A couple of weeks ago, I discovered Jeff Gothelf’s blog. Check it out, there are some interesting articles about UX, Lean, and Agile. Among these, the first one to catch my attention was about an original Agile game, card aging. In short, it introduces a time parameter to stories so that corrective actions can occur before the iteration ends — read the article, it is playful and well written.
After contacting Jeff on twitter, I decided we would try the game soon. Having done that for two iterations, here are our thoughts.
While we found the game productive and fun, it does not quite make sense for us because our iterations are 1-week long, whereas Jeff’s team has 2-week long ones. Therefore, this new timeboxed approach to individual stories is too long for our weekly span. In the game, a story has 4 days to be completed. We already do that by default when commiting to 1-week iterations.
Naturally, we could make the cards age faster, but it still does not change much: 4 steps in a 10-day range translates as 2 steps in a 5-day range. Again, we already do that by commiting to showing progress in a shorter span. Far from me to say that our team is faster or more productive than Jeff’s, I merely justify the game mechanics and further, our iteration length.
The ensuing question is, what is the rationale for iteration length? Obviously, for a workflow to be considered agile, one must introduce enough checkpoints and reviews, or risk falling back into a waterfall cycle. But what length is optimal? I think it depends on the complexity of your work, and the availability of your client. 2 weeks seems a perfectly reasonable timebox when you must produce complex code clear of bugs and still be able to show any progress to your client. Additionally, not all clients are as available or as involved as one may dream… At graphility, we deal more with design and photography than with development and interaction design. Since our work is so visual, clients are excited to see it fast, and we strive to show it as soon and as often as possible to get feedback. 1 week iterations make sense to us.
My parting words would be to consider your iteration length and reflect on why you chose a particular timebox. This simple musing may give you insights on how to make your process work for you. Be sure to give the game a spin, though ;)