Velocity Chart - How to verify that the data is correct?

[Rashmi Ratnani]

Hello! 

I feel there is something amiss in the velocity charts i.e. committed vs completed hours reflected. So I tried to download the values in excel and match them. They do not match - clearly I am missing something. Can you please help me understand what's the issue or how to better read the reports?

How we add estimates:

• Story, Bugs : We look at the Sum of Remaining Estimate field as we have sub-tasks
• Task: We look at the Remaining Estimate as we do not use sub-tasks

When I downloaded the parent items from Sprint 55, I would expect the following:
Committed value = Sum of Original Estimate

Completed value = Sum of Time Spent

What am I missing?

/RR

[Vasyl Krokha]

Hi @Rashmi Ratnani 

You cannot verify whether the calculated velocity in Jira is correct. You can only trust it or not.

As mentioned earlier, the complete calculation of sprint metrics includes processing historical data - issue change history.

Overall, the following math must work:

Final commitment = Initial commitment + Added work - Removed work + Estimation change where:

• Added work - issue added into sprint after its start
• Removed work - issues removed from sprint after its start
• Estimation change - issue estimate changes inside active sprint

In our reporting apps Velocity Chart and Agile Reports we implemented such historical data processing in order to add those metrics explicitly to Velocity report/gadget, but I can tell you, that's more than a hundred lines of code)

When we perform a cross-check with Jira numbers, in most cases if we have a diff, it is our issue, not Jira's.

Hope this helps.

Vasyl

[Dick]

I'm not sure which calculation you refer to (the excel is not quite clear)
But I've seen calculations with and without the weekend giving rise to differences.

[Rashmi Ratnani]

Hello @Dick 

Thanks for your inputs. I guess the weekend could affect the timelines of deliveries but when we are looking at the estimate numbers only, then I would expect the numbers to match. 

What the excel is about: 

I pulled out the sprint backlog for Sprint 55, and took only the parent items from the backlog. Downloaded it to excel to analyse. Then summed up the columns. So I would expect 

Committed value (in first screenshot) = Sum of Original Estimate

Completed value (in excel screenshot) = Sum of Time Spent

Hope this clarifies. 

[Dick]

Original estimate of an issue is a historical value (that could've changed upon refinement and later adjustments prior to the issue being in progress)

[Rashmi Ratnani]

Agree. Therefore I would expect that it is committed value.

As the sprint progresses, the remaining estimate should be updated which means that some time is spent. Time Spent is, therefore, calculated by Jira as Original minus Remaining. 

[Dick]

I see "original" as the first entry of the time estimate. This value could change prior to the ticket being "in Progress" .

Hence, you should poll the time estimate at the moment the ticket is pulled into the sprint, that could differ from the "original" !

[Rashmi Ratnani]

Hmm. That would be difficult and don't know how precise!

Is there any other way to check if this velocity is correct?