Net Promoter Score (NPS) calculator

How the Net Promoter Score calculator works

Net Promoter Score is the percentage of promoters minus the percentage of detractors from a survey that asks how likely someone is to recommend you, on a scale of 0 to 10. Promoters score 9 or 10, passives score 7 or 8, and detractors score 0 to 6. Enter the counts in each bucket and this calculator returns the score, the percentage breakdown, and, unlike most, a confidence interval that tells you how much to trust it.

The default numbers show the pattern. With 62 promoters, 28 passives, and 10 detractors out of 100 responses, the score is 62 minus 10, which is 52. The passives sit out of the subtraction, but they still count in the denominator, which is why adding passives lowers both percentages and pulls the score toward zero.

The formula, and what the score can be

The score is percent promoters minus percent detractors, and it lands somewhere between -100 and +100. You hit +100 only when every respondent is a promoter, and -100 when every one is a detractor. A score of 0 means promoters and detractors cancel out, so anything positive means more people would recommend you than would warn others off.

The choice of a 0 to 10 scale is deliberate. It gives respondents enough room that they don't default to a lazy top mark, and it draws a hard line: a 7 or 8 is not counted as a win, because those passives are, in Bain's phrasing, satisfied for now and easily tempted away. That strictness is why a Net Promoter Score reads lower than a plain satisfaction percentage on the same customers.

How much to trust the number

A Net Promoter Score carries a wide margin of error, because each response is scored as plus one, zero, or minus one, so small samples swing hard. The 95% margin is 1.96 times the square root of the variance divided by the sample size, times 100, where the variance is the promoter proportion plus the detractor proportion, minus the score fraction squared. This calculator reports that margin and the resulting interval next to every score.

The practical effect surprises people, and it is bigger than most published guidance admits. On the default 62 promoters, 28 passives and 10 detractors, 30 responses carry a margin of about 24 points, so a score of 52 honestly means somewhere between 28 and 76. At 100 responses the margin is about 13 points. The margin also depends on the split: an evenly divided audience swings harder than a lopsided one, because the variance is larger.

Those numbers are roughly half again as wide as the ones you will find on most sites, and the gap has a single cause. A survey percentage has a margin of 1.96 times the square root of 0.25 over n, which gives the familiar 9.8 points at 100 responses. An NPS response is not a percentage: it is worth +1, 0 or -1, so its variance is larger and its margin is wider. Anyone quoting about 10 points at 100 responses has reached for the percentage formula. An earlier version of this page did exactly that, which is why the table above now carries the split as well as the sample size.

A few hundred responses is usually the point where a quarter-on-quarter move starts to mean anything at all.

Is a change in your score real

Comparing two surveys is where the interval earns its keep, though the test people reach for is not quite the right one. The proper comparison is on the difference: the change is real at 95% confidence when the gap between the two scores is bigger than 1.96 times the square root of the two standard errors squared and added together.

This is the trap behind most Net Promoter Score dashboards. A jump from 48 to 53 looks like progress, but on 150 responses each the gap of 5 points sits against a critical value of about 15, so the honest read is that nothing changed. The calculator gives you the margin precisely so you can make that call instead of celebrating noise.

One refinement worth knowing, because the shortcut runs the other way too. Two intervals that fail to overlap always mean a real difference, but overlapping intervals do not always mean the opposite. Run the same 48 against 53 on 1,500 responses each and the intervals still overlap, 44.6 to 51.4 against 49.6 to 56.4, while the difference of 5 points clears its critical value of 4.8 and is significant. Eyeballing the overlap is a conservative rule of thumb that will tell you nothing happened when something did.

What counts as a good score

There's no universal good Net Promoter Score, and any tool that hands you a single pass mark is misleading you. The usual advice is to compare against benchmarks for your own sector. Reasonable, except that the published benchmarks disagree with each other to a degree that makes them close to unusable.

Financial services, the same metric, five sources fetched on the same day:

Insurance is wider still: 71 in one set, 46 in another, and a third reports a range of 23 to 80 while calling it the most variable industry it measures. That is a spread of nearly 50 points on a scale where, as the table above shows, a 5-point move is usually noise.

The bands labelled good disagree too. Two of these sources call 0 to 30 a score that needs improvement; another calls the same range good.

None of that is dishonesty. Each vendor benchmarks whichever companies happen to be its customers, draws the line between B2B and B2C differently, and publishes a different year. It does mean that a benchmark is a loose orientation and never a pass mark, and that a tool printing one authoritative industry number would be selling you the same false precision this page spends its length arguing against.

What survives is smaller and more useful. Above 0, promoters outnumber detractors. Your own score, measured the same way on a comparable sample, is the only series you can trust to move for real reasons. The gauge here bands the range as a rough visual guide, and the label is a starting point rather than a verdict.

