What Is Planning Poker?

Planning poker — also called scrum poker, pointing poker, or story point estimation — is a technique agile teams use to estimate the relative effort of user stories, bugs, and technical tasks. Instead of one person assigning a number while everyone else listens, each participant picks an estimate privately, all votes are revealed at once, and the group discusses the differences before agreeing on a final value.

That small change in sequence — vote first, talk second — is the whole point. It turns estimation from a monologue into a conversation, and it surfaces disagreement that a quick verbal poll would hide.

A quick example of a planning poker round

Imagine a team of five estimating the story “Add password reset by email.” Everyone selects a Fibonacci card at the same time. The reveal looks like this:

The 3 and the 13 are the interesting votes. Dana thought the endpoint was done; Priya knew that email delivery and token security were not. That gap is not noise — it is the team discovering that the story is bigger and riskier than it looked. After a two-minute discussion, they re-voted and landed on 8. The estimate matters, but the shared understanding matters more.

How a planning poker round works, step by step

A round follows a repeatable pattern that works in person with physical cards or online with a tool like PlanITPoker:

1. The Product Owner or facilitator presents one backlog item.
2. The team asks clarifying questions until the scope is understood.
3. Each participant selects an estimate card privately.
4. On the facilitator's signal, everyone reveals their card at the same time.
5. If estimates differ widely, the highest and lowest voters explain their reasoning first.
6. The team re-votes until estimates converge, or the facilitator records the agreed value.

Where planning poker comes from

The technique was described by James Grenning in 2002 and popularized by Mike Cohn, whose company later attached the “Planning Poker” name to it. It builds on the Wideband Delphi method: gather independent expert estimates, reveal them, discuss the outliers, and repeat. The core insight from Delphi is that anonymous, simultaneous input produces better group judgments than open discussion, because it prevents the first or loudest voice from setting the anchor.

Why agile teams use it

• It reduces anchoring bias. Private votes and a simultaneous reveal stop the senior engineer's number from dominating the room.
• It surfaces unknowns early. A wide spread is a signal that a story needs refinement before the team commits to it.
• It builds shared ownership. Everyone votes, so the estimate belongs to the team rather than to one lead.
• It keeps estimation fast. Most items resolve in a minute or two; only the contested ones get a longer discussion, which is exactly where the time is worth spending.
• It works remotely. Online rooms replicate the same flow without passing physical cards around a table.

What the numbers actually mean

Planning poker uses story points, which measure relative size — a blend of effort, complexity, and uncertainty — rather than hours. Most teams use a modified Fibonacci deck (1, 2, 3, 5, 8, 13, 21) because the widening gaps reflect the reality that large items are inherently harder to estimate precisely. A story is not “5 hours”; it is “about the same size as that other 5 we did last sprint.” For the full breakdown, read story points explained and Fibonacci estimation.

When to use planning poker — and when not to

Teams most often run planning poker during sprint planning and backlog refinement. It is most valuable when multiple disciplines contribute to delivery and when real uncertainty exists.

It adds little value when work is already broken into fixed hourly tasks, when a single specialist owns an item that needs no team consensus, or for trivial changes where the discussion costs more than the estimate is worth. In those cases, skip the ceremony.

Planning poker vs other estimation methods

See the full comparison of agile estimation techniques for a deeper look at each.

Common mistakes to avoid

• Treating points as hours. The moment someone asks “how many hours is a 5?” the relative scale collapses.
• Averaging the votes. A 3 and a 13 do not average to an 8 — they mean the team disagrees, and the disagreement needs a conversation.
• Letting the loudest person anchor first. Keep votes hidden until everyone has chosen.
• Estimating stories that are not ready. If nobody can describe what “done” looks like, refine the story instead of guessing.

Frequently asked questions

How long should a planning poker session take?

Keep it to 60–90 minutes and timebox each item to a few minutes. If a single story eats ten minutes, park it for refinement rather than letting it stall the whole session.

Who should be in the room?

The people who will do the work: developers, QA, and design, with the Product Owner presenting the items. Managers who will not build the story should observe, not vote.

Do I need special cards or software?

No. You can use physical cards, or a free online tool. PlanITPoker gives you a shared room, Fibonacci and T-shirt decks, and instant vote analytics with no signup required.

Run planning poker online for free

PlanITPoker lets you create a room instantly, share a link or QR code, add stories, vote with Fibonacci or T-shirt decks, and view vote distributions — no account required. For facilitation tips, continue with our planning poker best practices guide.