NPS Calculator: Check Your Score and Missing Responses

Use one eligible invited cohort

Enter the number of known-eligible customers invited and the disjoint counts of valid promoter, passive and detractor answers. Each customer contributes at most one answer under your documented deduplication rule. Counts are whole numbers from zero to one million; their sum cannot exceed invitations. The target accepts −100 to 100 with two decimals. It is your threshold, not a recommended benchmark.

What the calculator measures

Let N be invited customers, P promoters, S passives and D detractors. R=P+S+D is the valid answer count, U=N−R the missing count. Respondent NPS is 100(P−D)/R. Valid-answer coverage is 100R/N. This is a simple known-cohort coverage fraction, not an AAPOR standardized response rate and not the percentage of all customers who responded.

At the lower extreme, all U missing answers are detractors: L=100(P−D−U)/N. At the upper extreme, all are promoters: H=100(P−D+U)/N. Width is H−L=200U/N NPS points. This interval describes possible completions of the invited roster, not sampling uncertainty, predicted scores or a bias correction.

For target T, missing answers need a net promoter-minus-detractor count of at least ceil[TN/100−(P−D)]. A required net count of −U or less is automatic; more than U is unattainable. A number between them means different missing answers can change the conclusion. The tool does not recommend selectively collecting positive answers.

Known-answer example

N=1,000, P=140, S=40, D=20 and T=30. R=200, U=800, respondent NPS=60, coverage=20%, L=−68, H=92 and width=160 points. The target requires a missing net count of 180. If everyone answers, the range collapses to the observed score. If nobody answers but N is positive, respondent NPS is undefined and the range is −100 to 100.

Validation and safe interpretation

Zero invitations produce undefined ratios and no cohort decision. Blank, negative, fractional, nonfinite and excessive counts fail validation. Changes clear the prior result. Counts are retained exactly at the accepted scale; displayed scores round to two decimals, while target decisions use integer hundredths to avoid rounding a boundary into a pass.

Copy/download includes inputs, the anonymous scope note and limitations. It runs locally and does not call AI or a survey provider. Do not enter weighted totals, incomplete eligibility estimates or multiple answers per customer. The model holds observed answers fixed and assumes no measurement error in them. It does not cover customers outside the invitation roster, estimate a confidence interval, identify dissatisfied individuals or establish a link to revenue.

Sources checked October 13, 2026: Bain’s NPS scoring definition and AAPOR Standard Definitions. Publication/update dates for these living pages are unavailable. They support scoring and caution about interpreting response rates; this finite-cohort worksheet is original synthesis, not an endorsed AAPOR response-rate formula.