Know how many responses you actually need. Enter your population, confidence level and margin of error, and this returns the sample size for results you can trust — before you send the survey.
What this free tool is great for: a quick, one-off job with no signup — it runs in your browser and there's nothing to manage.
Its honest limit: you can compare up to three scenarios during this visit, but they are not kept after you leave, they won't update as your real numbers change, and this tool does not connect to your live accounts.
A free tool is fine for a one-off. If survey sample size calculator is a weekly job, these take it off your plate.
Best for teams whose long survey forms get few answers
Sounds like you? “Long survey forms put people off, and too few of them finish.”
Plan that fits: Basic, $19 a month on yearly billing (no monthly option); European visitors see €19
Best for WordPress site owners building contact and lead forms
Sounds like you? “My WordPress site needs contact and lead forms, and I can't code them.”
Plan that fits: Basic, $49.50 for the first year (1 site), renews at $99; WPForms Lite free
Best for small marketing teams with web forms already live
Sounds like you? “Leads fill in my forms, but I can't see what they read first, and I have no budget for a big tool.”
Plan that fits: Free plan; it stores 200 lead profiles
Best for small EU businesses collecting form data
Sounds like you? “Our customers' form answers must stay in the EU, and I'm tired of tools that bill per response.”
Plan that fits: Pre-paid, €4 per form bought in packs of 5, excl. VAT
Picked from our review dossiers: who each tool suits, their problem in their words, and the plan that fits. “Try” buttons are affiliate links: we may earn a commission if you sign up, at no extra cost to you. It never changes our verdict.
The instinct when you want to understand a group is to try to hear from all of them — but you don't need to, and you can't. The whole power of surveying is that a well-chosen sample of a population tells you, within a known margin, what the whole population thinks. A national poll of a few thousand people can predict an election involving millions. This calculator tells you how many responses you actually need for a given confidence and margin of error, which stops you from either wasting effort chasing an impossible census or, worse, drawing bold conclusions from a handful of responses that can't support them.
Four concepts set your sample size. The population is the whole group you care about. The sample is the subset you actually survey. The margin of error is how far your results might be from the truth — the "plus or minus" you see on polls. And the confidence level, usually 95%, is how sure you are that the truth falls within that margin. Together they answer the real question: how many responses do I need so that my results are close enough to reality, often enough, to act on?
Once you see surveying as buying a certain precision at a certain confidence, the sample-size number stops being mysterious.
The relationship between sample size and margin of error is the key insight, and it's not linear. To halve your margin of error, you don't need twice the responses — you need roughly four times as many. So the first few hundred responses buy you a lot of precision, and each additional batch buys less. Getting from a wildly-uncertain result to a decent one is cheap; getting from decent to extremely precise is expensive.
This is why most surveys aim for a margin of error of around three to five percent rather than chasing perfection — beyond that point, you're spending a great deal of effort to shave off a sliver of uncertainty that rarely changes the decision.
Here's what surprises everyone. Above a certain point, the size of your total population has almost no effect on the sample you need. Getting a reliable read on a town of 20,000 and a country of 200 million requires a strikingly similar number of responses — often just a few hundred to a couple of thousand for a small margin of error at 95% confidence. It feels wrong, but the statistics are clear: what matters is the absolute size of your sample, not what fraction of the population it represents.
This is liberating, because it means even huge audiences can be understood with a manageable, affordable number of responses.
The number this calculator gives you is completed responses — and that's very different from how many people you must invite. Most surveys get a response rate well below half, often far lower, so to end up with your target number of responses you have to send to many times that number. Plan for it: if you need several hundred responses and expect a modest response rate, your invitation list has to be several thousand.
Forgetting this is why so many surveys stall short of a usable sample. Always work backwards from your required responses through a realistic response rate to the size of the audience you actually need to reach.
Hitting your sample size is necessary but not sufficient, because who doesn't respond can quietly poison your results. If the people who ignore your survey differ systematically from those who answer — happier customers respond while frustrated ones don't, or vice versa — your sample is biased no matter how large it is. This is non-response bias, and it's more dangerous than a small sample because it's invisible in the numbers.
A big, biased sample gives you a confident, precise, wrong answer. Guarding against it — through good survey design, follow-ups, and a healthy suspicion of who's actually replying — often matters more than squeezing out extra responses.
Flowing from that, a smaller sample that genuinely represents your population is worth more than a huge one that doesn't. A thousand responses all from your most engaged power users tell you about power users, not your whole customer base. The goal isn't just quantity; it's a sample whose mix — of demographics, tenure, usage, whatever matters — mirrors the population you're trying to understand. When you can, aim for a spread that reflects reality rather than whoever happens to be easiest to reach.
A well-balanced few hundred beats a lopsided few thousand every time, because representativeness, not raw count, is what lets you generalise.
Not every question needs statistical rigour. If you're doing qualitative research — trying to understand why people feel something, discover unexpected problems, or gather rich feedback rather than measure a precise percentage — a small number of in-depth responses can be genuinely useful. A dozen thoughtful interviews can reveal a usability issue no quantitative survey would catch. The trap is confusing the two: don't quote a percentage from five responses as if it were a measurement, and don't demand a huge sample when you're really after understanding, not numbers. Match the sample size to whether you're measuring how much or exploring why.
It depends on your confidence level and margin of error, not just population size. For 95% confidence and ±5%, you often need only a few hundred responses — this tool gives the exact number.
The range around your result where the true answer likely sits. A ±5% margin on a 60% result means the real figure is probably between 55% and 65%.
Send your survey with a tool like SurveySparrow, which runs engaging, conversational surveys and gathers the answers for you.
The pricing behind our reviews is published as an open dataset, with the sentence behind every cost-trap flag and the date it was researched. Free to use with attribution (CC BY 4.0).
Blogger, teacher or toolmaker? Put this calculator on your own page — free forever, no strings. Copy the snippet below (the credit link is appreciated and keeps the tool free):
The credit link in the snippet points to this tool’s page on AIBuilder Marketplace, our own site. It is not an affiliate link.
This tool is free and runs entirely in your browser.
New dossiers, cost-traps we found, and tools that earned a keep — no hype, no sponsored-disguised-as-advice. Unsubscribe anytime.
