- My organization has just started with Scrum and we are in the 2nd
real sprint. On a daily basis we are gathering task level data for
the sprint burndown chart and further analysis.
At the end of each sprint, it would be nice to look back at the work
accomplished and look at sprint metrics and how a particular sprint
compares to previous sprints. Hopefully, a positive trend will
emerge, which proves that the team is getting better.
Examples of metrics I can think of are:
total hrs estimated
total hrs consumed
total hrs. remaining
act % of estimate (remaining + worked vs. estimated)
days in sprint
sprint capacity (people)
sprint capacity (hrs consumed)
capacity per day and person
# of stories
# of tasks
# tasks completed
% of tasks completed
What metrics have you found to be useful and what data are you
gathering to calculate them?
- Not necessarily that people do more of what you measure, though measuring will have impact on their behavior. For example, Arbitron asked if I'd participate in their gathering of ratings for radio listening. Basically, all I have to do is write down in a journal when I listen to the radio, what station, the time and where I was (at home, in the car, etc.) for a week. There have been times this week when, rather than turn on the radio, I've listened to a CD or podcast...just so I wouldn't have to write the information down.
A lot of times there is a reward, even if it isn't explicit or monetary. If the data is collected and publicly presented with an interpretation (more is better, less is better, close to this line is better, etc.) that is understood, then there is a reward of attention/praise/satisfaction, even if it wasn't explicit or monetary or external to the team.
So, you need to be careful when you decide what you're going to measure that you understand how the effort to collect the data will affect behavior, how the use of the data will affect behavior and if those perturbations of the "system" and the costs associated with data collection/storage/presentation provide enough value to making decisions.
michaelOn 6/5/07, Peter Hundermark <peterh@...> wrote:
--- In email@example.com, "Matt Truxaw" <mtruxaw@...>I have observed that people do what you measure. So the answer to
> I am late to this conversation, but I would like to comment on the
> discussions and to reframe the question slightly.
> First off, I know that you do not HAVE to measure something to
> improve it. You only have to measure it if you want to know
> or not you improved it. Sometimes, you accidentally do something
> that makes an improvement in a situation, but if you do not have
> some way of measuring that, you will never know that the
> happened nor will you be able to repeat and build upon it.
> Secondly, measurements can be subjective but objective measurements
> are much more easy to compare. That doesn't mean objective
> measurements are necessarily more accurate or valuable, but they
> much easier to work with.
> My reframed questions are:
> * Are you gathering any metrics or other objective measurements
> about your sprints?
> * What are they and how do you measure them?
> * Do you find that measuring these things changes the behaviour of
> your individuals and teams?
> * Do you reward/punish based on the metrics?
> * Do you find that rewarding/punishing these things changes the
> behaviour of your individuals and teams?
> * Do these changes in behaviour help or hinder your team efforts?
> How do you know?
your 3rd question is always 'yes'. It does not even require that you
have explicit punishment or reward. The corollary is that you need to
be very careful what you measure. Do others have this experience too?