Sample size: how many responses do you actually need

How many responses does a poll actually need? There’s no single number, and anyone who gives you one is guessing. It hangs on two things: how much the decision matters, and how tight the race is between your options. A casual call might be settled by a few dozen votes. A high-stakes one can demand hundreds or thousands. The real skill is knowing the trade-offs and staying honest about what your sample can and can’t tell you.

Why does sample size matter?

Sample size drives how much you can trust the result. A tiny sample can miss the broader mood entirely; a bigger one gives you firmer footing. But size alone isn’t the whole game. Quality and representativeness count just as much. A huge sample that doesn’t reflect the people you care about will still hand you skewed answers, just with more confidence than it deserves.

How many responses do you need for different types of decisions?

For casual decisions, like picking a lunch spot or choosing a movie, a small sample is plenty. A few dozen responses give you the general lean of the group. The stakes are low and a wrong call costs almost nothing, so there’s no reason to chase statistical significance.

For important decisions, like a new product feature or a marketing direction, you want more. Hundreds of responses buy you the confidence to commit. When the outcome actually matters, a bigger sample is what lets you stop second-guessing.

For critical decisions, like policy changes or major investments, go bigger still, often into the thousands. The consequences are real, so you need to be as sure as you reasonably can. A large, representative sample is what keeps you from betting big on shaky data.

What is the margin of error, and why does it matter?

Margin of error tells you how much wiggle room sits around your number. Say 60% of respondents favor option A with a margin of error of ±5%. The true figure likely lands somewhere between 55% and 65%. Tighter margin, sharper result. It’s the difference between “roughly this” and “pretty much exactly this.”

Generally, the bigger the sample, the smaller the margin, because more responses mean less uncertainty. But sample size isn’t the only lever. How varied the population is and the confidence level you pick, commonly 95%, both feed into it too.

How can you improve your poll’s representativeness?

Size gets the headlines, but representativeness is what keeps you honest. A representative sample actually mirrors the group you’re studying. A few ways to get closer:

• Diversify your distribution channels: Share your poll across multiple platforms (e.g., social media, email, website) to reach a broader audience.
• Target specific demographics: Use demographic filters to ensure your sample includes the groups you’re interested in.
• Avoid bias: Be mindful of the language and framing of your questions to prevent leading respondents toward a particular answer.
• Use random sampling: If possible, employ random sampling techniques to ensure every member of the population has an equal chance of being selected.

Never forget the catch: a big sample that isn’t representative still lies to you. For more on building polls that hold up, see our guide on creating a poll.

Frequently asked questions

How do I know if my sample size is large enough?

It depends on how much the decision matters and how tight a margin of error you can live with. Low-stakes calls are fine on a few dozen responses. Critical ones need enough to hit a margin you’re actually comfortable with, which often means hundreds or thousands.

Can a small sample size ever be valid?

Yes. For a low-stakes decision, or when the population is small and clearly defined, a small sample holds up fine. It’s the broad, high-stakes questions where you really need the numbers to back you up.

What if I can’t get a large sample size?

Then put your energy into representativeness. Target your distribution, spread across channels, and keep the questions unbiased. A small sample carries a wider margin of error, sure, but a representative one can still tell you something genuinely useful.

The bottom line

Picking a sample size is more judgment call than formula. You’re balancing the stakes, the margin of error you’ll accept, and the plain reality of how many responses you can actually gather. And it’s never just the count. Quality and representativeness matter every bit as much. Keep those in view and your decisions rest on firmer ground.

OXPoll’s whole stance is to be straight with you about what online polls can’t do. We work to make polling easy, but we’ll also nudge you to keep one eye on your sample size and how representative it really is. For more on building polls that actually work, head to our create a poll page.