π Agile estimation, explained
What Is Planning Poker? The Complete Guide
Planning poker is a gamified, consensus-based way for agile teams to estimate story points. This guide covers how it works, the rules, the cards, and how to run a session β then lets you try it free.
Planning poker in one sentence
Planning poker is an estimation technique where each member of an agile team privately picks a card representing the effort of a user story, then everyone reveals at the same time and discusses the differences until they reach consensus.
It was popularised in the agile community in the early 2000s and remains one of the most widely used estimation methods because it is fast, collaborative, and surprisingly accurate.
How planning poker works, step by step
The Product Owner or facilitator reads out a user story and answers clarifying questions. Each member of the development team then selects a card face-down. On a count, all cards are revealed simultaneously. If estimates cluster, the team records the agreed number. If they diverge, the highest and lowest voters explain their reasoning and the team re-votes. This repeats until consensus emerges.
Why reveal simultaneously?
Revealing at once prevents anchoring bias β the well-documented effect where the first number mentioned skews everyone else's judgement. Private, simultaneous voting keeps each estimate independent.
Why discuss the outliers?
A high or low outlier usually means someone sees a risk or a shortcut others missed. The conversation, not the number, is where most of the value lives.
Who takes part?
The development team (engineers, QA, UX) does the voting because they do the work. The Product Owner clarifies scope and acceptance criteria but typically does not vote. The Scrum Master facilitates, keeps the session timeboxed, and protects the simultaneous-reveal rule.
What are you actually estimating?
Planning poker estimates relative effort, not hours. A story worth 8 points is roughly twice the effort of a 4-point story, regardless of who picks it up. Story points bundle complexity, uncertainty, and volume into a single relative number β which is why teams forecast more reliably with points than with raw time estimates.
Frequently asked questions
βΈIs planning poker only for Scrum?
No. It works for any agile team that estimates work relatively, including Kanban teams sizing backlog items before pulling them into flow.
βΈWhat numbers are on the cards?
Most teams use the Fibonacci sequence (1, 2, 3, 5, 8, 13, 21β¦). The widening gaps reflect that larger items are inherently harder to size precisely.
βΈHow long should a session take?
Keep it timeboxed. Most refinement sessions estimate a backlog in 30β45 minutes; spend the time discussing, not debating single points.
βΈCan we play planning poker remotely?
Yes. Online tools like Planning Poker Hero let distributed teams vote in a shared room in real time, free and with no signup.
βΈWhat if the team can't agree?
If two re-votes don't converge, split the story, park it for more discovery, or take the higher estimate and move on. Perfect precision isn't the goal.
Related guides
Try planning poker with your team
Reading about it only goes so far. Open a free room and run a real estimation round.