NPS (Net Promoter Score) calculator
Enter the response counts or paste the list of responses. See the NPS, the promoter/passive/detractor shares and the confidence interval; compare two segments (two sites, two periods) and check whether the difference is just chance.
The calculation runs in your browser; the responses you enter are not sent anywhere.
Enter the response counts, paste a response list or load the sample.
Let us read your customer feedback together with your operational data and look at what moves your NPS.
Request a call01
How to use
A
Enter the number of responses for each score (0–10) or paste your response list; you can also load the sample data.
B
Read the NPS, the promoter/passive/detractor shares, the score distribution and the 95% confidence interval.
C
To compare two sites or two periods, open the second segment; check whether the difference and the p-value are significant.
02
How is NPS calculated?
Responses to the question "How likely are you to recommend us to a friend or colleague?" on a 0–10 scale fall into three groups: 9–10 promoters, 7–8 passives, 0–6 detractors. NPS is the percentage of promoters minus the percentage of detractors and is a number between −100 and +100; it is written without a percent sign.
For example, if 34 of 100 responses are promoters, 43 passives and 23 detractors, NPS = 34 − 23 = +11. Passives do not enter the formula directly but they change the shares indirectly because they are part of the total.
03
Confidence interval and sample size
NPS is calculated from a sample and carries uncertainty. If you think of each response as +1 (promoter), 0 (passive) or −1 (detractor), NPS is 100 times the mean of these scores. The variance is σ² = pP + pD − (pP − pD)², the standard error is SE = √(σ²/n) and the 95% interval is NPS ± 1.96·SE.
With 100 responses the interval is typically around ±15 points; with 30 responses it can exceed ±25 points. A difference of 5–10 points between two NPS values is therefore often just chance. The normal approximation is also rough for small samples; below 30 this tool warns about it explicitly.
04
Comparing two segments
For the NPS difference between two independent segments (for example two sites, or an earlier and a later period with different customers) the standard error is the square root of the sum of the squared standard errors of the two segments; z = difference / SE and the two-sided p-value come from the normal distribution. A small p says the difference is hard to explain by chance alone.
A significant difference does not tell you the cause and does not show that it matters in practice. If the same people answered in both periods (paired data), this comparison, which treats the segments as independent, is not appropriate. Comparing many segments with each other also increases chance "significant" findings.
FAQ
- What is a good NPS?
- There is no fixed "good" value; it varies a lot by industry, product and country. A positive NPS says there are more promoters than detractors. The most meaningful comparison is with your own history and with values measured the same way in the same industry.
- Should I write NPS as a percentage?
- No. NPS is the difference of two percentages but is written as a score between −100 and +100 without a percent sign: "NPS = +11".
- How many responses do I need?
- Below 30 the result only indicates a direction. For a reliable NPS and a meaningful comparison at least 100 responses per segment, preferably a few hundred, are recommended. The tool shows the confidence interval and margin of error according to the number of responses.
- Why are passives not part of the calculation?
- NPS is designed to measure the balance between those who recommend and those who speak negatively; passives are considered neither. They do affect the shares indirectly because they are part of the total: more passives lower both the promoter and detractor shares.
- If the difference between two segments is not significant, are their NPS equal?
- No. Not significant does not mean no difference; it means that with this sample size you cannot separate the difference from chance. More responses narrow the interval and give a clearer result.
- In what format should I paste the responses?
- Each response must be a whole number from 0 to 10; the separator can be a space, new line, tab, ; or comma. You can copy a single column from Excel. A header line is skipped; decimals or out-of-range values are skipped and their number is shown.
Turn feedback into decision support
Let us read your survey, complaint and operational data together to see which factors move customer satisfaction and how you can track them.