Free Sample Size Calculator
How Many Survey Responses Do You Need?
Find the minimum number of survey responses needed for statistically significant results - set your confidence level, margin of error, and population size.
Leave blank if unknown or over 100,000
Minimum Responses Needed
at 95% confidence · ±5% margin of error
Standard precision - appropriate for most business decisions.
Invitations needed by response rate
Sample Size Reference Table · Infinite population
| Margin of Error | 90% Confidence | 95% Confidence | 99% Confidence |
|---|---|---|---|
| ±1% | 6,766 | 9,604 | 16,590 |
| ±2% | 1,692 | 2,401 | 4,148 |
| ±3% | 752 | 1,068 | 1,844 |
| ±5% | 271 | 385 | 664 |
| ±10% | 68 | 97 | 166 |
What is Sample Size and Why Does It Matter?
Sample size is the number of survey responses you need to draw statistically valid conclusions about your broader population. Without a sufficient sample, your results are unreliable - a lucky or unlucky batch of responses can lead to completely wrong conclusions.
For customer feedback surveys, sample size determines whether your NPS, CSAT, or CES score reflects reality or just noise. A sample that is too small means your confidence intervals are too wide to act on - your "results" could swing 20 points in either direction by chance.
Two factors control how accurate your results are: confidence level (how certain you want to be that your results are repeatable) and margin of error (how much variation you're willing to accept). Tightening either one requires more responses.
Sample Size Formula
There are two formulas - one for large or unknown populations, one with a finite population correction:
Infinite / Large Population
n = (Z² × p × (1−p)) / e²
Finite Population Correction
n_adj = n / (1 + (n−1) / N)
Variable Key
Choosing Confidence Level and Margin of Error
There is no universally "correct" setting - the right choice depends on the stakes of your decision and your ability to collect responses.
95% confidence · ±5% margin
Recommended for most surveysThe industry standard. Requires ~385 responses (infinite population). Suitable for NPS, CSAT, and product feedback surveys where you are making meaningful business decisions.
95% confidence · ±10% margin
Directional / early-stageOnly ~97 responses needed. Good for quick temperature checks or when you cannot reach large samples. Do not use for benchmarking or high-stakes decisions.
99% confidence · ±3% margin
High-stakes researchRequires ~1,842 responses. Use when making major product investments, pricing changes, or publishing results publicly. Expensive to collect but highly reliable.
90% confidence · ±5% margin
Budget-consciousOnly ~271 responses needed. Acceptable for internal tracking where some uncertainty is tolerable. Avoid for external reporting or competitive benchmarking.
Frequently Asked Questions
How many responses do I need for a reliable NPS survey?
At 95% confidence with ±5% margin of error, you need a minimum of 385 responses. For ±10% (directional tracking), you only need 97. Most SaaS companies target 200–400 NPS responses per measurement cycle, which gives reliable results with a ±5% margin at 95% confidence.
Does my population size matter if I have thousands of customers?
For populations over 20,000, the population size has almost no practical effect on required sample size. The finite population correction only meaningfully reduces your sample size when your population is small - under 5,000. If you have 10,000+ customers, you can use the "infinite population" formula.
What is the difference between margin of error and confidence level?
Confidence level is how certain you are that your results are repeatable - 95% means that if you ran the same survey 100 times, 95 results would fall within your margin of error. Margin of error is the acceptable range of variation around your result. At 95% confidence with ±5% margin of error, your true NPS is within 5 points of your measured NPS 95% of the time.
What if I cannot reach the minimum sample size?
Collect what you can and report wider confidence intervals. 50 responses at 95% confidence have a ±14% margin of error - directional but not precise. Be transparent about sample size in your reporting. As you collect more responses over time, your cumulative data becomes more reliable even if individual cycles are small.
Should I include all customers or use a random sample?
For accurate results, use a random or representative sample of your customer base. Surveying only active users, recent signups, or unhappy customers creates selection bias that distorts your score. If you cannot survey everyone, randomize who receives the survey to avoid systematic skew.