Measuring Business Value With Metrics
by Kelly Waters
Discussing software development metrics at my place of work, a colleague (Derek Morrison) came up with a neat concept - a way of measuring Business Value.We're using Scrum as an agile management framework, estimating in points and measuring Velocity to help plan future Sprints.
For those not familiar with Velocity, it's the total estimated cost (in effort or points) of the features that are 100% complete (i.e. signed off and delivered) in a Sprint or iteration. If you're familiar with Earned Value Analysis in more traditional project management methods, it's basically the same concept.
Each feature, or user story, listed on the Product Backlog has two Fibonacci scores. One for Size and one for Business Value. This is potentially a useful way to prioritise the Product Backlog, as a simple formula (value/size) can be used to order the backlog with the quickest/biggest wins first.
Derek's idea was to measure the Business Value of what we're delivering, as well as Velocity for the team's planning purposes. We referred to this metric as Business Velocity.
We haven't started using this metric in our teams yet, so it's too early for me to comment on how it's working in practice. However I was very interested to see James Shore's comments on the same.
Personally I think it has merit. So long as - as with all metrics - people don't play the system.
1 January 2008 00:56
We've attempted to perform the same thing. I'm afraid it didn't work out for us that well. The business users didn't really grasp the concept of relative estimation (8 point story is twice as valuable as a 4 point story and half as valuable as a 16 point story).
What we ended up with was different areas of the business trying to game the system in order to make their story land as high up the stack as possible.
We didn't (and still don't) have a formal product owner role, so I think that is a large part of it, and we also didn't ensure that everybody agreed on BV and took an average, which makes the system easier to game. We have since tabled the project and moved on to higher priority projects.
I'm interested in hearing how well this works for your team.
1 January 2008 10:52
Yes, a single Product Owner for your team - or at least for each product - is an essential ingredient for so many of the benefits of agile development. And certainly for this metric to work.
In fact, a single, engaged product owner is the key to solving so many development problems, even with a more traditional development approach.
Kelly.
6 December 2009 23:53
Take a look at classic function point analysis (http://en.wikipedia.org/wiki/Function_point) It actually encapsulates what you are trying to achieve and has a solid economic and statistical basis