Skip to main content
Skip to content

Planning Poker: How Scrum Poker Estimation Works, Step by Step

Planning Poker

Planning poker, also called scrum poker or agile poker, is a consensus-based estimation technique that agile teams use to size user stories. Each team member privately picks a card showing their estimate in story points, everyone reveals at the same time, and the team discusses any big differences until they agree. The goal is a shared, well-understood estimate, not a precise prediction.

Planning Poker

What is planning poker?

Planning poker is a technique for estimating the relative size or complexity of work, such as a user story, in story points. Story points are a relative unit: a story of 8 points is considered about twice as big as one of 4. Because the estimate is relative, teams compare stories with each other instead of trying to predict hours. It is used in agile development and especially in Scrum.

A short history of planning poker

Planning poker was described by James Grenning in 2002, and later popularised by Mike Cohn. It borrows its core idea, independent estimates followed by group discussion, from Wideband Delphi, an estimation method with roots in work done at the RAND Corporation in the 1940s and 1950s.

The planning poker cards

Cards usually follow a modified Fibonacci sequence: 0, 1, 2, 3, 5, 8, 13, 20, 40, 100, plus a question mark for “I cannot estimate this” and sometimes a coffee cup for “I need a break”. The gaps get wider as the numbers grow, which reflects that larger items are less certain. You do not need physical cards: many teams use a free online planning poker tool or a plugin for their agile board.

How to run a planning poker session

  1. Prepare the backlog. The product owner brings a short list of stories with acceptance criteria. Good criteria make estimating easier. See how to write acceptance criteria.
  2. Agree a reference story. Choose a small, well-understood story as a baseline, such as “2 points”, and compare everything else against it.
  3. Present a story. The product owner explains it and answers questions. Keep this short.
  4. Everyone picks a card privately. Nobody reveals yet, so one person’s number does not influence the others (this is called anchoring).
  5. Reveal together. All cards are shown at the same moment.
  6. Discuss the outliers. The people with the highest and lowest estimates explain their reasoning. This uncovers hidden assumptions, missing requirements and risks.
  7. Re-estimate. Repeat until the team converges on one value, usually within two or three rounds.
  8. Record the estimate and move on. If a story cannot be agreed, split it or park it for more information.

Who takes part?

The whole delivery team estimates: developers, testers and other people who will do the work. The product owner attends to answer questions but does not usually estimate. The scrum master keeps the session on time. Involving testers matters, because testing effort is part of the real size of a story.

Benefits of planning poker

  • Collaboration. Every team member contributes, so estimates reflect the whole team’s knowledge.
  • Better estimates. Group estimates tend to be more balanced than one person’s guess, and relative sizing avoids false precision.
  • Surfaces misunderstandings early. A large gap between estimates usually means people understood the story differently.
  • Less anchoring and louder-voice bias, because estimates are revealed at the same time.
  • Speed. Sizing in points is quicker than estimating in detailed hours.

When to use planning poker

  • Sprint planning: estimate stories that might go into the next sprint.
  • Backlog refinement: size upcoming stories so the backlog can be prioritised.
  • Release planning: estimate larger features or epics to set realistic goals.
  • Change requests: judge the effort of a late change before agreeing to it.
  • Project kick-off: get a rough size of the main pieces of work.
  • Regular re-estimation: revisit estimates as the team learns more.

Planning poker vs other estimation methods

Method Unit Best for Notes
Planning poker Story points Agile teams sizing stories Relative size, group consensus
Three point estimation Hours or days with a range Tasks where risk and uncertainty matter Gives an expected value and a confidence range. Try the 3 point estimation calculator
Wideband Delphi Hours or days Larger projects, anonymous rounds The ancestor of planning poker
Test estimation techniques Effort in hours or days Planning testing work See software test estimation techniques

Teams often use planning poker for day-to-day sizing and a three point estimate when a release needs a date with a confidence range. If you want an end-to-end guide to estimation, see our estimation resources and tips for agile testing teams.

Common planning poker mistakes

  • Turning story points into hours. Points are relative. Converting them one-to-one defeats the purpose.
  • Letting senior people speak first, which biases everyone else. Reveal together.
  • Long debates. If two rounds do not converge, take the larger estimate, or split the story.
  • Estimating unclear stories. If nobody understands it, refine it first.
  • Leaving testers out, which hides testing effort.
  • Using points to compare teams. Points mean different things to different teams.

Planning poker FAQ

What is planning poker?

An agile estimation technique where team members privately choose cards showing their story point estimates, reveal them together and discuss differences until they agree.

What is the difference between planning poker and scrum poker?

There is none. They are two names for the same technique, and agile poker is a third.

Why does planning poker use the Fibonacci sequence?

The growing gaps reflect that larger items carry more uncertainty, so the team is not pretending to tell 20 from 21.

How many people should take part?

Typically the whole delivery team, often five to nine people. Larger groups slow the discussion.

How long should a planning poker session take?

Aim for a few minutes per story. If a story takes much longer, it is probably not ready or is too large and should be split.

Can planning poker be done remotely?

Yes. Online planning poker tools and agile board plugins let distributed teams reveal estimates at the same time.

Is planning poker accurate?

It produces shared and reasonably consistent estimates, but not exact forecasts. Treat the output as a size to plan with, and track the team’s velocity over time.

What is the difference between planning poker and three point estimation?

Planning poker sizes stories relatively with a group. Three point estimation estimates one task with best, most likely and worst cases to get an expected value and a range, as in our 3 point estimation calculator.

Want more testing guides in your Google results? Add Testmetry as a preferred source.

Share it :

Leave a Reply