The first thing I tried was to use retrospectives to find real pain points in the team... but not everyone agreed on what the highest priority items were. Then I tried using positive reinforcement of goals that I'd chosen. I wish I could say we chose these goals but there was no common consensus, so these goals were usually the collaborative output of Olivier and I, or the output of several team members... but for simplicity's sake I'm saying that I worked for the following. Please note that most of these changes were introduced to the team one at a time, and allowed to become habitual before moving on to the next item:
- more pair-swapping
- shorter story cards
- avoidance of black holes
- lighter-weight story card estimation
- emphasis on TDD rather than Test-After
- up-front face-to-face discussion with the client for every story card
- introduction of fixed-cost story cards
- focus on fewer changes so we can do those few well
- Agile Thought of the Day
None of this really created a gelled team, in fact much of it was probably just a distraction from our real problem (at least from my current perspective). It probably wasn't a waste, since these changes have become a valued part of our daily work, and have also helped to establish a level of trust in the work I'm doing. Still, this more effective group work didn't translate into a performing team. Why?
I felt like if I could get everyone to see and focus on one single problem simultaneously, that would become the shared vision that defines a performing team. I concluded that I needed to construct an unambiguous definition of success... which, again, I pushed for. I believe it would have been more effective to have the team define this success, but I couldn't get the group to agree on one goal. So, I said that an iteration was not a success if we hadn't moved all the story cards over to the DONE column by the end of the iteration. I shared this vision consistently and regularly at stand-ups, my turn as story-observer, at retrospectives, and with my pair. Slowly the team adopted this vision... slowly means it took 10 weeks! Since then, however, there's been clear consensus on whether an iteration is done or not... yet that didn't make a gelled, performing team! Now what?
I needed more information, more ideas. So I started doing one-on-one interviews with every person (12 total) in the team, every week. That's been effective at understanding how people feel, and in making other types of changes (like finishing our iteration before 8pm Friday nights... now we're done Thursday afternoon or Friday morning). So with these interviews, I discovered I wasn't alone in feeling like we weren't a gelled team... people wanted closer cooperation, stronger teamwork, but we just didn't know how to get there.
As I said before, XP Day France was a great opportunity for me. I now know we need to get through forming, and I figured that if I shorten the cadence, we'd learn quicker, we'd have more pressure, and communicate more... all setting up fertile ground for team forming and agile growth. So I suggested we try the Pomodoro technique (as a team) last Thursday. Afterwards, we decided that the practice was too complicated to evaluate in just one day, and Olivier decided we should try it for an iteration. Well, that iteration began this Monday, and few people actually read the free 40-page book (hooray--some people are giving it an honest shot), but for the rest I've been using the first long pause of the day to explain the practice in more detail.
Everyone uses the same, 25 minute timebox. After each pomodoro, we pause, then have a mini-standup, where the sole objective is to find out how the team is going to attack the remaining story cards. If someone is blocked, we might send help to that pair; if someone has something interesting to share, they do; otherwise, we may resume the task we were on before. The object is to efficiently select the most important tasks that remain.
The breaks are ad-hoc, and may involve a team-building game, drawing on the white board, or a walk outside. This has become self-organizing already--different people propose games or set up the timers. Just to note, we play games that stimulate the creative, left-brain, so that when people go to sit back down they're more likely to see creative alternatives in the new pomodoro.
On a larger scale, the objective is to add so much discipline that we start to storm.
I'm hoping that before this month is over, our team will just be able to sense when a pair goes dark, will know when to swap pairs, will know how to focus on the core need inside a story, and will hum like a hive, producing a continuous flow of honey (err, business value).