How to calculate NPS (Net Promoter Score)

By Thisys team · Last reviewed 7 October 2026

Net Promoter Score turns one question into a single number from −100 to +100. The arithmetic takes a minute. Reading it well takes a little longer: knowing what the number hides, how far it can move by chance, and what to ask next to it.

The question

NPS comes from one question, answered on an 11-point scale from 0 to 10:

“How likely is it that you would recommend [organisation] to a friend or colleague?”

Label the ends (0 = Not at all likely, 10 = Extremely likely) and show every point. Keep the wording and the scale the same every time you ask, or this quarter's score cannot be compared with last quarter's.

The three groups

Promoters answer 9 or 10.
Passives answer 7 or 8.
Detractors answer 0 to 6.

Passives count in the total number of answers but in neither group, so they pull the score towards zero without adding to it or taking away from it.

The formula

NPS = % promoters − % detractors

Both percentages are of everyone who answered the question.

The result is reported as a whole number, not a percentage: a score of 32, not 32%. It runs from −100, when everyone is a detractor, to +100, when everyone is a promoter.

A worked example

Worked example: 200 customers answer the question

110 answer 9 or 10, so they are promoters: 110 ÷ 200 = 55%.
50 answer 7 or 8, so they are passives.
40 answer 0 to 6, so they are detractors: 40 ÷ 200 = 20%.
NPS = 55 − 20 = +35.

The passives still matter. If 20 of those 50 passives had been detractors instead, detractors would be 60 ÷ 200 = 30%, and the score would fall to +25.

How far can it move by chance?

An NPS from a sample has a margin of error, and it is wider than most people expect, because the score is the difference between two shares. For a simple random sample, the usual approximation at 95% confidence is:

Margin of error ≈ 1.96 × √[(p + d − (p − d)²) ÷ n] × 100

p is the share of promoters and d the share of detractors, both as decimals, and n is the number who answered.

Worked example: The margin for the example above

p = 0.55, d = 0.20 and n = 200.
p + d − (p − d)² = 0.75 − 0.1225 = 0.6275.
0.6275 ÷ 200 = 0.0031375, and its square root is about 0.056.
1.96 × 0.056 × 100 ≈ 11 points.
So +35 from 200 people is consistent with a true score anywhere from about +24 to +46.

That is why a five-point move between two small samples is usually noise. Look at the number of answers before you look at the score, and compare like with like.

Mistakes that make NPS misleading

Averaging the 0 to 10 answers. An average is a different measure; NPS counts groups.
Changing the scale. A 1 to 10 or 1 to 5 scale is not NPS, and the groups no longer line up.
Comparing your score with another industry's or another country's as if they were the same. Expectations and answering styles differ; your own trend is the more useful comparison.
Asking for the number alone. Without a follow-up question you know the score but not the reason.
Reporting the score without the base. +40 from 25 people and +40 from 2,500 people are not the same finding.
Asking only customers you expect to be happy, or nudging people towards a 10. Both inflate the score and hide the problems it exists to find.

What to ask next to it

Follow the score with one open question, such as “What is the main reason for your score?”. If you will act on the answers, ask whether you may contact the person about them. The score tells you where you stand; the reasons tell you what to change.

Use this template

Each opens its own page, where you can preview every question before you use it.

Use cases: NPS surveys and Customer satisfaction (CSAT).

Sources and notes

The method: Fred Reichheld, “The One Number You Need to Grow”, Harvard Business Review, December 2003.
Net Promoter, NPS and Net Promoter Score are trademarks of Bain & Company, Inc., NICE Systems, Inc. and Fred Reichheld.

Put it into practice. Free to start, and nothing to install.

Build a survey free Browse templates