Net Promoter Score, or NPS, summarises answers to a likelihood-to-recommend question on a scale from 0 to 10. Responses are grouped into promoters, passives and detractors, and the score subtracts the percentage of detractors from the percentage of promoters. It ranges from −100 to +100.
The labels describe score bands. They should not be treated as complete accounts of a customer’s behaviour: someone classified as a promoter has expressed a high likelihood of recommending, not necessarily made a recommendation. The distinction matters when a simple score becomes a claim about loyalty or future growth.
Calculate it using all valid responses
Promoters select 9 or 10, passives select 7 or 8, and detractors select 0 through 6. Passives remain in the denominator even though their percentage is not subtracted or added directly.
For an illustrative set of 100 valid responses containing 50 promoters, 30 passives and 20 detractors, NPS is 50 minus 20, or +30. It is conventionally reported as a score, not 30%. Excluding the passives from the denominator would produce the wrong result.
Bain’s explanation of NPS measurement sets out the method. Keep the question, scale and calculation consistent when comparing waves, and report how incomplete or ineligible responses were handled.
Decide which relationship or event is being assessed
A periodic question about the overall company relationship is different from a survey triggered immediately after a support interaction. Both may use the recommendation format, but their timing and context shape the response. Combining them without distinction can obscure what the trend means.
Recruitment matters as much as arithmetic. If only recently successful customers receive the survey, the result cannot stand for everyone’s experience. Examine non-response bias and changes in the invitation process before attributing movement to a product improvement.
Comparisons with other organisations need compatible audiences, channels, timing and methods. A benchmark from a different context can be interesting without being a fair performance target.
Read the reasons and the distribution
Ask a neutral follow-up about what informed the rating. Review the answers systematically, including mixed and contradictory accounts. A respondent may give a high score because of a helpful support team while describing a recurring product failure that still deserves attention.
The same NPS can arise from different distributions of responses. Keep the underlying bands visible and report sample size and appropriate uncertainty, rather than narrating every small change as a shift in customer sentiment. Coaching people to select a particular score or selectively sending invitations undermines the measure.
NPS does not diagnose a cause or guarantee future retention or growth. Compare it with relevant behaviour and other evidence. Customer satisfaction and customer effort address different questions, while a Voice of the Customer programme can connect feedback to investigation and action. The score becomes useful when it leads to a better understanding of the experience behind it.