What is the best way to work with estimating and assigning tasks and sub-tasks in JIRA?

Jonathan Crowe August 7, 2017

Our current setup which is causing us some headaches as it's not as clean.

  • Epics
  • Stories
  • Sub-Tasks

For this example, let's just talk about Stories and Sub-Tasks.  We assign story points to Stories for planning but then assign hour estimates to sub-tasks.  

Problems:

Assume the scenario is setup as such: one user story with two sub-tasks.  Each sub-task is 4 hours and each assigned to a different developer.  

  1. Before we start the sprint, we'll look at what each assignee has allocated to them in in regards to time.  If you assign the story to a developer, the total workload for that developer becomes 12 hours (4 hours for sub task + total sum of the remaining estimate on the parent shich is 8 hours as it takes into account both sub-tasks).  Therefore the remaining estimates are double counting themselves
  2. If we leave the stories unassigned, it'll fix our planning and estimate problem, but then nobody has accountability to close out the story.  We have filters on our board that lets users see their own issues
  3. Our board has quickly become overloadeded with items and they aren't grouped in a nice fashion.  We tried swimlanes but wasn't a fan.  What are some ideas how keeping the board in a better and more organized fashion versus having sub-tasks spinkled throughout the board.

1 answer

0 votes
Jack Brickey
Community Leader
Community Leader
Community Leaders are connectors, ambassadors, and mentors. On the online community, they serve as thought leaders, product experts, and moderators.
August 7, 2017

My recommendation is to leave stories unestimated but assign to someone - either the scrum master or one of the sub-task developers. Check, Include Sub-tasks in the Time Tracking area to ensure the story reflects some of the parts and progress.

Generally speaking you should have a 'primary owner' of every story as they need to ensure the pieces come together and then to ensure the story is marked as done. Another approach is to assign the Story to a 'QA' person to verify the story and close it. Of course they too should have a sub-task in the story for this effort.

Bottomline - estimate on the sub-task not the story. You can always pre-estimate on the story until the details (sub-tasks) are fleshed out and then remove estimate on story. 

Jonathan Crowe August 9, 2017

Thanks @Jack Brickey Understood and I agree, always estimate on the sub-task, not the story.  That's exactly what we're doing.

We want to assign the story to the developer to take ownership and as you said will be the 'primary owner' but that is the exactly problem we are facing.

Scenario 1

Story - Unassigned
Sub-Task (2h) -- Dev 1
Sub-Task (2h) -- Dev 2

Planning:
Dev 1 has 2 hours of work estimate remaining
Dev 2 has 2 hours of work estimate remaining

Scenario 2 (what we want but causing a problem)

Story - Dev 1
Sub-Task (2h) -- Dev 1
Sub-Task (2h) -- Dev 2

Planning:
Dev 1 has 6 hours of work estimate remaining (2 + all sub task time)
Dev 2 has 2 hours of work estimate remaining

Question: How can we perform scenario 2 without doubling up the remaining estimate?

Jack Brickey
Community Leader
Community Leader
Community Leaders are connectors, ambassadors, and mentors. On the online community, they serve as thought leaders, product experts, and moderators.
August 9, 2017

@Jonathan Crowe, thanks for the additional info, it helps me see more clearly and indeed S2 is undesirable. The described behavior seems odd on the surface for a couple of reasons. First I was thinking that if there was not time estimated on the story then there would be no time allocated to the assignee. Second you are seeing 6hrs when the total of the sub-tasks is only 4. So there is double counting. So i'm perplexed but likely because of a lack of understanding on my part. It has been awhile since I was using the scrum boards and sprints so my memory is stale. We are using kanban for everything lately and not doing any burndown activities. I will see if I can dedicate some time to look at this more closely. In the meantime maybe others will chime in.

For now, if it were me, I would stick w/ S1.

Jonathan Crowe August 9, 2017

@Jack Brickey apparently this is a common issue that has been around since 2013 and Atlassian has neglected. Not sure why as it appears to be a top request. 

https://jira.atlassian.com/plugins/servlet/mobile#issue/JSWCLOUD-9167

Suggest an answer

Log in or Sign up to answer