A rant on estimation

This is a transcript of a team discussion we had around 'estimation', why we do it, what its for etc. The format for estimation is 'Relative Effort' using a fibonacci scale (1,2,3,5,8 etc). We had a team meeting to go over our estimation after we had 2 sprints with wildly different points delivered.

Transcript and outcomes we summarised by me, below (in a team Skype session)

Yes, the decision yesterday was to continue estimating, we just use certain estimates as triggers for further action, i.e. 5/8's are indicators that the story needs to either be preceeded by a spike, [Ed's note - this is a team decision, we have agreed within the team to try to standardise card sizes as much as possilbe, and our team standard was a '3'] or broken down if possible, we will try to get stories to be round about the same size wherever possible

No 2's as per current (team chose to drop 2's as they spent too long quibbling about whether somehting was a 2 or a 3, and this was deemed a waste of time and that estimation is a tool used to help business priorities, as well as us to visualise our sprint (to an extent) and that there may well end up being a range of 3's (much laughter about this) [Ed's note: At a subsequent retro, some team members observed that they had been under pressure to estimate everythings as 3's ]

Bottom line though, we need to be clear what the estimates at pre-planning/release planning are to be used for, as we can't (in 2-4 weeks) be held to estimates if product have used them to 'promise' delivery as in these are early estimates, not 'promises' and by product I mean YOU [Product owners name]..sorry need to stop using that term in the 3rd person

Essentially the team were a bit nervous that by providing 'number' estimates at this stage, they might be mapped to a previous sprint velocity and we might be held to a set of stories as a deliverable.

What we know about estimation is that we have seen swings within sprints, and that makes velocity a guess at best. Mainly we are indicating by estimating a story as a 3 that we are comfortable with the size and complexity of that story, that it along with some others is doable within a 2 week sprint.

A 5-8 point estimate indicates sufficient doubt in the story (unknowns) to conclude whether it is achievable without significant risk of its delivery within a shorter sprint, and that this story either needs breaking down, or a spike prior to commencement to reduce the unknowns

Comments

Popular posts from this blog

A survey on release processes with organisations that 'sef-identify' as folowing agile practises.

Software and Elephants - you won't believe how thin an elephant can get!

Using an agile storyboard as a scale of certainty