Introduction
For a while (3 years) we've been using a board with post-its, and through time it changed a bit. It started as a traditional Scrum board, with the To Do, Doing and Done columns and until last week it was much more elaborated. We were using story mapping on the backlog and on for the development process waiting queues and Limited WIP and swimming lanes but still we where a bit stuck with scrum, and with a board that needed refactoring.
A small change
We have discuses for a few months about abandoning scrum completely, for many reasons, but I'm not going that way on this post. Coincidence or not Alisson Vale and Rodrigo de Toledo (that is involved in our project for years) where about to present the course "From Scrum to Kanban" (http://www.scrumban.com.br/).
The course was excellent and gave me a lot of insights. A good one came from the drawing in Dorian Taylor post about the principle of one degree. The idea is that we are constantly assuming and validation theories about our knowledge work, and the graph gives us a clue about what we know and are trying to find out and what are our next assumptions.
Another good idea came from the fact that the board could be seen as a game, where you have decisions to make every day, and each movement affect the other players.
The Layout
The first thing we decided was that the board needed to change, it was hard to maintain, hard to read and definitely too boring. We though about it for a while and decided to try using the iceberg analogy to represent the backlog. It is a good idea because it is visually appealing and uses the physical space on the board to limit WIP. But then we realized that it would be nice to group a few features and business values, so we needed more than one iceberg, 3 or 4 maybe, but that is a variable quantity. So a member of the team came up with the idea off joining them in a circle came, with the icebergs as slices of pizza, and each slice being one specific area.
To maintain the same philosophy and symmetry we decided to use the same structure on the development cycle part of the board.
The other two small ones are for bugs and retrospectives, that are also limited by the circle size.
The Game Flow
So how are we moving post-its along the board? The main goal is to move post-its towards the center of the big circles but respecting the limits of each phase. So the outer ideas/BVI of the backlog, that are vaguely described, tries to move to the next step where we have story-like descriptions of a determined feature. At this point we should have more post-its than we originally had, but the inner area is smaller then the outer one, should it be the other way around? we are trying this way to help the job of cutting scope and let the most important and immediate stories closer to the center, in order the keep our focus on them, but still capable of looking on the future of each of each slice of our pizza.
The next step are estimated the stories that fall in our PMG sizes, if it doesn't we need to make it small enough. On the final one are the selected stories. They are ordered clockwise and are the next ones that should be made.
On the right side we have four slices, since we had four swimming lanes on our original board. on each one we place a story, use smaller post-its for tasks if it is appropriate, and they go through: To Do, Doing, Review, Done. Notice that this way each slice could have their own structure to go towards done, which is a nice thing if you have lanes dedicated to specific kinds of work (which is our case). We didn't include waiting queues just yet, just to see how this goes, but we could simply create them.
After finishing a story development cycle the story is approved and then it is buffered on the release box on the top of the board.
Conclusions
These are still very recent changes, we in fact will star to use them next week to see how it goes. Cesare D'Amico stated on twitter "Beautiful! But is it as clear as columns?" this is in fact a big issue when we change the board structure, a think it could be less clear then simple columns, but it should be more fun and thats the idea. When we use spreadsheets, the raw data are clearly visible with columns and rows, but if we represent them in a different way such as a graph we can read more details and patterns than looking the rows and columns individually.
I should soon post the results of our experience, and for now I would appreciate any kind of feedback from the community, preferring the opposing ones, or any new idea.
