Research appendix
Full Data: State of Planning Poker 2026
This page contains the complete anonymized dataset behind the State of Planning Poker 2026 report. All findings in the main post are derived from these numbers. Publishers and researchers are welcome to reference the data with a link back to the original report.
Sessions
2,465
Filtered planning poker sessions.
Teams
208
Distinct real ScrumJam teams.
Votes
103,343
Individual votes in PostHog.
Median duration
41.8 min
Session start to last reveal.
1. Dataset Overview
| Metric | Value | Source |
|---|---|---|
| Total sessions analysed | 2,465 | Firebase |
| Total teams | 208 | Firebase |
| Total reveals / estimates | 14,119 | PostHog |
| Total individual votes | 103,343 | PostHog |
| Data window | Q4 2025 - May 2026 | ScrumJam |
| Session filter | 2+ reveals with 2+ participants | ScrumJam |
2. Day Of Week
Session creation grouped by weekday.
| Day | Sessions | % of total |
|---|---|---|
| Monday | 1,844 | 17.2% |
| Tuesday | 2,350 | 21.9% |
| Wednesday | 2,308 | 21.5% |
| Thursday | 2,660 | 24.8% - peak |
| Friday | 1,308 | 12.2% |
| Saturday | 158 | 1.5% |
| Sunday | 110 | 1.0% |
3. Hour Of Day
Session creation grouped by UTC hour.
| Hour (UTC) | Sessions | % |
|---|---|---|
| 06:00 | 297 | 2.8% |
| 07:00 | 580 | 5.4% |
| 08:00 | 1,073 | 10.0% |
| 09:00 | 1,251 | 11.7% - peak |
| 10:00 | 818 | 7.6% |
| 11:00 | 915 | 8.5% |
| 12:00 | 986 | 9.2% |
| 13:00 | 1,015 | 9.5% |
| 14:00 | 882 | 8.2% |
| 15:00 | 550 | 5.1% |
| 16:00 | 425 | 4.0% |
| 17:00 | 462 | 4.3% |
| Outside 06-18 | ~600 | ~5.6% |
Times are UTC. Most teams appear to be in European timezones, placing the real peak at 10-11 AM local time.
4. Session Duration
| Metric | Value | Source |
|---|---|---|
| Median duration | 41.8 min | Firebase |
| Average duration | 48.3 min | Firebase |
| P25, fastest quarter | ~20 min | Firebase |
| P75, slowest quarter | ~62 min | Firebase |
| Sessions under 15 min | ~12% | Firebase |
| Sessions over 60 min | ~28% | Firebase |
Duration is measured from session start to last reveal. Sessions shorter than 1 minute or longer than 4 hours are excluded.
5. Tasks Estimated Per Session
| Tasks per session | Sessions | % |
|---|---|---|
| 2-3 tasks | 858 | 33.6% |
| 4-6 tasks | 961 | 37.6% - most common |
| 7-10 tasks | 524 | 20.5% |
| 11-15 tasks | 132 | 5.2% |
| 16+ tasks | 80 | 3.1% |
| Average tasks per session | 5.4 | |
| Median tasks per session | 5 |
6. Team Size Analysis
| Team size | Sessions | Median duration | Tasks/hour | Any perfect consensus |
|---|---|---|---|---|
| 2-3 people | 89 | 20.8 min | 51.5 | 51.7% |
| 4-5 people | 423 | 38.0 min | 21.5 | 56.3% |
| 6-7 people | 536 | 38.1 min | 20.8 | 40.5% |
| 8-10 people | 281 | 44.8 min | 19.0 | 31.0% |
| 11+ people | 100 | 46.5 min | 15.3 | 25.0% |
Tasks/hour is estimation throughput regardless of point values. Perfect consensus means a session had at least one round where all participants voted identically.
7. Vote Distribution
Based on 103,343 individual votes across all sessions.
| Card | Votes | % |
|---|---|---|
| 3 | 31,073 | 30.1% - most common |
| 2 | 23,660 | 22.9% |
| 5 | 21,097 | 20.4% |
| 1 | 11,160 | 10.8% |
| 8 | 7,544 | 7.3% |
| 13 | 2,148 | 2.1% |
| 0.5 | 686 | 0.7% |
| 21 | 682 | 0.7% |
| 4 | 658 | 0.6% |
| Other / T-shirt | ~4,635 | ~4.4% |
Cards 1-5 make up 84.2% of all votes. Cards 8+ make up 10.7%. Average individual vote is 4.73; median individual vote is 3.
8. Highest Estimate Per Session
Highest average vote recorded in any single round, per session.
| Highest estimate | Sessions | % |
|---|---|---|
| 1-2 pts | 80 | 21.9% |
| 3 pts | 137 | 37.5% |
| 5 pts | 90 | 24.7% |
| 8 pts | 31 | 8.5% |
| 13 pts | 8 | 2.2% |
| 21+ pts | 19 | 5.2% |
84.1% of sessions never produce an estimate above 5 points. Only 15.9% include an estimate of 8 or above.
9. Deck Usage
| Deck | % of sessions | Source |
|---|---|---|
| Fibonacci Sequence | 84.5% | Firebase |
| Fibonacci Flat Sequence | 4.9% | Firebase |
| T-Shirt Sizes | 2.3% | Firebase |
| Simple Numbers | 2.1% | Firebase |
| Hours Sequence | 0.6% | Firebase |
| Doubled Sequence | 0.5% | Firebase |
| Custom decks | ~5.1% | Firebase |
Fibonacci variants combined: 89.4%. Notable custom decks included half-point linear sequences, large-scale sequences, Yes/No decks, and novelty sequences such as 1, 2, 3, 5, 8, >9k.
10. Sprint Cadence
Among teams that have run three or more sessions.
| Sessions per month | % |
|---|---|
| 5+/month | 36.1% |
| 3-4/month | 39.6% - most common |
| 2-3/month | 24.3% |
75.7% of active teams run three or more sessions per month.
11. Consensus And Alignment
| Metric | Value | Source | Notes |
|---|---|---|---|
| Weighted consensus score | 71 / 100 | Firebase | Step-distance based, not binary |
| Perfect consensus rate | 15.6% | PostHog | All votes identical |
| Teams with measurable outlier voter | 31.9% | PostHog | 2+ point gap between two team members |
| Banana rains per session | 0.88 | Firebase | Rounds where all votes were identical |
Consensus score is a 0-100 weighted score based on Fibonacci step distance. Two votes one step apart, such as 8 and 13, score 85/100.
12. Session Overview
Firebase aggregate fields used in the report.
| Metric | Value |
|---|---|
| Average reveals per session | 5.4 |
| Average resets per session | 4.21 |
| Average banana rains per session | 0.88 |
| Median session duration | 41.8 min |
| Average session duration | 48.3 min |
Data Quality And Filters
Sessions included if:
- At least 2 estimation rounds with 2+ participants voting.
- Session duration between 1 minute and 4 hours.
- Associated with a real team, not anonymous single-user tests.
- Not from ScrumJam internal test accounts.
Data sources:
- Firebase: session-level aggregates, deck usage, duration, consensus, and reveal counts.
- PostHog: event-level votes, participant counts, timestamps, and outlier analysis.
- Where sources overlap, Firebase is primary because it captures all sessions.
How To Cite
ScrumJam, "State of Planning Poker 2026", scrumjam.app/blog/state-of-planning-poker-2026.