November 9, 2009

Playing Planning Poker at Perforce


As humans, we're better at relative estimates than absolute ones. As a project manager, I make schedules based on estimates. (I consider an estimate to be an informed guess.) I need to estimate many aspects of a project:

  • planning staffing, scheduling & budget
  • how many stories can fit into one release?
  • how many stories can be completed in a sprint?
  • know how much is left to do

It turns out that "Planning Poker" provides a terrific model for estimating tasks, generating useful estimates with a minimum of effort. It's not a casino game, but rather a planning tool for software projects.

In Planning Poker, each team member is  given a set of non-linear cards. For example, 1, 2, 3, 5, 8, 13, 20. The reason behind these cards being non-sequential is that it's a gentle reminder that precision isn't critical.  The objective is to emphasis relative comparison, so the numbers 1, 2 and 3 might indicate that the tasks are fairly well understood and do-able within the sprint time frame.  A number like 13 indicates it's uncertain and the number 20 says that this is too big and needs to be split up into different, smaller tasks. In a moderated process, the team is briefed on the task, then plays a series of rounds in which consensus is reached on an estimate for the task.

For example, if you are designing a new website to track lost cell phones, the team decides on the tasks required to create the website. Tasks are then prioritized. For example, logging onto the website with your phone number is high-importance, so it's a 1. Seeing a photo of phone, by comparison, might be a 13. Finally, the team plays Planning Poker to estimate how much work can be accomplished in a cycle.

Planning Poker is definitely a fun way to plan, get everyone involved and have a bit of excitement at work.  As my almost 80-year-old dad reminded me recently, "Rachael, scrum master, but still a pig as your bacon is on the line."