Estimation Consensus Meter
Enter each teammate's estimate after a planning-poker vote and instantly see the spread, median, mode, an agreement percentage, and a clear re-vote-recommended signal. Works offline.
Built by the AbraCalc team
How to play
- Type each teammate's estimate into the entry rows; add more rows with + Add Estimate if your team is larger than three.
- Watch the distribution bars, spread, median, and mode update live as you enter or edit numbers.
- Set the re-vote threshold to the spread you're willing to accept without discussion (defaults to 5).
- Read the agreement percentage and the accept/re-vote signal below it to decide whether to move forward or run another round.
Estimation Consensus Meter takes the raw numbers from any estimation round — a planning poker vote, a show of hands, a spreadsheet of story points — and turns them into a clear read on whether the team actually agrees. Type in each estimate (a name is optional, the number is what matters), and watch a live distribution chart, the min, max, spread, median, and mode update instantly as you type. An agreement percentage scores how tightly the estimates cluster, from 100% when everyone lands on the same number down toward 0% as the spread widens, and a configurable re-vote threshold turns that into a plain accept-or-re-vote recommendation. Non-numeric or blank entries are simply ignored, so a stray '?' or empty box never throws off the math.
Frequently asked questions
- How is this different from Planning Poker?
- Planning Poker runs the full estimation game itself: choosing a card deck, hiding picks until everyone's in, and revealing together. This tool has no card flow or hidden state at all — it's a standalone analysis meter you can use after ANY estimation method, not just planning poker, to see the spread, median, mode, and an agreement percentage with a re-vote signal.
- How is the agreement percentage calculated?
- Agreement is 100% when every estimate is identical (zero spread) and decreases as the spread between the highest and lowest estimate widens relative to the group's median, so a spread of 2 among estimates clustered around 5 scores much higher agreement than the same spread of 2 among estimates clustered around 1.
- What exactly triggers the re-vote recommendation?
- The re-vote signal fires when the spread (max estimate minus min estimate) is strictly greater than the threshold you've set, which defaults to 5; a spread exactly equal to the threshold is treated as still acceptable, only spread that exceeds it recommends discussing outliers and re-voting.