How it works
Respondents rate their likelihood to recommend on a scale from zero to ten. Nines and tens are promoters, sevens and eights are passives and count for nothing, and everything from zero to six is a detractor. Promoter percentage minus detractor percentage gives a score between −100 and +100.
The design is deliberately blunt. Its value was never precision; it was giving a whole organisation one comparable number and, more usefully, a free-text follow-up asking why.
Its known weaknesses
The bucketing throws away information — a seven and a zero are recorded identically — and it makes the score jumpy in small samples. Response bias is severe, since the delighted and the furious answer while everyone else ignores the survey. Cultural response norms shift scores across regions, which makes international comparison unreliable. And when NPS becomes a target attached to compensation, it stops measuring anything, because asking is easy to game.
Cross-company comparison is the weakest use of all: two companies surveying different populations at different moments in the customer lifecycle are not measuring the same thing.
When to ask, and of whom
Two design choices decide whether the result means anything. The first is timing: a score collected immediately after onboarding measures the onboarding, and one collected at renewal measures the relationship. Both are legitimate and they are not comparable.
The second is the respondent. Surveying the day-to-day user and surveying the budget holder produce different answers about the same product, and the second is the one that predicts renewal. Programmes that survey whoever is easiest to reach tend to track satisfaction with the interface rather than with the outcome.
The useful part
The verbatim comments are worth more than the number. That is also the part that generalises competitively: public reviews are effectively the same signal collected without a survey, from people who chose to explain themselves at length.
Read for the reason rather than the score, and NPS becomes a serviceable trend indicator for your own base — is the same complaint growing? — rather than a benchmark to argue about.
The detractor comments are the most competitively useful part of the whole exercise. They are, almost word for word, the objections your rivals will raise about you in their next deal, delivered early and free by people who wanted the product to work.
How IndustryLens handles this
IndustryLens does not run surveys. It reads the public equivalent: what a competitor's customers write on review sites, what the recurring themes are, and how the sentiment behind them moves over time.
Net Promoter Score: common questions
What is Net Promoter Score?
Net Promoter Score measures customer advocacy from a single question about likelihood to recommend, scored zero to ten. Nines and tens are promoters, sevens and eights are passives that count for nothing, and zero to six are detractors; promoter percentage minus detractor percentage gives a score between −100 and +100.
What are the weaknesses of NPS?
The bucketing throws away information, since a seven and a zero are recorded identically, and it makes the score jumpy in small samples. Response bias is severe because the delighted and the furious answer while everyone else ignores the survey, and cultural response norms make international comparison unreliable. Cross-company comparison is the weakest use of all, since two companies surveying different populations at different lifecycle moments are not measuring the same thing.
What is the most useful part of an NPS programme?
The verbatim comments are worth more than the number, and the detractor comments are the most competitively useful part of the whole exercise. They are, almost word for word, the objections your rivals will raise about you in their next deal, delivered early and free by people who wanted the product to work.