How To Play Planning Poker Right
Planning Poker is an estimation technique often used in agile software development. All Scrum teams I’ve worked with or met at gatherings know planning poker and use it on their projects. So I assume it is very widely spread.
However, often enough I’ve seen that Planning Poker isn’t played the right way. There are certain rules for playing Planning Poker that need to be respected in order to get the full benefit of playing it. Team members reveal their estimation to each other before everyone has made an estimate for themselves. In case Story Points are used, there is often a debate about Story Points interfering the actual estimation and exchange of information in Planning Poker. This is just two of the things that could harm Planning Poker.
Before you start playing Planning Poker, ensure that everyone knows the rules and understands it’s purpose. In case Story Points are used as your estimation unit, ensure everyone has a good understanding of it. For more information on Story Points see my previous post on this topic. Also consider setting one or two reference estimates at the beginning of the project, so that everyone gets a similar understanding of the estimation unit used.
–
The origin of Planning Poker is the Wideband Delphi method, which is derived from the Delphi method.
The Delphi method is a forecasting method in which a panel of experts answers a questionnaire anonymously in several rounds. A facilitator then provides anonymous feedback about the results of the previous round to the participants and another round of answering is held. This process is repeated several times until the forecasts of all experts reach a consensus. The Delphi method was developed in the 1950s at the RAND Corporation.
From the Delphi method the so called Wideband Delphi method was derived. The Wideband Delphi method follows the same process as the Delphi method. Additionally experts are allowed to discuss their estimation with each other. Experts discuss differences in their estimations in several rounds and by doing so reach a consensus on estimates.
–
Let’s now look at how Planning Poker works, which is developed based on the Wideband Delphi method.
As well as the Delphi or the Wideband Delphi method, Planning Poker is a consensus-based estimation technique. Before you can actually play Planning Poker, all your work items or User Stories need to be prepared and introduced to the team. The team then asks questions for clarification until everyone has fully understood the requirements.
All participants get a card deck handed out. The cards show values like 0, 1, 2, 3, 5, 8, 13, 20, 40 and 100 if you are using Fibonacci number Story Points range. Each participant then makes an estimation and chooses one card without letting the group know which one it is yet. When all participants have their individual estimation, they reveal their cards all at once. The reason for this is, to catch each participants point of view without being influenced by other estimations or thoughts.
In the event that everyone shows the same number in the first round already, that would be the estimate. Usually this doesn’t happen and estimates vary. If there are different estimates after the first estimation round, this is the starting point for exchanging thoughts and ideas about what led to these estimates.
Let’s assume you have estimates ranging from 3 to 8 Story Points and the majority of participants estimate a 5. You may ask the participants that estimated 8 Story Points questions such as “What makes this work item an 8 for you?” or “What is the main effort you see that this User Story is estimated with 8?”. The difference between 3, 5 or 8 Story Points are then discussed. Some participants might have chosen an 8 because they were faced with uncertainty or wanted to add an extra buffer. There might also be technical issues which were not considered by the ones who estimated a 3 or 5.
Discussions should be limited to a fix timebox in order to keep Planning Poker efficient. After the given timebox has ended, the next round of estimation may take place. Again everyone should make their estimates individually without letting others know what the estimation is. Once they have finished their individual estimation, all participants may reveal their estimation at once again. If there are still differences in estimation, the same procedure as described above takes place. If estimates are the same, you can move on to the next work item or User Story.
Thats basically how Planning Poker works. Once consensus is reached, you have a single indicator as an estimation for your User Story. As mentioned Planning Poker is often played with Story Points but other estimation units may be used as well.
–
A few thoughts on the maturity of User Stories or work items that are estimated. You have read about setting a fix timebox for discussions. The purpose of this is to see whether discussions come to an end in reasonable time. If this is not the case, there might be too much uncertainty, open questions or lack of understanding of the work item. This is a sign that the User Story or work item is not ready for estimation yet and needs to be refined. Also limit the estimation rounds to a maximum of 5. If no consensus is reached after 5 rounds, this is also an indicator that the User Stories or work items are not ready for estimation. They need to be refined and become clearer to the participants before estimation.
Following are some of the strengths and benefits of Planning Poker at a glance:
- consensus in estimation is reached within short time
- structured exchange of thoughts and ideas establishes common understanding of work items
- Planning Poker is an effective way to get estimates
- Planning Poker reveals whether your User Stories or work items are ready or need further refinement
Reads:
Tools:
- Agile Sizing Cards – Planning Poker Like – Perfect For Estimating/Sizing Work!
- Top Quality Agile Poker Estimation Planning Cards
Sources & Links:
Make, Nice summary. The point is not the estimate – it’s the conversation to drive discovery.
Thank you, Jay!