Agile Metrics

Joan Díaz
3 min readApr 26, 2018

If you are an addict to metrics and all this kind of stuff is such an interesting topic, isn’t it?! After reading last David J. Anderson’s and Alexei Zheglov’s book “Fit for Purpose: How Modern Businesses Find, Safety, & Keep Costumers” — which I get it at the very beginning of this year at the KMP Course (Kanban Management Professional) at the LeanKanban University — I started thinking about a couple of interesting things… To be honest, this book is more related with business but it made me link what they talk about with other ideas.

Which KPIs do you think a team should use? Well, it depends on the team, but mixing what Anderson & Zheglov says plus other sources like Daniel S. Vacanti and Jurgen Apello, we could say that metrics must taking into account: fitted for a purpose, useful to do predictions and help the team to learn and improve. If any of the metrics you are currently using or you are thinking to use, if don’t have these points then you are using vanity metrics (which only helps to calm down anxiety on particular people, that’s it).

First of all, forget having one single number because we are in front of a very complex problem and, as you might know, for every complex problem there are a simple and easy answer, which is wrong. Then you should build a set of metrics, as a beginning we could classify these metrics like: KPI, Goal or Improvement driven and Health metrics.

  • KPIs:

Few metrics, close related with purpose and also related about what upstream or business are expecting about you. As a beginning could be Lead Time or anything related with Quality — functional (what you do) or non-functional (how you do it), like % of bugs per feature or something like this. They have a threshold, because you could giving a bad service or surpassing expectations, both sides are important because you want to fit the team to a concrete purpose.

  • Goal or Improvement driven:

Are those metrics that helps you to get the KPI inside the right threshold and obtain a clear outcome, for example: Throughput (you will try to increase throughput in order to reduce lead time) or number of new automated tests (useful to improve non-functional quality and reduce % of bugs). Once the goal are attained, then you can move it as a Health metric, I mean having a high throughput or new tests is also healthy.

  • Health metrics:

A subset of metrics that: were goal driven or also helps the team to know how healthy they are. If you improve them then you became more healthy but not better, you should use them as a global picture of the team not as a goal driven, observe how they are moving, tendencies, etc. As an example could be: testing % of code coverage, happiness ratio (with Nico-Nico calendar), number of interruptions, peer recognition (with Kudo Cards), etc.

Here is an example that I think is quite easy and I hope it will help you to understand this classification, actually is one of the examples that Anderson & Zheglo use on their book. It is about an athlete who likes running Marathons:

  • KPI: is the time he does running Marathons, at domestic competitions.
  • Goal driven: He wants to run the Marathon with 3 hours, so lets say that he have to reduce about 20 minutes his last marks. Then we can use increase pace (minutes per kilometer) to reduce overall time.
  • Health: Heart rate, number of weekly practices, average of kilometers run per week, pace (minutes per kilometer) at short runs or on slopes…

Which metrics do you use? Please share them with us! :)

Thank you for reading this text, if it likes you please applause it, share it, comment it!!



Joan Díaz

Agile, Lean, company culture, new ways of management or self-management enthusiast.