How to make a poll people will actually answer
A practical guide to making a poll: choosing the format, writing the question, sharing the link, and reading the result without fooling yourself.
Making a poll takes about thirty seconds. Making one that produces a useful answer takes slightly longer, and almost all of the extra time goes into the question rather than the tool.
This is the whole process, start to finish.
1. Decide what the poll is for
There are really only three reasons to run one, and they call for different questions.
To decide something. You have two real options and you are going to act on the result. This is the strongest use, and the question should name the actual trade-off.
To find out what people think. You are not committing to anything; you want a read on the room. Here you can ask about preferences rather than actions.
To start a conversation. The result matters less than the argument it causes. Aim for the questions where you genuinely cannot guess the split.
If you cannot say which of the three you are doing, the poll will read as aimless and people will scroll past it.
2. Reduce it to two options
Multiple choice feels more thorough and usually is not. Every extra option splits the vote, lowers the winner's share, and gives you a result you have to interpret rather than act on. Four options at 30/28/24/18 tells you nothing.
Two options force you to decide what the real question is before anyone else has to. If you genuinely have five candidates, run several rounds rather than one crowded poll — people answer a two-choice question faster and more honestly.
3. Write the question
The rules that matter most, in order:
- One decision per question. "Should we launch in March with the new pricing?" hides two disagreements inside one result.
- Make the options mutually exclusive. If someone can reasonably want both, your number is noise.
- Put the difference at the front. "Icon left" and "Icon right" beat two options that share their first eight words.
- Do not signal the answer you want. If you would be unhappy with one side winning, you are asking for permission, not an opinion.
There is a longer version of this with examples in how to write a poll question people actually answer.
4. Choose where it lives
The format matters more than people expect.
| Where | Good for | Limits |
|---|---|---|
| Group chat | Fast decisions with people you know | Scrolls away; hard to count |
| Story sticker | Casual engagement with followers | Gone in 24 hours, followers only |
| Survey form | Long, detailed research | Heavy; low completion for one question |
| A poll on its own page | One question, any audience, a link you can put anywhere | Needs a link to travel |
For a single decision you want an answer to, a poll with its own link is usually the right shape: it works in a group chat, a story, an email or a document, people can answer without signing up, and the result stays there afterwards instead of vanishing.
5. Give it a deadline
An open-ended poll gets a slow trickle and no moment. A poll that closes tonight gets a burst, a result and a decision. The deadline is not administrative — it is what converts opinions into an outcome, and it is the single most reliable way to increase the number of responses.
6. Share it where the answer actually is
Post it where the people who care already are, not everywhere. A poll about your team's sprint belongs in the team channel; a poll about which cover to use belongs in front of readers. Twenty responses from the right people beat two hundred from strangers.
If you are sharing to more than one place, use a separate link per channel so you can see which audience actually turned up.
7. Read the result honestly
Three traps.
Small numbers look precise. With 20 votes, a 55/45 split is a coin toss wearing a suit. Treat anything inside about ten points as "no clear preference" until the numbers are much larger.
Your audience is not the population. People who follow you, or who are in your team channel, are not a random sample of anyone. A poll tells you what your audience thinks. That is genuinely useful, and it is not market research.
A poll is an input, not a verdict. You are allowed to go against the result. Say so, and say why — people respect a decision far more than a number, and it teaches them their vote informs you rather than replaces you.
8. Close the loop
Say what you decided. This is the step almost everyone skips, and it is the one that determines whether anyone votes next time. The reason people take part is to find out what happened; if nothing visibly happens, they stop.
Make one now
Pick the decision you have been going back and forth on, write it as two options, and put it to a vote. It is free, it takes about thirty seconds, and the people you send it to can answer without making an account.
If you are stuck for a question, the this-or-that list and the would-you-rather list are both sorted by who you are asking.