Results from a Kanban learning game

1 minute read

I work with 6 experienced teams helping a UK government department build their new website. Some of the teams are moving moving from using Scrum to using Kanban instead. In practice, we visualise our work and attempt to manage flow but don't attempt to limit work in progress (WIP).

Limiting WIP is hard. It's not enough for 1 or 2 people to say "we should do this", the entire team needs to understand the benefit.

I setup a session to play "Pass the Pennies", a often used training game for teams starting out with Kanban. More information available:

I gave an informal talk about what WIP represents and why limiting it makes sense. To me, WIP represents commitment to deliver something of value to the client but they cannot realise it yet.

I explained the game and then asked for volunteers. I chose role names of "flippers" and "timers" rather than "workers" and "managers" because this indicates a negative connotation for the former role. There was some reluctance (no-one likes to be the first volunteer...) but we found our 9 people and started.

We played the game over 4 rounds and our findings were:

  • the flippers treated it as a race - it was quick (and difficult to time!)
  • on reflection, the flipper's pushed coins and didn’t pull them. We only realised this after the game finished because it was so quick. I should've been clearer about this from the start.
  • the customer's timing was off (e.g. for the batch size of 1 round, the total time measured was less than the quickest flipper's time). Luckily, we had a back up timer to correct this.
  • the total time fell from 1m 17s to 47s after reducing the batch size from 20 to 5. Some people asked how this was possible given that each flipper's elapsed time had increased. We discussed why this happened.

The last - and best - result though was that those teams who want to use Kanban went back to their boards and started to discuss what WIP limits they should use.

A good result.