How do you know that is a 5 point story (cont)?

In my previous blog article on this I implied that you can do this in two ways

• estimating relative size, risk and complexity
• using a normalised scale

For the latter we need to use hours (how else can we normalise?)

For the former we don’t necessarily need to use hours to compare relative size, complexity and risk  – but then how can we check how we did?

If the team are happily estimating using story points alone and at the end of the iteration managing to get all stories accepted then you don’t have a problem – unless at some stage you need to compare this team with other teams as their story point yardstick will differ, possibly by orders of magnitude. But how can you tell this if one or both teams don’t use hours?

If the team are regularly struggling to complete an iteration cleanly then using story points alone may not be sufficient to provide the data as to what is going wrong. If the team estimate using hours and have a task that requires 20 hours still to be done by a single resource and there are only 2 days left of the iteration then you can predict that that story is unlikely to get completed and accepted. If you have just a task within a story then how can we predict that? Adding task estimates in hours may often be a key part in helping a team to understand where effort is going and so help them to correct carry-over issues that they may have.