Your Net Promoter Score, counted properly
One question (“How likely are you to recommend us, from 0 to 10?”), three groups, one subtraction. This page classifies every response by the published definition (0 to 6 detractors, 7 to 8 passives, 9 to 10 promoters), shows the counts and percentages behind the score, and then answers the follow-up everyone actually asks: what would it take to move it.
Enter how many respondents gave each score. Leave a score at 0 if nobody chose it.
Disclaimer: This calculator gives estimates for general information only. It is not financial, medical or other professional advice. Net Promoter Score is a survey metric defined by its published method; what a score means for a specific business depends on its market, sample and how the survey was run.
What this calculator does
It does the small arithmetic of NPS exactly, and then refuses to stop there. Beyond the score itself you get the response total, the count and percentage in each of the three groups, the full score-by-score distribution when you enter counts or a list, and a “what moves the score” panel that prices improvement in responses rather than in slogans: how much one conversion is worth at your sample size, how many detractors would have to become promoters to hit a target, and how many fresh promoter responses the same target would take if nobody changed their mind. The two routes differ by a factor of four in the worked example, which is the most useful thing most teams learn from their first honest NPS calculation.
How to use it
- The groups are fixed by the method, not by taste. Scores of 9 and 10 are promoters, 7 and 8 are passives, 0 to 6 are detractors. A business that counts its 8s as promoters is not running an optimistic NPS, it is running a different metric with the same name, and its scores cannot be compared with anyone’s, including its own past if the definition changed.
- Passives count in the total but in neither percentage. They dilute both camps. A score can fall while promoters grow, if detractors grow faster or if the passive middle swells at the promoters’ expense; the group table here is what tells you which happened.
- Small samples swing wildly. At 10 responses, one person changing groups moves the score by up to 20 points; at 1,000 responses, by 0.2. The what-moves panel shows your own sample’s sensitivity, which is the honest way to decide whether a 4 point change is news or noise.
- NPS is not an average. The same responses that score 40 here average about 7.6 out of 10, and the two numbers answer different questions. The average measures central tendency; NPS measures the balance of enthusiasm over unhappiness, ignoring the middle. Report both, label both, and never average a set of NPS scores from different surveys into one figure without weighting by responses.
- Compare like with like. Scores differ by industry, by country, by channel (email surveys run hotter than in-product ones) and by when in the relationship you ask. A benchmark is only useful if its method matches yours, which is a reason to record your method alongside every score you publish internally.
How it is calculated
- Classification: each response is a detractor (0 to 6), a passive (7 or 8) or a promoter (9 or 10), per the published NPS definition.
- NPS = percentage of promoters – percentage of detractors, reported as a whole number from -100 to +100. It is not a percentage itself, and the exact unrounded difference is shown in the results so the rounding is transparent.
- One detractor converted to a promoter moves the exact score by 200 / total responses points: the detractor percentage falls and the promoter percentage rises at once. One passive converted moves it by half that.
- Target by conversion: the number of detractors who must become promoters, total held constant, is the gap to the target divided by 200 / total, rounded up, and it cannot exceed the detractors available.
- Target by new responses: adding k new promoter responses solves (promoters + k – detractors) / (total + k) at the target percentage; the count is rounded up, and a target of 100 is unreachable by addition while any detractor remains.
Worked examples
The balanced hundred. 100 responses: 20 detractors spread across 0 to 6, 20 passives, 60 promoters. Promoters are 60.00%, detractors 20.00%, and the score is 40. To reach 50 by conversion takes 5 detractors becoming promoters (each conversion is worth 2.00 points at this size). To reach 50 by collecting new promoter responses instead takes 20, because each new response also grows the total it is divided by. Persuasion is four times cheaper than volume here, and the panel shows that trade for your own numbers.
The edges. Every response a promoter gives a score of 100; every response a detractor gives -100. Zero responses give no score at all, and the calculator says so rather than printing a zero that would look like data.
The pasted list. Ten ratings (10, 9, 8, 7, 6, 10, 9, 5, 10, 9) hold 6 promoters, 2 passives and 2 detractors: 60% minus 20% is again 40, on a tenth of the sample. Same score, far less certainty, which the sensitivity rows make concrete: at this size a single conversion is worth 20 points.
Groups you already have. A survey tool reporting 45 promoters, 30 passives and 25 detractors gives 45% minus 25%: a score of 20, with the passives quietly forming the largest single group after promoters, a fact the headline number alone hides.
Reading your result
Read the three group percentages before the score. A score of 40 built from 60/20/20 (promoters/passives/detractors) is a healthy business with a loud minority of unhappy customers. A 40 built from 45/50/5 is a contented but unenthusiastic customer base that a competitor’s launch could scatter. The score is identical; the strategies are opposites, one fixing pain, the other building delight. Then read the sensitivity rows: if one conversion moves your score by more than the change you are celebrating or mourning, you are reading noise, and the correct response is more responses, not a new initiative.
On the common guide bands (below 0 needs urgent attention, 0 to 30 is a reasonable range, 30 to 70 is strong, above 70 is rare air): treat them as folklore with some truth in it, not as standards. They come from cross-industry habit, not from the method itself, and a score of 25 can lead one market while 45 trails another. Your own trend, measured the same way on comparable samples, is the benchmark that cannot lie to you.
Limitations
- Sampling decides what the score can mean. Who was asked, who bothered to answer, and when, all bend the result. Self-selected respondents run more extreme than the customer base; a score from 4% of customers is a measure of the motivated. The calculator computes the score of the responses entered; it cannot repair how they were gathered.
- The score compresses a distribution into one number. Two businesses with the same score can have very different distributions, including a polarised base of lovers and haters that averages out to calm. The distribution table is included so the compression is never the only view.
- No significance testing is offered. The sensitivity rows show the arithmetic weight of one response, which covers most practical judgements, but formal confidence intervals for NPS are contested and are not claimed here.
- Transactional and relationship surveys are not mixed here on purpose. Scores asked right after an interaction measure that interaction; periodic relationship scores measure the brand. Entering both into one count produces a number about neither.
Frequently asked questions
How is NPS calculated?
Classify each 0 to 10 response: 9 and 10 are promoters, 7 and 8 are passives, 0 to 6 are detractors. Take the percentage of promoters, subtract the percentage of detractors, and report the result as a whole number between -100 and 100. Passives matter only through the total they are part of. The worked examples above show the full chain three ways.
Why are passives ignored in the score?
By the method’s design: passives are counted in the total (so they dilute both percentages) but are treated as neither advocacy nor damage. The logic is that a satisfied-but-uncommitted customer does not recommend you and does not warn others off. Whether or not that matches your market, changing the treatment changes the metric, and the scores stop being NPS.
What is a good NPS score?
Above 0 means more promoters than detractors; above 30 is generally read as strong and above 70 as exceptional, but the bands are convention, not standard, and they vary sharply by industry and survey method. The defensible comparisons are your own past scores on the same method, and direct competitors measured the same way. A single absolute number, quoted without its sample and method, is closer to advertising than measurement.
Is NPS the same as average rating?
No. The average of the balanced-hundred example is about 7.6 out of 10 while its NPS is 40, and the two can move in opposite directions: converting passives to promoters raises NPS without moving the detractor problem at all, while soothing detractors into passives raises NPS with no new advocates. Each metric is honest about one thing; using both, as this page’s group table encourages, is honest about more.
Related calculators
Method: the published Net Promoter Score definition (promoters 9 to 10, passives 7 to 8, detractors 0 to 6; score as percentage of promoters less percentage of detractors), with conversion and new-response target arithmetic as defined on this page. Last reviewed: 13 October 2026.