Last year I was managing a small team that was managing software delivery. We were doing small-scale software development. It was nothing fancy but coding, stories, and priorities were needed anyhow. I had decided that we’d do it in the spirit of #noestimates. I needed a way to structure the stories.
I ended up with a board like this. Let me explain how it works.
First of all, it is based on the notion that cheap and valuable is great. That’s of course a no-brainer. You’ll have to figure out yourself what is valuable, but I’m sure you can hack it.
Second, the cheapness is the inverse of how much time you guess you’ll spend there. However, here’s the #noestimates twist to avoid excessive estimation.
All stories on the board are forced-ranked on the two axis: value and cost. You’ll all the time rank the stories against each other. That is: story A is more valuable than story B. And the same on the cost side. That’s how you end up with a two-axis forced rank. The green dots would be your stories.
Once you start doing that, you’ll quickly realize that it is quite doable. For instance a simple bubble sort will put the stories where they “belong”.
The top right corner is “magic” (the green area). That’s what you want to do next. It is sorted to be your best options. You can quickly trash anything on the bottom-left (the red area), and anybody can see why you did that.
Then we have two quadrants which are not as obvious. The left-top (cheap, not that valuable; yellow area) gets also rather quickly trashed as you see that there’s so much better stuff on the right side. Again, people are quick to agree on that. If there’s no agreement, maybe you missed some point in the value-sorting. Revisit the forced rank on value of that story.
The right-bottom (again green) is the “very valuable, very hard” corner. You’ll want to achieve them, but you either don’t know how to do it yet or you know it is going to be costly. For the latter category, try splitting the prime value out with different cost. These stories should bubble upwards. For the dilemmas, spike them. Maybe the cost isn’t as bad as you thought. For those stories that don’t seem to yield any options with less cost, you’ll end up keeping there as reminders of the great ideas.
The #noestimates development engine is all the time fed with the smallest possible stories. I personally noticed also that the selection becomes biased towards making the small stories with demonstrable benefit. Maybe it is my personal preference on that, but I think it could be even universal. It is easier to bet on a item that has low cost and some benefit, rather than a larger more undefined item with possibly more benefit. Those you want done for sure, but you’ll split them first so well that also the cost side is in control.
So, you’ll end up:
- Throwing out all fluffy and non-valuable stories from your board quickly
- See visually where to focus on the splitting work (likely the right hand side, just below the magic corner)
- Making the most value out of your efforts
- Doing it without the need of formal estimation as all you need to know is the rough relative size of the effort
Okay, one will argue that there is estimation going on here. Yes there is. However, that’s going on pretty much only in the top-right corner. And there as well to the extent that which of these stories is smaller. It doesn’t really matter if the rank on cost (or even value) is precisely right outside of the magic corner. Those items are not getting to development without splitting and re-evaluation. The ranking (i.e. cost estimation) rather simple for stories that take a few days to complete. And it doesn’t really need to be accurate again as you’ll end up doing all the stories on top-right corner sooner or later.