Results
Reading poll results without fooling yourself
A live tally looks like an answer. Here is what it actually tells you — and the four ways it misleads.
The sample is the result
Twenty votes from one group chat describe that group chat. They do not describe the company, the class or the customer base. Before quoting a percentage, note who could realistically have voted and how many of them actually did.
Small samples move a lot
With a handful of votes, one person changing their mind swings the leader. Early percentages are noise; watch the shape of the result near the deadline, and treat a two-vote lead as a tie.
One person can be several votes
Polls can allow multiple votes per person, and anonymous polls cannot tell a returning voter from a new one. If a result looks implausible, check the rule you set before you suspect the voters.
Hidden results change behaviour
Showing results from the start lets people vote with the crowd; hiding them until the poll ends reduces that effect but also reduces the sense of momentum. Neither is wrong — just know which one you chose.
Export before you discuss
Download the results as PDF, Word, Excel or CSV before the meeting. An export freezes the numbers at a point in time, which ends the “it said 40% five minutes ago” argument, and gives you something to attach to the notes.
Ready to try it?
Create a poll