What Is NPS, and Why Averaging the Scores Is Wrong
NPS is not the average of the scores; it's the percentage of promoters minus the percentage of detractors. Take the average and you produce a number that has no interpretation.
NPS is built from a single question: "How likely are you to recommend us to a friend?" answered on a scale of 0 to 10. But the number that comes out of it is not the average of those answers.
The correct calculation
Respondents are split into three groups:
| Score | Group | Meaning |
|---|---|---|
| 9 and 10 | Promoters | Actively recommend you |
| 7 and 8 | Passives | Satisfied but indifferent; not included in the calculation |
| 0 to 6 | Detractors | May actively discourage others |
The formula: NPS = % promoters − % detractors
The result is a number between −100 and +100, not between 0 and 10.
A worked example
Out of 20 responses: 5 promoters (25%), 7 passives, 8 detractors (40%).
NPS = 25 − 40 = −15
Now compute the average of the same responses: about 6.8 out of 10. A 6.8 looks "acceptable." A −15 is a warning. Same data, two completely different messages, and −15 is the right one, because it says you have more detractors than promoters.
Why passives are left out
This is deliberate. Someone who gives a 7 or an 8 is satisfied but does nothing: they neither recommend you nor speak badly of you. NPS was designed to measure growth through recommendation, and passives play no part in that.
Interpreting the number
- Above 50: excellent
- 20 to 50: good
- 0 to 20: acceptable but fragile
- Below zero: detractors outnumber promoters. This is a warning, not just a bad score
But comparing NPS across industries is meaningless. A mobile carrier's NPS can't be compared with a café's. What is meaningful is the trend of your own NPS over time.
Three common mistakes (we see them a lot in Iran)
- A small sample. With 20 responses, two people moving between groups shifts the NPS by 20 points. Below 100 responses, don't take the trend seriously.
- Asking at the wrong moment. Asking for NPS right after a successful interaction inflates the number artificially. If you want the real figure, ask a random sample, not the people whose problem was just solved.
- Not asking "why." The number alone doesn't tell you what to do. An open-ended follow-up ("What was the main reason for your score?") is worth more than the number itself.