Finance⏱ 4 min read
How to Calculate Net Promoter Score (NPS)
NPS is one of the most widely used customer satisfaction metrics in business. Here is the calculation, how to interpret your score, and its well-known limitations.
Net Promoter Score measures customer loyalty with a single question: "How likely are you to recommend us to a friend or colleague?" on a 0-10 scale. The calculation is straightforward — interpreting it is more nuanced.
The NPS Calculation
Respondents are classified by their rating:
Promoters: 9-10 (enthusiastic advocates)
Passives: 7-8 (satisfied but indifferent)
Detractors: 0-6 (unhappy, may spread negative word of mouth)
NPS = % Promoters - % Detractors
(Passives are excluded from the calculation)
Example: 200 survey responses
Promoters (9-10): 90 → 45%
Passives (7-8): 60 → 30%
Detractors (0-6): 50 → 25%
NPS = 45% - 25% = +20
NPS ranges from -100 (all detractors) to +100 (all promoters)
What NPS Scores Mean
NPS ScoreClassificationContext
Below 0PoorMore detractors than promoters — urgent attention needed
0-29GoodAcceptable — room for improvement
30-69GreatStrong customer loyalty
70+World-classExceptional — Apple, Amazon territory
Industry context matters enormously. An NPS of +30 is world-class in telecoms or insurance; the same score would be mediocre in software. Always compare to your sector benchmark, not absolute standards.
Statistical Reliability
NPS needs sufficient sample size to be meaningful.
Margin of error = 1.96 x sqrt((P x (1-P)) / n)
(where P = proportion of promoters or detractors, n = sample size)
At 200 responses, 45% promoters:
Margin of error = 1.96 x sqrt(0.45 x 0.55 / 200) = 1.96 x 0.0352 = ±6.9%
NPS of +20 with 200 respondents: roughly +20 ± 14 (95% CI)
This is a wide range — NPS requires large samples for precision.
Minimum for reliable NPS: 200-250 responses
Reliable change detection (to track improvement): 400-500 responses
NPS Limitations
- Single question oversimplifies: A customer might give a 9 because the product is good but the support is terrible — NPS won't tell you that
- Cultural bias: US respondents rate more extremely; European respondents give fewer 9s and 10s — cross-country comparison is unreliable
- Correlation vs causation: High NPS correlates with growth in some industries, not all — it's not universally predictive
- Gaming risk: Staff incentivised on NPS will ask customers to give high ratings, inflating scores without improving experience