NPS: what the Net Promoter Score is and how to calculate it

NPS (Net Promoter Score) is a loyalty index that sums up, in a single number, how willing customers are to recommend a company, product or service. It comes from a single question: "on a scale of 0 to 10, how likely are you to recommend us to a friend or colleague?". Based on that score, each customer falls into one of three groups, and the final number ranges from -100 to +100.

The metric was popularized by Fred Reichheld, of the consultancy Bain & Company, together with Satmetrix, around a simple thesis: a spontaneous recommendation is a strong signal of loyalty, because we only recommend what we trust enough to put our own reputation on the line for. That is why NPS measures loyalty and word of mouth, not satisfaction with a specific interaction, which is the territory of CSAT.

The calculation has two steps. First, classify each response by its score; then, subtract the percentage of detractors from the percentage of promoters. Passives are part of the total base, but do not add or subtract in the result.

• Promoters: scores 9 and 10.
• Passives: scores 7 and 8.
• Detractors: scores 0 to 6.

NPS = % of promoters minus % of detractors. Example: in 200 responses, 120 promoters (60%), 50 passives (25%) and 30 detractors (15%). NPS is 60 minus 15, that is, 45. The result is always a whole number between -100 (all detractors) and +100 (all promoters), and by convention it is written without the percent sign.

The three groups are not arbitrary labels, they describe distinct behaviors. Promoters (9-10) are loyal and enthusiastic: they tend to buy more, stay longer and bring in new customers through word of mouth. Detractors (0-6) are unhappy, may speak badly of the brand and hold much of the churn risk.

Passives (7-8) are satisfied customers, but without loyalty: they like you enough to stay, but not enough to recommend, and they switch providers easily when a better offer appears. They are left out of the calculation precisely because NPS wants to measure the extremes of loyalty, not the lukewarm middle. Ignoring passives in the score, but tracking them in the base, helps you see how many customers are one step away from becoming promoters or detractors.

There are two ways to run the survey, and they answer different questions. Relational NPS measures the overall perception of the brand and is usually sent at fixed intervals (for example, every quarter), regardless of a recent interaction. It is used to track loyalty over time and compare periods.

Transactional NPS is triggered right after a specific event, such as a purchase, a support ticket or onboarding, and measures the experience of that moment. While relational NPS answers "how is our relationship?", transactional NPS answers "how was this interaction?". Mature companies use both: relational for the pulse of the base, transactional to find the exact points in the journey that create promoters or detractors.

NPS, CSAT and CES are the three classic customer experience metrics, and they are easily confused. NPS measures loyalty and willingness to recommend, a long-term relationship view. CSAT measures point-in-time satisfaction with a specific interaction or product, usually right after it happens. CES measures the effort the customer had to make to get something done.

• NPS: "would you recommend us?", loyalty and word of mouth over the long term.
• CSAT: "were you satisfied?", immediate satisfaction with a touchpoint.
• CES: "was it easy to solve?", customer friction and effort.

None replaces the others. NPS tells you whether the relationship is healthy, CSAT and CES point to where, in the journey, the experience gets better or worse.

There is no universally "good" NPS: what is excellent in one sector can be average in another, and comparisons across industries or countries mislead, because culture shapes how people give scores. As a rule of thumb, a positive NPS (above zero) means more promoters than detractors, and scores above 50 are usually seen as strong, but the most useful move is to compare your number against your own history and direct competitors.

NPS also has known limits. Being a single number, it tells you "how loyal" customers are, but not "why", so it only pays off when paired with the open-ended question that explains the score. It is sensitive to who responds and to how the survey is run, and should not be read in isolation. That is why it usually feeds into a customer health score, alongside product usage and support, and gets cross-checked against real churn so it drives action rather than being just a vanity number.