The NPS calculation fits in one sentence: the percentage of promoters minus the percentage of detractors. Nothing is averaged or weighted, and there is no hidden maths. What deserves more attention is everything around that subtraction: who counts as a promoter, why passives affect the result without appearing in it, and which mistakes turn an honest score into a decorative number. This article walks through the formula step by step with a real example, covers the classic traps, and includes a calculator, just below, that does the work for you.
If you are still getting to grips with the Net Promoter Score itself, start with our complete guide to NPS, then come back here for the numbers. And if you already have your responses to hand, enter them straight away:
NPS calculator
Enter the number of responses per group; the score updates instantly.
The NPS formula
The Net Promoter Score rests on a single question: "How likely are you to recommend us to a friend or colleague?", scored from 0 to 10. The responses fall into three groups:
- Promoters (9 and 10): your ambassadors, the customers who actively recommend you.
- Passives (7 and 8): satisfied but unenthusiastic, and ready to switch for a better offer.
- Detractors (0 to 6): disappointed, at risk of churning and of spreading negative word of mouth.
The formula: NPS = % of promoters − % of detractors. The result runs from −100 (everyone is a detractor) to +100 (everyone is a promoter). Although it looks like one, the result is not a percentage but a balance, and it is written without the % sign.
One detail that matters: passives never appear in the subtraction, but they still affect the result. They count towards the response total, so they dilute your percentage of promoters. Two companies with the same number of promoters and detractors but a different number of passives will have different scores, and that is deliberate: a large pool of lukewarm customers is bad news in its own right.
A step-by-step example
Say your survey collected 200 responses:
- Sort the responses. 120 customers gave a 9 or 10, 60 gave a 7 or 8, 20 gave between 0 and 6.
- Work out each group's share of the total. Promoters: 120 / 200 = 60%. Detractors: 20 / 200 = 10%. (Passives: 30%; they play no part in the subtraction.)
- Subtract. 60 − 10 = NPS +50.
That is exactly the calculation the calculator above performs, with these values pre-filled. Change a single number and watch the score move: turning 20 passives into promoters lifts the score by 10 points without a single detractor changing their mind. The mechanics of NPS reward enthusiasm, and mere satisfaction earns nothing extra.
The five most common calculation mistakes
The formula is simple; the mistakes around it are surprisingly widespread.
- Averaging the scores. An average of 8.2 is not an NPS and does not behave like one: the average hides polarisation, the NPS formula exposes it. The two numbers tell different stories; never mix them in the same report.
- Removing passives from the denominator. The percentages are always calculated as a share of the total number of responses, passives included. Excluding them artificially inflates the score.
- Mixing relational and transactional. A relational NPS that measures the overall relationship and a transactional NPS measured right after a delivery never belong in one combined figure: they answer different questions, on different clocks.
- Publishing the score without its denominator. "NPS +38" means nothing on its own; "+38 from 412 responses, a 22% response rate" is a result. A score based on 15 responses moves 13 points when two customers change their mind.
- Comparing periods with very different response rates. Response rates are falling everywhere: Qualtrics XM Institute data shows survey response rates have dropped 7 to 8 percentage points since 2021, and only 16% of customers strongly believe their feedback leads to change. If your score is "improving" while your response rate collapses, it may simply mean that your unhappy customers have stopped answering.
How to interpret your result
A positive NPS means your promoters outnumber your detractors; above +30 you are in the better half of the pack across industries; above +50, excellent; beyond +70, you are playing in the league of the most loved brands. Bain & Company, where Fred Reichheld created the metric in 2003, reports that sustained value creators post Net Promoter Scores twice as high as the average company. You will find detailed benchmarks by sector in our guide to what counts as a good NPS score.
The most useful comparison, though, is your own: the same score, on the same touchpoint, a quarter earlier. Your own trend line is the benchmark. NPS is a trend instrument before it is a ranking instrument.
Calculating NPS in Excel or Google Sheets
In practice, you have exported your responses and all the scores (0 to 10) sit in column A, one score per row. You need three counts: how many scores are 9 or above, how many are 6 or below, and how many scores there are in total. That is exactly what the COUNTIF function does: it counts the cells that meet a condition.
Set up four formulas, one per cell:
| Cell | Formula | What it counts |
|---|---|---|
| B1 | =COUNTIF(A:A,">=9") | your promoters (scores 9 and 10) |
| B2 | =COUNTIF(A:A,"<=6") | your detractors (scores 0 to 6) |
| B3 | =COUNT(A:A) | the response total, passives included |
| B4 | =(B1-B2)/B3*100 | your NPS: (promoters minus detractors), divided by the total, in points |
With the example from this article, B1 shows 120, B2 shows 20, B3 shows 200, and B4 returns 50. Round to the nearest whole number, and always keep the B3 total next to the score in your reports. If a colleague uses Excel in French, the same functions are called NB.SI and NB, with semicolons instead of commas; the logic is identical.
And if spreadsheets are not your thing, scroll back up to the calculator at the top of this page, enter your three totals, and you get the same result without any formulas.
After the calculation: the why
The score tells you where you stand; it never tells you why. The why lives in the open question that follows the score, and that is where the real work starts: grouping the answers by topic, reading the sentiment behind the words, spotting what turns customers into detractors and what to do with them while recovery is still possible.
That is exactly the work our AI engine ISAAC automates: every open answer is grouped by topic, with a sentiment per topic, in your customers' own language, and impact analysis shows which topics actually move your score. Any spreadsheet can do the calculation. Understanding is another matter.
Frequently asked questions
How do you calculate NPS?
Sort the answers to the recommendation question (0 to 10 scale) into promoters (9 and 10), passives (7 and 8) and detractors (0 to 6), work out promoters and detractors as a percentage of the total number of responses, then subtract: % promoters − % detractors. The result runs from −100 to +100.
Can NPS be negative?
Yes. If your detractors outnumber your promoters, the score is negative, down to −100 at worst. A negative NPS is an alarm signal: the priority is no longer measuring but investigating the causes.
Do passives count in the NPS calculation?
They play no part in the subtraction, but they belong to the total on which the percentages are calculated. Excluding them from the denominator artificially inflates the score; it is one of the most common mistakes.
What is a good NPS?
Above 0, your promoters dominate; above +30 you are above the cross-industry median; above +50, excellent; beyond +70, exceptional. The most reliable benchmark is your own trajectory, touchpoint by touchpoint.
How often should you recalculate NPS?
Relational NPS is measured once or twice a year; transactional NPS is calculated continuously, after every measured interaction. In both cases, always compare periods with comparable response rates.
The calculation takes a few seconds; you have just done it above. Turning that number into customers who stay is the work that comes after, and that work is ours. Book a demo and bring your verbatims: we will show you what your score is trying to tell you.