Planning poker for remote teams
Planning poker is a simple way for a team to estimate work together. Everyone votes in private, then all the votes are revealed at once. Done well, it works just as well online as in a meeting room.
What planning poker is
The technique was described by James Grenning in 2002 and popularised by Mike Cohn. Each team member picks a card showing their estimate for a story, usually a story point value from a Fibonacci-style scale. When everyone has chosen, the cards are turned over together and the team discusses any differences.
Why votes stay hidden until the reveal
If people see each other's numbers as they vote, the first estimate given tends to pull everyone towards it. That is called anchoring, and it means a confident voice can decide the outcome without the team's real views surfacing. Voting privately and revealing together keeps every opinion independent, which is the point of the exercise.
Before the session
- Prepare the stories. Each one should have a clear description and acceptance criteria. Vague stories waste session time.
- Invite the people who'll do the work. Developers, testers and designers all see different risks.
- Agree a scale and a reference story. For example, a Fibonacci scale with a known story as a "3".
- Time-box it. Around an hour is plenty. Stop before people get tired, because estimate quality drops.
Running it online
- Start a call. Use voice or video alongside the estimation tool, so the discussion stays natural.
- Share the room link. In Estimate US, the owner creates a room and sends the link. Team members join with a name, and nobody needs an account.
- Start a vote. The owner enters the story title and starts the round. Participants pick a card.
- Reveal. When everyone has voted, the owner reveals the votes and sees who chose what.
- Discuss and re-vote. Talk through any spread, then re-vote if the estimates were far apart.
- Save the estimate. The owner records the agreed value, and it goes into the estimate table for the room.
When people disagree
A wide spread is useful information, not a problem. Ask the person with the highest estimate and the person with the lowest to explain their reasoning. Often one of them knows about a dependency, a tricky integration or a simpler approach that the rest haven't considered.
Limit yourselves to two or three rounds. If the team still can't converge, take the higher number, split the story into smaller pieces, or agree a short investigation first.
Tips for remote teams
- Don't let seniors or the product owner speak first. Let the votes come first.
- Treat the "?" card as a signal that the story needs more information, and pause to clarify it.
- Write down key decisions and assumptions in the ticket while they're fresh.
- Keep cameras on if possible. Facial reactions to a reveal are part of what makes the exercise work.
- Rotate who leads the session, so it doesn't rest on one person.
Try it with your team
Estimate US is a free planning poker tool. Create a room, share the link and start voting. No sign-up needed.
Create a room