How NPS is calculated
Net Promoter Score is the percentage of promoters minus the percentage of detractors. Everyone answers one question, usually "How likely are you to recommend us to a friend or colleague?", on a scale from 0 to 10, and each answer falls into one of three groups.
- Promoters answer 9 or 10.
- Passives answer 7 or 8. They count toward the total but don't add or subtract.
- Detractors answer 0 to 6.
NPS = % promoters - % detractorsExample: 200 people answer. 120 are promoters (60%), 50 are passives (25%), and 30 are detractors (15%). The score is 60 - 15 = 45. Scores run from -100, when everyone is a detractor, to 100, when everyone is a promoter.
What the margin of error tells you
A score from a sample is an estimate. The margin of error is how far the true score could plausibly sit from the one you measured, at 95% confidence. The calculator works it out from the share of promoters (P), the share of detractors (D), and the number of responses (n):
margin of error = 1.96 × √((P + D - (P - D)²) ÷ n) × 100In the example above that comes to about 10 points, so a score of 45 from 200 answers is consistent with anything from roughly 35 to 55. Collecting four times as many answers roughly halves the margin.
Is a change between two periods real?
Scores move from month to month even when nothing changed, because each month is a different sample. The Compare two periods tab checks whether a change is bigger than that noise. It combines the standard error of each period and treats a change as real when it is larger than 1.96 times the combined figure, a standard two-sided test at 95% confidence.
The default example goes from 23 to 45, with 200 answers in each period. The change of 22 points is larger than the combined margin of about 15 points, so it is unlikely to be chance.
The test assumes the two groups of respondents are independent and were reached the same way. If you changed who you ask, or when, the comparison tells you less.
Getting NPS answers you can trust
A precise calculation can't fix a biased sample. A few habits make scores comparable from one period to the next:
- Ask at the same moment each time, such as a set number of days after sign-up.
- Keep the question wording identical between periods.
- Ask inside your product, where active users already are, instead of relying only on email.
- Add one open follow-up question, so each score comes with a reason.
- Cap how often anyone sees the survey, so the same people aren't asked again and again.
Last reviewed