Survey Sample Size: How to Calculate It (+ Free Calculator)

Survey sample size is the number of completed responses you collect, and it determines how closely your results reflect the whole group you care about. Too small, and a handful of unusual answers can swing your numbers; too large, and you burn time and budget for precision you don’t need. This guide explains how to calculate the right sample size, what each input actually means, and how to turn “responses needed” into “people to invite” — starting with a calculator you can use right now.

Calculate your survey sample size

Set your confidence level and margin of error, add your population if you know it, and factor in your expected response rate. The calculator returns both the number of completed responses you need and how many people to invite.

If you just want a rule of thumb: for a large or unknown population at 95% confidence and a ±5% margin of error, aim for about 385 completed responses.

What is survey sample size?

Sample size is how many people from your target population actually respond to your survey. You rarely survey everyone — instead you survey a representative subset and use it to draw conclusions about the whole. The sample size controls three things:

• Precision — how tight your margin of error is around each result.
• Reliability of subgroups — whether you can trust a breakdown by segment, region, or plan.
• Signal vs. noise — whether a difference you see (say, 72% vs. 68%) is real or random.

Get it right and you can make confident decisions. Get it wrong and you’re either guessing or over-spending.

The four inputs that decide sample size

Every sample size calculation comes down to four variables. Understanding them makes the calculator above far more useful.

1. Population size (N)

The total number of people you want to learn about — all your customers, all users on a plan, everyone who bought last month. For very large or unknown populations, this barely affects the result. It matters most when the group is small (a few hundred), where you can afford to survey a larger share of it.

2. Confidence level

How sure you want to be that your result falls within the margin of error. 95% is the standard — it means that if you repeated the survey 100 times, about 95 of those results would land within the margin of error of the true value. Higher confidence (99%) needs a bigger sample; lower confidence (90%) needs less.

Each confidence level maps to a z-score:

3. Margin of error

The plus-or-minus range around your result. If 60% of respondents are satisfied with a ±5% margin, the true figure is likely between 55% and 65%. Smaller margins mean more precision — and a larger sample. ±5% is the common default; tighten it only when a decision truly depends on that precision.

4. Response rate

The share of invited people who actually complete the survey. This doesn’t change how many completes you need, but it drives how many invitations you send. Email surveys often see 20–30%; well-targeted in-product or post-purchase surveys can go much higher. Always plan your send list around the expected rate.

The survey sample size formula

Under the hood, the calculator uses Cochran’s formula with a correction for finite populations. First, the base sample size for a large population:

n₀ = (Z² × p × (1 − p)) / e²

Where:

• Z is the z-score for your confidence level (1.96 for 95%)
• p is the estimated proportion — use 0.5, which requires the largest sample and is the safe default when you don’t know the split
• e is your margin of error as a decimal (0.05 for ±5%)

At 95% confidence and ±5% margin, that gives n₀ = (1.96² × 0.5 × 0.5) / 0.05² ≈ 385.

Then, if you know your population size (N), apply the finite-population correction so you don’t oversample a small group:

n = n₀ / (1 + (n₀ − 1) / N)

Finally, to account for non-responders, divide by your response rate:

invitations = n / response rate

Sample size reference table

Here’s how many completed responses you need at the most common setting — 95% confidence, ±5% margin of error — for different population sizes. Notice how the number levels off as the population grows.

The key takeaway: beyond a few thousand people, chasing a bigger population barely moves the requirement. Once you can collect ~385 solid responses, you’re in good shape for most decisions.

How to choose your confidence level and margin of error

There’s no universally “correct” setting — it depends on the stakes of the decision:

• High-stakes, board-level or financial decisions — 95–99% confidence and a ±3% margin. You want precision you can defend.
• Everyday CX and product decisions — 95% confidence and ±5% margin is the sweet spot of rigor and effort.
• Directional, exploratory reads — 90% confidence and ±10% margin is fine when you just need a sense of the trend and will follow up later.

Remember: tightening the margin from ±5% to ±3% roughly triples the responses you need. Spend that effort only where the decision justifies it.

Probability vs. non-probability sampling

Sample size is only half the story — how you select respondents (sampling method) determines whether your sample is representative in the first place.

• Probability sampling uses random or structured selection (simple random, stratified, cluster) so every member of the population has a known chance of being included. It’s the gold standard for generalizing to the whole population.
• Non-probability sampling (convenience, quota, snowball) is faster and cheaper but riskier — you reach whoever is available, which can bias results. It’s fine for exploratory work, pilots, and quick reads.

A large sample drawn with a biased method is still biased. If you need to generalize confidently, invest in representative sampling, not just volume. For context on defining who you’re actually studying, see our guide to the population of interest.

What if you can’t hit the target sample size?

It happens — niche audiences, low-traffic pages, or hard-to-reach segments. You still have good options:

1. Report the real margin of error. With fewer responses, your margin widens; state it honestly rather than overclaiming precision.
2. Lean on qualitative signal. Open-ended answers provide rich context even when the numbers aren’t statistically bulletproof.
3. Use recurring pulse surveys. Repeated smaller reads over time can reveal trends that a single undersized survey can’t.
4. Raise your response rate. Better timing, shorter surveys, incentives, and the right channel (email, SMS, WhatsApp, in-app) all lift completion — which shrinks the gap without needing a bigger send list.

Below-target data isn’t worthless; it’s just less precise. Treat it as directional and validate the big calls with more responses.

Common sample size mistakes to avoid

• Confusing sends with completes. You need 385 completed responses, not 385 invitations. Always divide by your response rate.
• Ignoring subgroups. A sample that’s fine overall can be far too small once you split it by segment. If subgroup analysis matters, size each subgroup, not just the total.
• Chasing precision you don’t need. Going from ±5% to ±1% multiplies the effort many times over for decisions that rarely require it.
• Big sample, biased method. Volume doesn’t fix a non-representative sample. Sampling method comes first.
• Forgetting to clean the data. Speeders, duplicates, and incompletes eat into your effective sample size. Plan for some attrition.

How Responsly helps you hit your sample size

Knowing the number is step one — collecting the responses is where a good survey platform pays off. With Responsly you can:

• Reach people on the right channel — email, SMS, WhatsApp, website, in-app, and QR — to lift response rates and reach your target faster.
• Track completions in real time so you know exactly when you’ve hit your required sample size.
• Filter and segment responses in analysis to check whether each subgroup has enough data to trust.
• Prevent duplicate and low-quality submissions so your effective sample size stays clean.

Once your responses are in, our guides on how to analyze survey data and cross-tabulation help you turn them into decisions — and survey response bias covers the pitfalls that no sample size can fix.

Conclusion

Survey sample size isn’t about collecting as many responses as possible — it’s about collecting enough to make confident decisions without wasting effort. Set a 95% confidence level and a ±5% margin of error, use the calculator to get your number, then plan your send list around your expected response rate. Pair the right sample size with a representative sampling method and clean data, and your survey results will hold up when it’s time to act on them.

Ready to collect the responses you need? Create a free Responsly account and launch your survey across every channel today.

FAQ