Sunday, February 24, 2013

The ScrumButt Test

Bas Voode originated a test.  7 or 8 items.  Jeff Sutherland and others liked it.  It got revised a bit.  Eventually it became known as the ScrumButt Test. 

They developed the test to check whether a team is really using Scrum or reverting back to waterfall.  Or, possibly, reverting to what I call Cowboy Agile (see wikipedia on cowboy coding). Or doing Agilefall (talking Agile terms, but really doing mostly waterfall).

The ScrumButt Test is in two parts.

First, are you doing Iterative Development?
  • Sprints must be timeboxed to less than 4 weeks
  • Software features must be tested and working at the end of each iteration
  • Sprints start with an enabling specification
The experience is that if you ask a lot of "Scrum" teams if they can pass this part of the test, they can't. If you are at a conference, often not a single team in the room can pass.

The next part of the test checks whether you are doing Scrum:
  • You know who the product owner is
  • There is a product backlog prioritized (mainly) by business value
  • The product backlog has estimates created by the team
  • The team generates burndown charts and knows their velocity
  • There are no project managers (or anyone else) disrupting the work of the team
My reaction:
I think this is an excellent way to start to address Cowboy Agile or Agilefall.

Let me say this loud and clear: a firm can't in good faith say "we tried Scrum" if they can't pass this ScrumButt test. And pass for a reasonable period of time. Of course they could say "we tried Scrum", but they did not. (I will not define right now what 'pass' would require.)

The ScrumButt Test is demanding. That is clear. But a test of this nature is a necessary (if not sufficient) condition to doing professional agile software development. (Or did we expect to get out of Fred Brooks' tar pit by doing unprofessional software development?)

There are some forces in a firm that want to do Cowboy Agile or Agilefall or want Agile to fail ("it's moving my cheese"). This is true in every firm that I have worked in, I believe. So, if you use the ScrumButt Test, expect to get some resistance.

The  Test is arguably too minimal. There are many other important parts of Agile or Scrum that it does not cover. Are the things in the test the most important parts? My thought is that this collection of principles and practices provides a good core of Agile/Scrum that will defend you from the most common dysfunctions. Not all. 

Is there a better test? Probably we could define one, but let's pass the ScrumButt Test now.

Should the test be significantly more detailed? I think not. First, to work with any test, we need the test to be simple enough to be comprehended easily by most of the people involved. In this way, it becomes self-policing. Second, Agile/Scrum needs to be adaptive, so defining too many rules would, at some point, be counter to self-organization. 

What are your thoughts? Are you aware of similar tests? How does your firm limit Cowboy Agile?

No comments: