Staring at a pile of survey scores and unsure what to do with them? Here is how to calculate NPS: subtract the percentage of detractors (0 to 6) from the percentage of promoters (9 or 10). The result runs from -100 to 100.
This guide belongs to our wider coverage of customer experience metrics.
| Item | Value |
|---|---|
| Answer scale | 0 to 10 |
| Promoters | 9 or 10 |
| Passives | 7 or 8 |
| Detractors | 0 to 6 |
| Formula | % of promoters – % of detractors |
| Possible NPS range | -100 to 100 |
These figures come from the SurveyMonkey NPS calculator guide and the SurveyMonkey calculation guide. Vendors can revise their methods and benchmarks, so check the source before you rely on them.
The NPS formula and the three groups
The Net Promoter Score starts with one survey question: how likely is it that you would recommend this company to a friend or colleague? Respondents answer on a scale from 0 to 10, as the SurveyMonkey definition and formula page explains. For the background before the maths, see our explainer on what is net promoter score.
Each answer then lands in one of three groups:
- Promoters: ratings of 9 or 10 (SurveyMonkey NPS calculator).
- Passives: ratings of 7 or 8 (SurveyMonkey NPS calculator).
- Detractors: ratings from 0 to 6 (SurveyMonkey NPS calculator).
The score itself is the share of promoters minus the share of detractors, which SurveyMonkey writes as % of promoters – % of detractors. Passives are left out of the subtraction because they neither strengthen nor weaken momentum, though they still count in your total number of responses.
The metric was created by Fred Reichheld in 2003, according to Salesforce. A result can fall anywhere from -100 to 100, per the SurveyMonkey calculator page.
How to calculate NPS step by step
The routine below follows the method set out in the SurveyMonkey calculator guide and the calculation guide. It works the same way for a small sample or a large one once you have the counts.
- Collect every response in one place, such as a spreadsheet.
- Sort each respondent into promoter (9 or 10), passive (7 or 8) or detractor (0 to 6) and count each group (source).
- Divide the promoter count and the detractor count by the total number of responses, then multiply each by 100 (source).
- Subtract the detractor percentage from the promoter percentage. The remainder is your NPS.
You can also do it in one line. SurveyMonkey gives the formula as (Number of Promoters – Number of Detractors) / (Number of Respondents) x 100, which returns the same answer.
Worked examples with real numbers
Example 1: 150 responses
Suppose a survey collects 150 responses: 80 promoters, 30 passives and 40 detractors. Promoters make up 80 / 150 * 100% = 53% and detractors 40 / 150 * 100% = 27%. Subtracting gives 53% – 27% = 26%, i.e. an NPS of 26, as in the SurveyMonkey calculator guide.
That guide rounds each percentage to a whole number before subtracting. A tool that keeps the decimals may therefore show a slightly different figure for the same data.
Example 2: 200 responses
Now imagine 200 responses: 20 responses were in the 0-6 range (detractors), 40 responses were in the 7-8 range (passives) and 140 responses were in the 9-10 range (promoters). Subtract 10% (detractors) from 70% (promoters) and the NPS is 60, according to the SurveyMonkey NPS guide. That guide adds that the score appears as an integer, not a percentage.
Example 3: 137 responses
A larger, messier sample shows why the arithmetic matters. With 137 total responses, made up of 56 promoters, 51 passives and 30 detractors, the SurveyMonkey calculation guide works it out as 40.9 – 21.9 = +19. Notice that the passives are the largest slice after the promoters, yet they never appear in the subtraction.

Common mistakes when you calculate NPS
Most errors come from the grouping or the denominator rather than the subtraction itself.
- Dividing by the wrong total. SurveyMonkey says to divide by total responses, so passives belong in the denominator (source).
- Counting an 8 as a promoter. Only ratings of 9 or 10 qualify, and 7 or 8 are passives (source).
- Mixing up scales. Not every description of the method uses the same answer range, so confirm which range your survey tool offers before you group the data.
- Comparing against the wrong peers. Scores can vary dramatically with industry and company size (source).
What is a good NPS?
A good score is a relative idea. SurveyMonkey says global benchmarks typically range from +30 to +70, with the ideal figure depending on sector, region and customer expectations.
Industry gaps can be wide. IBM cites a 2024 Retently study in which the average NPS for an insurance company is 80, whereas it is 39 for cloud and hosting. The same page says the best performers receive an average NPS score of 80 or better.
Because of those gaps, track your own score over time and compare it with organisations that resemble yours.
Where CSAT fits and what to ask next
NPS measures willingness to recommend, whereas the customer satisfaction score CSAT asks how satisfied people are. IBM describes CSAT as the percentage of respondents who claim to be satisfied (4) or very satisfied (5).
The number alone rarely tells you what to fix. Salesforce notes that companies typically follow up with an open-ended question to learn why a customer chose their score. Read those answers by group, starting with the detractors.
Frequently asked questions
Can a Net Promoter Score be negative?
Show answer
Yes. Whenever detractors outnumber promoters, the result drops below zero. IBM describes the full span as -100 (all 6’s or less) to +100 (all 9’s and 10’s).
Do passives count in the calculation?
Show answer
They do not appear in the subtraction, but they enlarge the total you divide by. A survey with many passive answers therefore shrinks both percentages and pushes the score toward zero.
Should I write NPS with a percent sign?
Show answer
No. The SurveyMonkey NPS guide says the score always appears as an integer and not a percentage, so report a plain whole number.
How often should I run the survey?
Show answer
Often enough to see a trend. IBM advises deploying surveys frequently, so you can set benchmarks and see whether changes to the experience move the score.
Do I need special software?
Show answer
No. Once the answers are tallied, the arithmetic is three simple operations, so a basic spreadsheet is enough. Survey platforms mainly save time as the response count grows.



