Feature

Planning Poker

Anonymous story point estimation with 6 deck types, instant statistical analysis, and zero setup. The fastest way to estimate with your team.

How it works

  1. 1

    Create

    Pick Planning Poker and choose a deck. Share the room link.

  2. 2

    Vote

    Everyone picks a card anonymously. Change your vote until reveal.

  3. 3

    Reveal

    Host reveals all votes. See stats instantly.

  4. 4

    Discuss

    Outliers explain their reasoning. Re-vote if needed.

  5. 5

    Next

    Move to the next story with one click.

6 built-in deck types

Modified Fibonacci

0, 1, 2, 3, 5, 8, 13, 20, 40, 100, ?, Pass

The most popular choice. Gaps grow with uncertainty.

Fibonacci

0, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, ?

Classic Fibonacci sequence for purists.

T-Shirt Sizes

XS, S, M, L, XL, XXL, ?

Great for rough sizing during backlog refinement.

Powers of 2

0, 1, 2, 4, 8, 16, 32, 64, ?

Clean doubling scale for teams that prefer it.

Linear (1-10)

1, 2, 3, 4, 5, 6, 7, 8, 9, 10

Simple numeric range.

Custom Numeric

Your own values

Define any set of numbers or labels.

Instant stats after every reveal

Average and median of all votes
Spread (difference between highest and lowest)
Min / Max values
Closest card to the average
Vote distribution chart
Consensus level: Full, Near, or Divided
Outlier detection with count

Why simultaneous voting beats going around the table

The trouble with estimating out loud is that the first number wins. As soon as somebody says five, nobody else is estimating the story any more. They are reacting to five. Psychologists call it anchoring, and it does not care how experienced your team is: the person who speaks first sets the range, and the two people who would have said thirteen quietly say eight instead.

Planning Poker removes the anchor with one rule. Nobody sees a vote until everybody has voted. In ScrumMastr the cards stay face down while the room fills in, and you can change your mind as often as you like before the host reveals. When the reveal comes, every card flips at once, so the spread you are looking at is what your team actually thought, not what the room agreed to after the loudest voice went first.

Picking a deck your team will not argue about

Modified Fibonacci is the default for a reason. The gaps between the cards grow as the numbers get bigger, which matches how confidence works: the difference between a 1 and a 2 is real and knowable, while the difference between a 40 and a 45 is imaginary. Forcing the team to choose between 40 and 100 stops a debate that was never going to produce a better number.

Use T-shirt sizes when you are sorting a long list quickly and precision would be false comfort, and switch to numbers once the work is close enough to plan. Powers of 2 suits teams who think in doubling. A custom deck is there for the team that has its own scale and no interest in changing it. Pick the deck when you create the room, tell everyone which one you picked, and leave it alone for the rest of the session. Changing scales mid-session makes every earlier estimate incomparable.

Reading the reveal

After every reveal ScrumMastr shows the average, the median, the spread between highest and lowest, the full distribution, and a consensus level of Full, Near, or Divided. It also flags outliers and counts them. The number worth watching is the spread, not the average. A room that voted 5, 5, 5, 8 has agreed and can move on. A room that voted 3, 3, 13, 20 has not, and averaging that to 10 would bury the disagreement instead of resolving it.

When the vote is divided, ask the highest and the lowest voter to explain their thinking, in that order, and keep it to a minute each. Nine times out of ten they are not disagreeing about effort. They are describing different stories: one of them included the migration and the other did not. Once that surfaces, re-vote. One re-vote is usually enough, and if a second one does not converge, the story is the problem rather than the estimate.

Keeping the session inside the hour you booked

Give each story about two minutes. Read it, vote, reveal, and move on if the team agrees. Discussion is for divided votes only, which in a healthy backlog is maybe one story in four. Teams that discuss every story regardless of the spread are the ones whose planning sessions run to ninety minutes.

Some stories will not converge no matter how long you talk. That is useful information: the story is too big, too vague, or depends on something nobody has looked at yet. Split it, write a spike for the unknown, and put it back in the backlog rather than pushing for a number the team does not believe. When you do land on a number, put it into whatever tracker you actually use. ScrumMastr is the room where the estimate happens, not another place to keep your backlog.

Keep reading

Try Planning Poker with your team. Free, no signup.

Create a room