operators to calculate the numbers.

You can calculate the numbers in an OLAP table using the following query:

Dimension 1: SOURCE ( "_CEL_P2P_ACTIVITIES"."ACTIVITY_EN" ) 
Dimension 2: TARGET ( "_CEL_P2P_ACTIVITIES"."ACTIVITY_EN" ) 
Dimension 3: MEDIAN ( MINUTES_BETWEEN ( SOURCE (  "_CEL_P2P_ACTIVITIES"."EVENTTIME" ), TARGET ( "_CEL_P2P_ACTIVITIES"."EVENTTIME" ) ) )

This would calculate the exact number of minutes, you can modify the query accordingly to get Hours or Days.

You need to first decide whether you want to display only direct edges like in the Process Explorer (then you need to go with SOURCE / TARGET), or if you also want to take indirect edges into account (then your CALC_THROUGHPUT approach is correct).

Best,
David

1 Like
#4

Thanks Maria and David for the very fast feedback! I’ve quickly set-up the SOURCE and TARGET table and figures are indeed matching the process explorer numbers. It’s good for me to understand the differences and decide what we really want to show and subsequently Analyse here :slight_smile:

1 Like