Free tool
Survey sample size calculator
The short answer
For a large population, a survey needs about 385 completed responses for a 5% margin of error at 95% confidence. A small population needs fewer: from a list of 5,000 customers, about 357. Divide by your expected response rate to see how many people to invite.
On this page 6 sections
Survey sample size
For example, your customer list
50 is the safe default
Responses needed
357
Without population limit
385
People to invite
1,785
Leave the expected proportion at 50% unless you have a good prior estimate; it gives the most conservative sample size.
Runs in your browser. Nothing you enter is sent to us; the page address keeps your numbers so you can share a result.
The formula
The calculator uses Cochran's formula, n0 = z² x p x (1 - p) / e², where z is 1.645, 1.96 or 2.576 for 90, 95 or 99% confidence, p is the expected proportion and e the margin of error. For a known population N it applies the finite population correction, n = n0 / (1 + (n0 - 1) / N).
Choosing the inputs
- Margin of error: 5% is common for customer surveys; 3% for decisions where a few points matter.
- Confidence level: 95% is the usual default.
- Expected proportion: 50% gives the largest, safest sample. Use a lower figure only if you have good evidence, for example from a previous survey.
- Response rate: be realistic. Email surveys to customers often get well under 30%; in-product surveys more.
What sample size cannot fix
The margin of error only covers random sampling error. If the people who answer differ from those who do not (the happiest and angriest customers reply most), a bigger sample just measures the bias more precisely. Send the survey to a random or complete list, keep it short, and send one reminder.
Analysing an NPS survey? The NPS calculator shows the margin of error for the score itself, which is wider than for a simple percentage.
Questions
- How many survey responses do I need?
- About 385 for a 5% margin of error at 95% confidence when the population is large. For a population of 1,000 it falls to about 278, and for 500 to about 218.
- Does a bigger population need a much bigger sample?
- No. Above a few tens of thousands, the required sample barely changes: about 385 responses give a 5% margin at 95% confidence whether the population is 50,000 or 50 million.
Sources
Related guides
How to measure NPS (Net Promoter Score)
How to ask the NPS question, calculate the score, choose when and where to survey, and turn the follow-up comments into decisions.
NPS vs CSAT vs CES: which customer metric to use
NPS, CSAT and CES compared: the question each asks, how to calculate it, and when each one is the right measure of customer experience.
A/B testing for small business: traffic and significance
How to run an A/B test on a small site, how much traffic you need, how to read statistical significance, and when not to test. With a free calculator.