What I Like about Poker Planning

Poker planning is one scheme often used in Scrum or other Agile methodologies as a way of helping to estimate tasks in a sprint.

The way I used it was to get the sprint team together (the people who will do work in the sprint), for them to come up with a proposed list of user stories and tasks they are thinking of undertaking (this is typically done before the poker planning meeting, possibly individually or in small teams), then you get out the deck of cards and sit around a table. The cards have numbers on them for legal estimates. E.g. ½, 1, 2, 3, 5, 8, 13… You must always round your estimate up, never down.

Someone then describes the next task to be estimated for say a minute. People can ask questions, but you try not to spend too long here. The first time you do this, it’s not uncommon for the meeting to run over time as people learn more about what others are doing. In later sprints, more knowledge is present so less time is needed to provide the background.

After getting the task description, everyone picks a card and plays it all at the same time. This is a good time to be a little mischievous. This is what makes it fun.

If the estimates are all close, you quickly move on to the next task. There is no need to discuss further. If there is any discrepancy, you ask some of the individuals why they picked their estimates. Often the tasks get rethought and re-estimated at this time.

At the end, you have a list of estimations (I like ½ day, 1 day, 2 day resolutions in particular). Long tasks get broken into smaller ones (I dislike tasks more than 2 days long).

I like poker planning because…

• The social side makes the planning process more fun.
• It encourages sharing of understanding of what other team members are doing.
• During standup meetings, people understand what other team members are doing better, reducing the length of the daily status reports.
• It shares the responsibility of estimation amongst team members.
• Team members can learn from each other as to what they look for in terms of estimates.
• It helps catch things that were missed. Often one team member thinks of something others did not. These are valuable catches.
• You spend time discussing estimates that are controversial, moving on from easy ones quickly.
• Everyone is involved and engaged.
• Playing cards all at once avoids the people adjusting their estimates based on what other people say.
• It helps reinforce that teams are responsible for delivery of sprint tasks, not individuals. If someone gets stuck or is behind, other team members can help more rapidly (they understand more up front), and the team is acknowledged for reaching goals (not only individuals).

I really suggest giving it a go if you have not tried it. At worst it can be fun. At best it shares the responsibility of estimation, and can really help improve the reliability of estimation. At least that was my experience with it.