Guide
Planning poker for distributed teams
The mechanics of planning poker — simultaneous reveal, group convergence, outlier-driven discussion — are independent of whether the team is in one room or spread across ten timezones. The risks change. The practice does not.
What actually changes
Three things shift when planning poker moves to a video call:
Hesitation signals get harder to read
Co-located, you can see the team member who voted a 13 hesitate when everyone else voted 5. Remote, that hesitation is gated behind a small video tile and only visible if cameras are on. The fix is non-optional cameras and a facilitator who calls out outlier voters by name ("Marie, you voted 13 — what made you pick that?").
Side-channel chat replaces room dynamics
In-room, the discussion is one conversation. Remote, half the team is in the Slack huddle while the other half is in the meeting chat. The estimate suffers because the gap-surfacing mechanic only works when everyone hears every signal. Close side-channels for the session.
Time-zone fatigue accumulates
A 90-minute estimation session at 10pm in Berlin and 6am in San Francisco is not a session; it is one team estimating and the other rubber-stamping. Either rotate the time or split refinement (async, written) from estimation (sync, short).
Practical setup
- Refinement first, sync. Send out the backlog 24 hours ahead so people read it, ask questions, and surface obvious clarifications in writing. Estimation then takes 30 minutes instead of 90.
- Tool sharing a single room link. Everyone clicks one URL. No screen-share required — each browser sees the same state.
- Cameras on, mics off-default. Cameras for the hesitation signal, push-to-talk on mics so the side conversation does not become the conversation.
- Time-box per ticket. Two to three minutes per ticket including discussion. If a ticket consistently overshoots, it is not ready and you split or postpone.
- Facilitator names outliers. Not the loudest voter — the outlier voter speaks first.
Tool features that matter remote
Most planning-poker tools work fine in-person; the differences show up under remote conditions. The features that matter:
- Simultaneous reveal. Anything less re-introduces anchoring.
- No account required. Friction kills attendance — pin a link, get votes.
- Real-time sync. Two seconds of lag between vote and reveal is enough for someone to ask "wait, who voted what?".
- Spectator mode. Stakeholders watch without voting; the team estimates without political pressure.
- CSV export. Estimates flow back into the issue tracker after the session.
Ace-The-Backlog ships all five — see the tool or the complete planning poker guide for context.
FAQ
- Do remote teams estimate less accurately than co-located teams?
- Not inherently. The accuracy of planning poker comes from simultaneous reveal and forced discussion of outliers — both of which work identically over a video call. The risks are different (distraction, side-channel chat) but the core mechanic is preserved.
- Should cameras be on during remote planning poker?
- Yes, where social norms allow. Faces carry the hesitation signals that surface when someone is unsure or unconvinced — those signals are what triggers the productive follow-up question. Without faces, the session collapses to a number-collection exercise.
- How do we handle timezones?
- Two options work: rotate the meeting time so the same people are not always inconvenienced, or split refinement (async, written) from estimation (sync, time-boxed). The second option is harder to set up but cuts the live session to 20 to 30 minutes.
- Can planning poker work fully async?
- Partially. Estimation works async — collect votes over 24 hours, surface outliers afterwards. The discussion step is what fails: written follow-ups take a day per round, and the value of planning poker is the discussion. Hybrid (async vote, sync discuss) is the practical compromise.
Run your next remote session
Share one room link in Slack, Teams, or Zoom. No installs, no logins, no plugins.
Start a planning poker session