How Many Survey Responses Do You Actually Need?

"How many people do we need to survey?" is one of the first questions in any research project — and one of the most misunderstood. Businesses often either survey far too few people to trust the results, or chase a huge number that drains budget for no extra benefit. The honest answer is that the right sample size is a calculation, not a guess.

This guide explains the three things that actually determine how many responses you need, so you can collect enough for reliable, defensible results without over-investing.

Sample size depends on three things

Forget the idea of a fixed target. The number of responses you need is driven by three inputs working together:

• Population size — how many people are in the group you want to understand (your customers, visitors, or a market segment).
• Confidence level — how sure you want to be that your results reflect the true population, usually expressed as a percentage such as 95%.
• Margin of error — how much wiggle room you accept around each result; a smaller margin of error means a tighter, more precise answer and requires more responses.

Together these define precision. A higher confidence level and a smaller margin of error make your findings more reliable — but both push the required sample size up. The skill is choosing a level of precision that matches the weight of the decision.

What confidence level and margin of error really mean

Imagine 60% of your survey respondents say they're satisfied. With a margin of error of plus or minus 4 points, the true figure across your whole audience is very likely between 56% and 64%. The confidence level — say 95% — tells you how often that range would hold true if you repeated the study. The two numbers always travel together: a result is only meaningful when you know the range around it and how confident you are in that range.

A single percentage with no margin of error attached isn't a finding — it's a guess wearing a number.

Why bigger isn't always better

Precision improves quickly as you add early responses, then flattens out. Going from 50 to 400 responses sharpens your results dramatically; going from 1,000 to 2,000 barely moves the margin of error. Past a certain point you're paying for responses that don't change your decision. That's why a well-scoped study sizes the sample to the precision required — and stops there.

Representativeness matters more than raw numbers

A large sample drawn from the wrong people is worse than a smaller, representative one. If your respondents skew toward one type of customer, age group, or channel, the results describe that subgroup — not your whole audience. Good research designs how respondents are selected so the sample mirrors the population, then segments the findings to reveal the differences that actually drive decisions.

A practical way to decide

1. 1Define the decision the research will inform, and how costly a wrong call would be.
2. 2Pick a confidence level and margin of error that match that stakes level — higher stakes justify tighter precision.
3. 3Calculate the required sample for your population using those inputs.
4. 4Check it against your timeline and budget, and adjust the precision (not the rigour) if you need to.
5. 5Design selection so the sample is representative, not just large.

The bottom line

The right number of survey responses is the number that gives you the confidence and precision your decision deserves — no more, no less. Get the inputs right and you collect reliable, defensible data efficiently. At ILIA, we size and design every study around the decision it serves, so the results hold up to scrutiny without wasting your budget.