Agile Planning - Planning Poker

"Estimations" is a buzz word in the software world. Every one are looking at the best ways to estimate their taks/stories/features. But is there some tried and tested technique to arrive at the magic number?


Yes there is one atleast in the agile world, and as a Agile coach I have extensively used this to demostrate a more pragmatic way of arriving at estimations. Here is a sneak preview of the same:

So how to get to the magic number ?????

Prerequisites:

Stories:

• Stories have to be analyzed and approved by business (SME)
• Stories must have clear assumptions and specific acceptance criteria

Estimation numbers:

• Lets take 1, 2, 4, 8 as estimation numbers
• 1 being the least and 8 being the maximum
• If any of the story is estimated 8 then it means
• The story is so large that it  cannot be played in one single iteration/sprint
• The story must be further broken down into smaller ones

The Game:

• The facilitator picks up one story from the story bucket and reads the story details loud to the team
• The story details must be functional and NOT tilt or lead towards solution’ing
• The team then discusses the tasks that are required to meet the acceptance criteria defined in the story
• During this time the team can clarify any queries related to the story with the facilitator/BA
• Its important to promote healthy discussion within the team during this time as it yields in better estimation
• Facilitator then asks the team if they are ready for the estimation
• If Yes, on a count of 3 the team starts to estimate
• This can be done in 2 ways:
• Use poker cards, so that each team member will show one card that he/she feels is the right estimate of the story
• I like to use "Shout out loud" technique, just use your fingers to point out the estimation(1,2,4,8)
• If the entire team agrees on a number then there is consensus and the team moves to the next story
• If the estimates differ, the high and low estimators defend their estimates to the rest of the team
• A healthy debate must be encouraged but the debate must not yield into chaos
• The debate must be time boxed
• After the discussion, the story is re-estimated like earlier unless a consensus is reached
• Repeat the same for all the stories