How to Run a Planning Poker Session (Facilitator Checklist)

Planning poker is simple to explain but easy to run badly. This checklist covers what experienced Scrum Masters do before, during and after a session so estimates stay fast and meaningful.

Before the session

• Refine first, estimate second. Stories should have a clear goal and acceptance criteria. If the team spends ten minutes figuring out what a story means, it wasn't ready.
• Pick a scale and stick with it. Fibonacci story points are the most common choice; T-shirt sizes work well for early roadmap sizing. See Fibonacci vs T-shirt sizing.
• Agree on reference stories. Keep two or three finished stories as anchors ("this was a 3", "this was an 8") so new estimates stay consistent sprint after sprint.
• Prepare the room. For remote teams, create the online room ahead of time and drop the link in the calendar invite.

During the session

1. Read the story aloud and give the team a minute for questions.
2. Vote silently. Ask people not to say numbers out loud. Anchoring is the most common way planning poker goes wrong.
3. Reveal at the same time. The facilitator reveals once everyone has voted.
4. Ask the extremes to speak first. "Sam, you said 2, what makes it small? Alex, you said 13, what worries you?"
5. Re-vote once or twice. If the team still disagrees after two rounds, take the higher estimate or split the story. Don't argue for ten minutes over a 5 versus an 8.
6. Timebox each story to around five minutes. Park anything that needs research.

Handling common situations

Someone always votes "?"

That's useful information: the story or the domain isn't clear to them. Pair them with someone on the next similar task, or schedule a short spike.

The team converges instantly on every story

Great, but double-check that people aren't just copying each other. A quick "what would make this bigger?" question can reveal hidden risk.

Stakeholders want to attend

Invite them as observers. In Planno, observers see every vote and the results but don't vote, so they can follow along without skewing the numbers.

Common mistakes

• Converting points to hours. Story points are relative; once they become hours, the benefits of relative estimation disappear.
• Letting the facilitator or tech lead vote first or out loud.
• Estimating stories that are obviously too big instead of splitting them.
• Averaging away a real disagreement. A spread of 2 to 13 needs a conversation, not a 7.5.

Run your next session

Planno gives you private voting, a simultaneous reveal, the average and agreement level, and an observer mode, all free and without sign-up. Start a session and share the link with your team.