What the score does not tell you

Whether the Net Promoter Score predicts growth is contested in the academic literature, not settled. Published research has questioned whether it is a reliable predictor of company growth at all, and has found no empirical evidence that the likelihood-to-recommend question predicts growth better than other loyalty questions, or that it measures anything the conventional ones do not.

There's a sampling problem underneath the statistical one as well. People at the extremes, the delighted and the furious, are the likeliest to answer a survey in the first place, so the sample that reaches your calculator is already skewed before any arithmetic runs. A widening margin of error describes the uncertainty in what you collected, and it says nothing about who declined to respond.

None of this makes the metric useless. It makes it one tracked number among several, best read as a trend on a stable sample, and worst read as a single figure compared against somebody else's industry table.

eNPS uses the same maths

The employee version, eNPS, runs the identical formula on a different question: how likely an employee is to recommend the company as a place to work. The same 0 to 10 bands apply, so promoters, passives, and detractors are counted the same way, and this calculator produces an eNPS from employee responses exactly as it does a customer score. The margin-of-error caveat matters even more here, since employee surveys often have smaller samples than customer ones.

Two different things called NPS

The abbreviation NPS points at two unrelated tools. This one is the customer-experience metric. If you were looking for India's retirement scheme, the National Pension System calculator projects a pension corpus instead. They share three letters and nothing else.

Net Promoter, NPS, and Net Promoter Score are registered trademarks of Bain & Company, Fred Reichheld, and Satmetrix. This is a free calculator for the metric and is not affiliated with them.

Frequently asked questions

What is a Net Promoter Score calculator? A Net Promoter Score calculator turns the counts of promoters, passives, and detractors from a recommend survey into a single loyalty score. It works out the percentage breakdown and the score, and here it also reports a 95% confidence interval so you can tell a real change from sampling noise.

How is Net Promoter Score calculated? Net Promoter Score is the percentage of promoters minus the percentage of detractors. On a 0 to 10 recommend question, promoters score 9 or 10, passives 7 or 8, and detractors 0 to 6. Passives count toward the total but not the score, so 62 promoters, 28 passives, and 10 detractors out of 100 gives 62 minus 10, which is 52.

What is the range of Net Promoter Score? Net Promoter Score runs from -100 to +100. You reach -100 only if every respondent is a detractor, and +100 only if every respondent is a promoter. A positive score means promoters outnumber detractors, and 0 means they balance out.

What is a good Net Promoter Score? Any score above 0 means you have more promoters than detractors, and higher is better, but there is no universal threshold. Published industry benchmarks disagree sharply: financial services appears at 56, 68, 41 to 44 and 10 to 30 across four sources fetched the same day, and insurance ranges from 46 to 71 to a reported span of 23 to 80. Each vendor measures its own client base in a different year, so treat a benchmark as loose orientation and track your own score on a comparable sample instead.

What is the margin of error on an NPS? Net Promoter Score has a wide margin of error because each response counts as plus one, zero, or minus one. The 95% margin is 1.96 times the square root of the variance over the sample size, times 100, where the variance is the promoter proportion plus the detractor proportion, minus the score fraction squared. On a typical split, 30 responses give a margin of roughly 24 to 28 points and 100 responses about 13 to 15. Guidance quoting around 10 points at 100 responses has used the formula for a survey percentage, which understates an NPS margin by about half again.

How many survey responses do I need for NPS? The more responses, the tighter the score, and the answer also depends on how split your respondents are. On the default 62/28/10 mix, 30 responses give a margin of about 24 points, 100 about 13, 400 about 7, and 1,000 about 4. An evenly divided audience is wider again, closer to 20 points at 100 responses. A few hundred responses is usually the point where quarter-on-quarter changes become trustworthy.

Is a change in my NPS statistically significant? Test the difference, not the overlap. The change is real at 95% confidence when the gap between the two scores exceeds 1.96 times the square root of the two standard errors squared and summed. On 150 responses each, a move from 48 to 53 is a 5-point gap against a critical value of about 15, so it is noise. Non-overlapping intervals always mean a real difference, but overlapping ones do not always mean the reverse, so the eyeball test is conservative.

What is eNPS? eNPS, or employee Net Promoter Score, uses the same formula on a different question: how likely an employee is to recommend the company as a place to work. Promoters, passives, and detractors are scored on the same 0 to 10 bands, so this calculator computes an eNPS in exactly the same way as a customer NPS.

Sources

• Qualtrics, how to measure Net Promoter Score
• Bain and Company, measuring your Net Promoter Score
• Genroe, how to calculate the margin of error for NPS
• Retently, 2026 Net Promoter Score benchmarks by industry
• SurveyMonkey, NPS calculator and industry benchmarks
• Wikipedia, Net promoter score, including the academic criticism