Hours or points for sprint planning seems to be quite an opinion splitter in the Scrum community. Even the Scrum heavyweights Mike Cohn and Jeff Sutherland have chipped in with their opinions with both landing on either side of the argument.
I’ve discussed this a few times on the Scrum Alliance LinkedIn group and whenever I put my point across I’m not sure if my point has ever really come across. To be clear I always estimate stories in points (i’ll save the reason why for another blog) and always use story points for velocity calculation. I then always break tasks into hours. At the start of any sprint planning session I will always show the team their historical velocity for the last few sprints (no more than 5) and ask the team if there is any known absence for the next sprint. At this point I will ask the team what they think they can achieve in the coming sprint. The Product Owner should know the teams velocity and should have a good idea of what this team usually achieves in points to help them decide what to bring in.
Once the team has agreed on what they can achieve the PO will then discuss the stories followed by the team breaking the stories into tasks. From my experience a team using points may choose not to put a time estimate on their tasks if they are happy with their estimation. I always encourage a new team to do this to get a correlation between hours and points. It is also really helpful for teams to compare stories of similar point estimate to the number of hours required to get it done.
Many people in the hours camp suggest that teams who estimate in points are only committing to doing that number of points when they may be capable of more, hence the use of capacity and hours. This is where i strongly disagree. I have NEVER worked with a scrum team who have met their agreed point allocation early and haven’t taken more work into a sprint. By planning a sprint with points you are using factual data of what the team historically achieves as a baseline, it certainly isn’t set in stone.
What I do like about capacity and hour planning (especially in TFS where this for me is one of its best features) is it helps the team and PO see if there is immediate scope by the amount of work vs remaining hours. I plan sprints based on a 60% capacity of hours to allow for changing requirements, meetings, distractions amongst other things. An over zealous PO may try and use this as pressure to push the team to over commit (especially with a visual indicator that a sprint is 60% full of time) but this is where the Scrum Master must protect the team.
In summary i think both approaches can be used in tandem. Use points for a good estimate and look at the hours to both check story points are correct (e.g. a 3 pointer with 4 hours work versus a 3 pointer with 15 hours would suggest the estimates are incorrect).
I would always encourage the team and PO to discuss the commitment and what they feel they can achieve. A PO will always want more done than is agreed but this is the nature of the role. A team and PO who understand each other on this commitment will always work better than a team and PO who are suspicious of each other.
What do you guys use? Points or Hours?