What is planning poker?
A short, practical guide to agile estimation with story points — and why simultaneous votes beat loud first opinions.
The short version
Planning poker (also called Scrum poker) is a consensus-based way for a team to estimate the relative size of work items. Everyone privately chooses a card, then reveals together. Large spreads spark discussion; the team re-votes or agrees on a number. The goal is a shared understanding of complexity and risk — not a perfect prediction of hours.
Story Point Poker is a free online room for that workflow: invite with a code, vote in realtime, reveal when the host is ready.
Why teams use it
- Less anchoring — if the most senior engineer says “this is a three” first, others often copy. Silent votes reduce that pressure.
- Surfaces unknowns — a split between 3 and 13 usually means the story is unclear or risks differ by specialty.
- Shared language — over time the team builds a mental catalog of “what an 8 feels like” for their codebase.
- Better backlog conversations — estimation meetings become refinement: acceptance criteria, edge cases, and dependencies come up early.
Story points (not hours)
Most planning poker decks use a modified Fibonacci sequence: 0, 1, 2, 3, 5, 8, 13, 21, and so on. The gaps grow on purpose. Humans are bad at distinguishing 14 from 15, but better at “about twice as hard as that other story.”
Story points usually mix effort, complexity, and uncertainty. A tiny config change might be a 1. A feature that touches several services with unclear requirements might be a 13 — or “?” until the team spikes it. Converting points to hours is optional and often misleading across people and contexts; velocity (points finished per sprint) is a team metric for planning capacity, not individual performance grading.
A typical workflow
- Pick a backlog item and ensure everyone understands the goal.
- Clarify questions; split the item if it is obviously too large.
- Everyone selects a card without showing others.
- Reveal together and discuss high and low estimates.
- Re-vote if needed, then record the agreed size.
- Move to the next item.
Online tools replace physical cards so remote and hybrid teams can run the same flow. In Story Point Poker the host controls reveal and the next round; see the how-to guide for product steps.
Common pitfalls
Estimating without a reference
Without a few known “this was a 5” stories, numbers drift. Keep a short reference list for the team.
Debating before the first reveal
Long pre-vote monologues reintroduce anchoring. Vote first, talk about the spread second.
Punishing “wrong” estimates
If people fear being the outlier, they stop bringing useful signal. Treat spreads as information about the work, not a test.
Pointing vague stories
If half the room votes “?”, improve the story or schedule a spike instead of forcing a number for the board.
When planning poker is a poor fit
Pure research with no deliverable shape, interrupt-driven ops work, or single-person tasks may not need a full poker session. Use judgment: the workflow costs meeting time, so reserve it for items where shared understanding matters.
Try it
Ready to run a session? Create a room, share the code, and estimate together. Questions about the product are on the FAQ.