Net Promoter Score Calculator

Written out

Net promoter score is the percentage of promoters minus the percentage of detractors, from a 0–10 likelihood-to-recommend question. Promoters score 9–10, passives 7–8, detractors 0–6.

NPS = %promoters − %detractors, giving a range from −100 to +100. Passives are counted in the denominator and contribute nothing to the score.

The scale is deliberately harsh: a 7 out of 10 is treated as neutral rather than positive.

In practice

From 400 responses: 180 promoters, 140 passives, 80 detractors. NPS = 45% − 20% = +25.

Now a different distribution with the same average rating: 100 promoters, 260 passives, 40 detractors. NPS = 25% − 10% = +15.

The second group is more uniformly satisfied and scores lower, because NPS rewards enthusiasm rather than absence of complaint.

That is the metric behaving as designed, and it is also why NPS moves in ways an average satisfaction score does not.

The limitations

Response rates are typically low and skew toward strong opinions in both directions, so the sample is not the customer base. A score built on a 4% response rate describes the people who felt strongly enough to answer.

Cross-industry and cross-country comparisons are also unreliable, since scoring conventions differ culturally and by category. A +25 means different things in different places.

Putting it to use

Track your own score over time rather than against a benchmark, and always collect the follow-up question asking why. The score is a number; the verbatim answers are the useful output.

Then act on detractors individually where the volume allows. A detractor contacted and resolved frequently becomes a promoter, and the intervention is cheap relative to acquiring a replacement customer.

Whether NPS predicts anything

The claimed link between NPS and growth has been contested for years, and the evidence is weaker than the metric's popularity suggests. It correlates with growth in some categories and not in others.

What it does reliably provide is a consistent, simple instrument that a whole organisation can understand, and a structured prompt for the open-ended feedback that actually contains the information.

Used that way it is genuinely useful. Used as a target to be raised, particularly when tied to compensation. It invites the score to be managed rather than the experience, and businesses that do this typically end up with a rising number and unchanged customer behaviour.

Response bias is worth quantifying occasionally by comparing the demographics of respondents against the customer base, since a score built on an unrepresentative sample will drift without anything changing.

Transactional NPS, asked after a specific interaction, is more actionable than relationship NPS asked periodically, because it points at a moment rather than a general feeling.

Following up with detractors individually converts a measurement exercise into a retention one, and the recovery rate on a personally handled complaint is high enough to justify the effort.

Sampling continuously rather than in periodic waves smooths the noise and catches deterioration sooner, and it avoids the artefact where a score moves because of when the survey was sent.

The score compresses a distribution into one number and discards most of it, which is worth remembering before acting on a change. A shift from 30 to 35 could be more promoters, fewer detractors, or movement within the passive group that happens to cross a boundary. The underlying distribution answers what changed; the score only says that something did.