Planning Poker / Delphi Estimating

Planning poker is a consensus estimate generated through multiple rounds of synchronised ‘reveals’. It’s great for thrashing out detail and shared understanding.

Pros/Cons:

  • Multiple people estimate a work item at the same time, avoiding 1 expert ‘leading’ the others
  • It highlights assumptions and misinterpretations
  • It encourages group discussion, which increases the accuracy of an estimate
  • It takes a long time to estimate a large backlog

Tools

  • 1 set of estimation units in card form (Fibonacci, T-Shirt Sizes) for each team member
  • Additional recommended cards (1 set for each team member):
    • Infinite (Too big to estimate)
    • Question (Unable to estimate with existing information)
    • Red, Amber and Green cards for confidence estimating
  • A number of items to estimate, printed or written on pieces of paper/card

Flow:

  • Someone reads the description of the work item to the team
  • The team consider the complexity of the work and their confidence level. They collect 1 estimate and 1 confidence level card (Red = Low, Amber = Medium, Green = High)
  • When someone feels they have a number they raise their 2 cards in the air to show they are ready to estimate
  • When everyone has their cards in the air, a rock-paper-scissors style 1-2-3 is called and everyone shows their estimate using their hands
  • If the estimate is inconsistent around the room, the highest and lowest estimates are compared and discussed*
  • The process is repeated until a consensus is made
  • The complexity estimate once agreed is written on the work item
  • Choose the next work item and follow the steps again

*If there is inconsistency, for example a ‘3’ and a ‘8’ estimate confidence this is a good cause for discussion, the team member estimating ‘8’ may have more detail on why a work item is more complex due to information that others may not be aware of or know that this would rely on another piece of work being done. Alternatively, the team member estimating ‘3’ may have knowledge of a similar work item that has been solved before and therefore makes this work item easier.

This is especially true when stakeholders estimate differently to team members as this can highlight a misunderstanding in requirements or complexity.