Usage data
Planning Poker in numbers
How agile teams actually estimate, measured on the sessions run with Planning Poker Online.
Period: March 11, 2026 to October 10, 202681 sessionsUpdated October 11, 2026
Key findings
Between March 11, 2026 and October 10, 2026, teams ran 81 sessions on Planning Poker Online, where 160 participants estimated 281 user stories. What the votes show:
- Team size: 9 participants per session at the median, 7.9 on average.
- First-round consensus: every voter picks the same card on 26.2% of first votes, and 59.5% stay within one step of the Fibonacci scale, such as 3 and 5.
- Spread of first votes: a median gap of 1 between the lowest and highest card, counted in Fibonacci steps.
- Voting rounds: 1.17 per story on average, and 84.3% of stories are settled in the first round.
- Cards played: 3 is the most played card; the ? card makes up 0.4% of votes and the coffee card 0.9%.
Every figure is an anonymous aggregate of the sessions run on the app; the sample and method behind each one are detailed below.
- sessions created
- 81
- participants
- 160
- stories estimated
- 281
- first-round consensus
- 26.2%
Sessions, participants and stories
- sessions created
- 81
- participants
- 160
- stories estimated
- 281
Volume over the period, the base of every ratio below. A session counts when it is created, a participant counts once even when joining several sessions, and a story counts once its estimate is validated.
Method: Sessions created, distinct participants and stories estimated over the period.
Sample: 81 sessions, March 11, 2026 to October 10, 2026
How many people estimate together
- participants at the median
- 9
- participants on average
- 7.9
Number of people who joined each session that revealed at least one vote, all roles included: Scrum Master, voters and observers.
Method: Distinct participants per session, over the sessions with at least one revealed round in the period.
Sample: 30 sessions, March 11, 2026 to October 10, 2026
Consensus on the first vote
- of first votes are unanimous
- 26.2%
- stay within one step
- 59.5%
Share of stories whose first vote puts every voter on the same card, then the share where the lowest and highest cards are neighbours on the scale, such as 3 and 5.
Method: First revealed round of each story on the Fibonacci scale, with at least two numeric votes; ? and coffee cards are ignored.
Sample: 237 first votes, March 11, 2026 to October 10, 2026
How far apart first votes are
- median gap, in Fibonacci steps
- 1
Distance between the lowest and highest card on the first vote of a story, counted in steps of the Fibonacci scale: 3 and 5 are one step apart, 3 and 8 two steps.
| Gap | Share of first votes |
|---|---|
| Same card | 26.2% |
| 1 step | 33.3% |
| 2 steps | 29.1% |
| 3 steps | 7.2% |
| 4 steps or more | 4.2% |
Method: Same first votes as above; the gap is the position of the highest card minus the position of the lowest.
Sample: 237 first votes, March 11, 2026 to October 10, 2026
Voting rounds per story
- rounds per story on average
- 1.17
- settled in a single round
- 84.3%
Number of voting rounds a story went through before its estimate was validated, and the share of stories settled at the first round.
Method: Rounds started for each story estimated in the period, over the stories voted at least once.
Sample: 281 stories, March 11, 2026 to October 10, 2026
Which cards teams play
- most played card
- 3
- of votes on 13 or 21
- 6.4%
Share of each card among all the votes revealed on the Fibonacci scale, re-votes included.
| Card | Share of votes |
|---|---|
| 0.5 | 1.7% |
| 1 | 9% |
| 2 | 12.5% |
| 3 | 27% Most played |
| 5 | 26.2% |
| 8 | 15.9% |
| 13 | 5.8% |
| 21 | 0.6% |
| ? | 0.4% |
| Coffee break | 0.9% |
Method: Every vote of the rounds revealed in the period on the Fibonacci scale, the app's default scale.
Sample: 1,957 votes, March 11, 2026 to October 10, 2026
The ? and coffee cards
- of votes are a ? card
- 0.4%
- of votes are a coffee card
- 0.9%
Share of the special cards among the same votes: ? means the voter cannot estimate, the coffee card asks for a break.
Method: Same votes as the card distribution above.
Sample: 1,957 votes, March 11, 2026 to October 10, 2026
Methodology
Every figure is computed from all the sessions run on Planning Poker Online over the period. There is no survey and no sampling: every session is included, then reduced to anonymous aggregates. No session name, story title or participant name is used.
- Period: from March 11, 2026 to October 10, 2026 included, in whole days. Figures generated on October 11, 2026.
- Card shares and first votes cover the rounds revealed on the Fibonacci scale, the app's default scale (0.5 to 21, plus ? and coffee).
- First-round consensus: on the first revealed round of a story, every numeric vote is on the same card. Rounds with fewer than two numeric votes are left out.
- A ratio is published only when it rests on at least 30 observations.
- Limits: no account is needed, so one person joining from two devices may be counted twice; sessions that visitors create just to try the tool are counted like the others; the figures describe teams using this tool, not every agile team.
How to cite this data
Planning Poker Online, “Planning Poker in numbers”, data from March 11, 2026 to October 10, 2026, https://www.planning-poker.online/data
Published under the Creative Commons Attribution 4.0 licence: you may reuse these figures with a link to this page.