Control Chart Analysis

Paul Clais April 3, 2013

Hi guys,

I'm trying to figure out the different time spend in each steps of our developement process;
I setup a board with 3 columns + done:

  • UpStream (aka To Do)
  • In Progress (aka in Progress)
  • DownStream (aka in review / to be release / qa )
  • Done (aka Closed)

I'm trying to segragated time Upstream, In Progress and Downstream, and while looking at the average values for each of them, I get figures :

  • Upstream : 6 days, 16 hours, 1 minute
  • In Progress : 15 days, 21 hours, 48 minutes
  • Downstream : 16 days, 5 hours, 13 minutes

Sum = 38 days, 19 hours, 2 minutes

(don't worry about the actual lenght of those, I'm actually more looking at Epics than sprint tasks -:) )

When I select my 3 columns directly, I got average time

  • Upstream + In Progress + Downstream : 42 days, 5 hours, 50 minutes

In my understanding, those 2 numbers should be exactly the same. I'm obviously using the same time frame and the same filters and swimlanes.

The same behaviour is reflected in the other metrics (standard deviation, running average....)

So my first question is :

Why are they not the same ? Where are those 4 days ?

My second question is about the "Done" column. There is stuff in there, mainly a few tasks that have been "repopened". Can I be certain that those tasks are nevertheless taken into account in the other columns ? Also, are they not introducing the issue above ?

Thanks;

Paul

2 answers

1 accepted

2 votes
Answer accepted
Ruchi Tandon
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
April 14, 2013

Hey Paul,

I think I understand your questions here. I believe the difference in numbers is expected behavior. It is a matter of adding up individual averages and considering it same as cumulative average. This will not be the same because of the way cumulative averages are calculated.

Individual averages are calculated like this : (Time elapsed in stage A)/(number of issues that were in stage A)

Total of individual averages will look like : (Time elapsed in stage A)/(number of issues that were in stage A) + (Time elapsed in stage B)/(number of issues that were in stage B) + (Time elapsed in stage C)/(number of issues that were in stage C)

Cumulative averages (which shows up when you select all 3 stages A,B and C on 'Refine' tab) are calculated like this : (Time elapsed in stage A, B and C)/(number of issues that were in stage A, B and C)

As you can see, there is a huge difference in the way these numbers are calculated and hence it is quite normal that they don't total up to be the same, especially if not all issues didn't make it to all stages (A, B and C).

I hope that explanation helps a bit...

0 votes
Paul Clais April 14, 2013

Hi Ruchi,

thanks for the feedback, it makes total sense. I'll check if this is effectively the way it's calculated, and I'll get back to you on this;
Cheers;
Paul

Suggest an answer

Log in or Sign up to answer