Sample Size Calculator

How to calculate your sample size

1. 1Pick your confidence level (95% is the market research standard).
2. 2Set the margin of error you can accept (±5% for most surveys).
3. 3Enter your population size if it's limited (customers, employees…).
4. 4Read the required number of respondents — then export or share the result.

Why sample size matters

Asking everyone is impossible; asking too few people makes your numbers unverifiable. Sample size is the bridge between the two: the minimum number of respondents that makes your percentages interpretable, with a known margin of error. A 200-respondent survey announcing “62% purchase intent” with no margin of error says almost nothing; the same figure at ±3.5% becomes a basis for decisions.

The good news: the required size barely depends on the size of your market. For a very large population, 385 respondents are enough for ±5% at 95% confidence — whether you target the whole US or a single state. It's counter-intuitive, and it's exactly what the formula below makes visible.

The formula used (and its limits)

This calculator applies the classic proportion formula: n₀ = z²·p·(1−p)/e², where z depends on the confidence level (1.645 for 90%, 1.96 for 95%, 2.576 for 99%), p is the expected proportion and e the margin of error. When your population is limited (a customer base of 1,000 people, say), the finite population correction lowers the required sample: n = n₀ / (1 + (n₀−1)/N).

The formula assumes simple random sampling and a binary or proportional question. It corrects neither recruitment bias, nor non-response, nor badly worded questions — sample size is a necessary condition for reliability, never a sufficient one. Also plan a safety buffer: if you expect 20% incomplete answers, recruit accordingly more.

p = 50%: the cautious choice

If you don't know the expected proportion, leave 50%. It's mathematically the worst case: p·(1−p) peaks at 0.5, so the required sample is the largest. Any other real value will make your results more precise than promised — never less. It's the convention used by virtually every research institute.

Frequently asked questions

What is the sample size for a standard survey?

385 respondents for a ±5% margin of error at 95% confidence, on a large population. That's this calculator's default setting, and the standard for most quantitative studies.

Does the size of my market change the result?

Barely, once the population exceeds ~100,000 people. Below that, the finite population correction reduces the sample: for 1,000 people, 278 respondents are enough instead of 385 (at 95%, ±5%).

Which margin of error should I choose?

±5% for everyday decisions (concept tests, satisfaction), ±3% for finer trade-offs, ±1–2% for critical measurements. Beware: precision is expensive — halving the margin quadruples the sample.

90%, 95% or 99% confidence: which one?

95% is the standard compromise. 90% is fine for quick explorations; 99% is warranted when an error would be very costly (health, regulatory), at the price of a much larger sample.

Does this calculation work for an A/B test?

The principle is close, but an A/B test compares two groups: you need the calculated size in EACH branch, and ideally a dedicated statistical power calculation accounting for the minimum detectable effect.

Does a large sample guarantee a reliable survey?

No. Size controls sampling error, not bias: a poorly recruited panel, leading questions or unbalanced response rates skew results regardless of size. Size is necessary, not sufficient.