Estimation Min. Max?

[Fabian Noethe]

Hi,
how do i teach Jira to work with Min. Max. estimation.

E.g. "I need 4-5 days for the user registration"

Thanks

[Beth Schaefermann]

I am seeking a similar solution and did not read the question as out of scope of the tool or related to dependencies.

Some approaches to estimation of a work unit use a range instead of a single original estimation figure. The value of using a range is that you can then see how confident the team is in their estimations. For example, you might have a story that has an estimation low of 2 story points and an estimation high of 4, while another story might have a low of 2 and a high of 10. This tells you that there is a lot of risk and "unknowns" in that second story, whether it be unclear requirements, untrained staff, or something never yet attempted. As a product owner, you can then assess the risk in relation to business value to decide whether to pick it up first or punt it. At the very least, you have identified it as one that needs some more preparatory attention before scheduling.

You then want to track logged time against the range instead of a single original estimate in order to understand whether your team is tracking first within the range or on one side or another with any frequency and second close to the aggressive estimate. Using a range also helps to manage client expectations, especially if you have one that fixates on an exact estimate number. Using a range helps to psychologically reinforce the definition of an estimate as a timespan in which you think you can deliver.

I have searched:

• Atlassian Answers
• Atlassian Plugins - https://studio.plugins.atlassian.com/secure/Dashboard.jspa
• The Jira project itself - https://studio.atlassian.com/browse/JST
• General online searches and in the Jira documentation (hoping to potentially use custom fields to solve the problem)

but I have not yet found either a feature request, work-around, or plug-in that might prove a viable solution to the stated need. Maybe someone in this community has figured out an approach for range estimation or can recommend resources? If so, please throw some pointers my way - it would be much appreciated. :)

My current options appear to be:

• Advise using a single estimate ("ideal days") approach which should represent the buffered midpoint between the range (you could have custom fields for "Estimation range - low" and "Estimation range - high", but then your team would have to manually do the math for the buffered "Estimation - original"; or
• Advise using a "confidence" custom field with a percentage that represents the team's view on risk inherent in the story; this I believe would require running the team through calbration exercises, which can be time-consuming and require a more mature team to "get it"

[Jeremy Largman]

Hi Beth,

I think you're looking for GreenHopper: http://www.atlassian.com/software/greenhopper/tour/sprint-planning.jsp

If you re-do your searches using 'GreenHopper' and 'Story points' and 'estimate', you'll find discussions around the Internet that might help a bit. Some that I uncovered:

• http://stackoverflow.com/questions/6242709/jira-greenhopper-how-to-do-agile-correctly
• http://forums.atlassian.com/thread.jspa?threadID=53329&tstart=0&messageID=257375280
• https://answers.atlassian.com/search/?t=question&q=estimate&Submit=search
• https://answers.atlassian.com/questions/7/what-method-does-the-greenhopper-team-use-to-do-estimation

Anyways, I hope that's a start.

[Beth Schaefermann]

Thanks, Jeremy, for the follow-up. While I am a newb and still learning, I don't think Greenhopper delivers the functionality I am looking for. Per my understanding, Greenhopper is a visualization of issue data with a few additional functionalities and hidden fields (such as Rank). I do, though, certainly appreciate you pushing me towards the Atlassian Forums which I had not yet explored.

Given that I could find nothing under "estimate range" or "estimate min max" on the forums, I think I will try to push a methodology change on our part as opposed to submitting a new feature request at this time. Jira's single estimate approach is, IMHO, simpler both in training and in tracking, although it does not relate perceived risk and estimation confidence. Perhaps there are other ways in which our team can accomodate the need, such as either using a custom Confidence percentage field or using the original estimate as the Estimate-low, and creating a custom field for the Estimate-high.

[Jim Birch]

Jira isn't a project planner with temporal dependencies ("A then B then C and D") so min/max esimates don't mean much, use your best guess and adjust. The work estimates are more aimed at answering the question how long to finsh with x developers working in parallel.