Agile Metrics

17 Jul
July 17, 2009

Agile Metrics

We all know the saying “measurement drives behaviour”. Therefore Agilists are understandably wary when management demands metrics to prove that their investment in Agile practices is paying off.

I’ve done some work to identify a small set of Agile metrics that I believe to be helpful. I have drawn particular inspiration from an open space workshop facilitated by Pete Behrens (CSC, CST) at the Orlando Scrum gathering and have also been helped by my colleague Mike Freislich.

This PDF contains my conclusions thus far. I must confess that not all of these metrics are yet in place at any of my coaching clients, so I do regard this as a work-in-progress and expect the power of empiricism to show me my mistakes over time. I would also value the input of my reader ;-).

Other blog Posts written by: Peter Hundermark

Tags: , , ,
  • http://carlokruger.com Carlo Kruger

    All of these are good metrics and easy to collect with one exception. I think Technical Debt is an important metric to track, but using only the story point value has two problems I can see.

    One, figuring out the story point value will be a difficult one since in most cases the true impact will only be felt by the development team (refactoring debt should have no impact on functionality). Planning poker-ing should mitigate some of this, my concern is that there “should” be no impact on the non-dev parts of the team. I mean you have a comprehensive automated regression suite which runs on check-in and hence the the test and analyst types can use this time to sharpen their pool playing skills.

    Two, it ignores the aspect of accumulating technical debt which is what the impact on the ability to deliver business value in the future will be. So I think it would be important to have the team estimate the ‘Business Value or Risk’ of the technical debt item. I would also tend to see the Technical Debt accumulated as a leading indicator of a drop in velocity of the team, and would try to forecast what that impact is, once you’ve accumulated some supporting data.