Survey sample size calculator

By the Gather team · Published August 18, 2026 · Last updated August 18, 2026 · Free, no signup, nothing stored

How many responses do you actually need? Enter your population, confidence level, and acceptable margin of error. The math is the standard normal approximation with a finite population correction, so it works for a list of 400 customers as well as for a general population.

How to use this

Population size is everyone you could theoretically survey. For a customer base of 3,000 accounts, enter 3,000. For a general market, enter something large like 100,000; above that the number barely moves.

Confidence level is how sure you want to be that the true value falls inside your margin. 95 percent is the standard in commercial research. 99 percent costs you a lot more responses for a small gain.

Margin of error is how much slop you accept. At 5 percent, a result of 60 percent means the true value is likely between 55 and 65. If a 10 point swing would change your decision, you need a tighter margin.

Expected proportion is your best guess at the split. Use 50 percent unless you have a strong prior, because it produces the largest and therefore safest sample.

The formula

Standard normal approximation, then finite population correction:

n0 = (z^2 x p x (1 - p)) / e^2
n  = n0 / (1 + ((n0 - 1) / N))

z = z-score for the confidence level
p = expected proportion
e = margin of error
N = population size

What this does not tell you

Sample size answers precision, not usefulness. A perfectly powered survey of the wrong audience is still the wrong answer. Three things matter more than the number this tool gives you:

• Who is in the sample. Screening and quotas matter more than raw count. 100 verified in-market buyers beats 1,000 people who clicked an ad.
• Whether you can act on it. If a 5 point difference will not change what you do, you do not need the precision.
• Whether you know why. A number without a reason is a headline, not a decision. That is why our interviews ask the follow-up right after the scale question.

Sample sizes people actually use

More on how we design and field studies on the platform page, or read a sample report.

Frequently asked questions

How many survey responses do I need?

For most commercial decisions, 385 completes gives you 95 percent confidence with a 5 percent margin of error on a large population. That number drops if your population is small: for 1,000 people it falls to about 278.

Is 100 responses enough for a survey?

It depends on the decision. 100 completes gives roughly a 10 percent margin of error at 95 percent confidence, which is fine for directional reads and message ranking, and not fine for anything where a 10 point swing changes your action.

What confidence level should I use?

95 percent unless you have a specific reason otherwise. It is the commercial standard, and moving to 99 percent nearly doubles your required sample for a small gain in certainty.

Does population size matter?

Only when it is small. Above roughly 20,000 the required sample barely changes. Below a few thousand the finite population correction reduces your requirement noticeably.

What is margin of error?

The range around your result where the true value probably sits. A 5 percent margin on a 60 percent result means the true value is likely between 55 and 65 percent.

Is this calculator free?

Yes. No signup, nothing is stored, and the calculation runs entirely in your browser.