Net Promoter Score®

Net Promoter Score® summarises customers’ stated willingness to recommend a brand or company. It can provide a consistent signal of relationship sentiment, but it is not a complete measure of satisfaction, loyalty, advocacy or growth. Its value comes from disciplined sampling, follow-up and improvement—not the number alone.

When to use it

• To track recommendation intent and support customer-experience improvement.
• To compare changes only when the population, question and collection method are consistent.

Origins

Companies long used satisfaction scales such as 1 to 10, 1 to 5, or 1 to 7 and asked many questions about product, delivery, sales, service and price. Fred Reichheld of Bain & Company popularised NPS through the 2003 Harvard Business Review article “The One Number You Need to Grow.”

The appeal was managerial simplicity, but the method was not intended to eliminate diagnosis. The recommendation question should be followed by “Why did you say that?” Later independent studies have not established NPS as universally superior to satisfaction or repurchase questions.

What it is

Ask: “How likely are you to recommend brand X to a colleague?” using a scale from 0 to 10, where 0 means not at all likely and 10 means very likely.

Promoters give 9 or 10 out of 10. Detractors give 6 or below. Respondents selecting 7 or 8 are called passives. NPS equals the percentage of promoters minus the percentage of detractors.

0 1 2 3 4 5 6 7 8 9 10

Not at all likely Extremely likely

NPS = % of promoters

(ie 9 & 10 scores) – % of detractors (ie 0–6 scores)

Responses often cluster between 6 and 10, with many at 7 or 8 out of 10. The calculation therefore discards distinctions within the 9 or 10 group and within the 6-and-below group. It also ignores movement among passives until a category boundary is crossed.

A high score may correlate with loyalty or growth in some settings, but correlation varies by category, culture and study design. Promoters are not automatically advocates and detractors do not always spread negative word of mouth.

The inherited B2B benchmark of 20 to 30 is not universal or current without a defined source and method.

A company might investigate customers scoring 7 or 8 out of 10 rather than merely trying to push them across a threshold. Likewise, the inherited rule that more than 20 per cent scoring 6 or below proves something is wrong is a diagnostic prompt, not a law. A score of 50+ does not guarantee an excellent product; verify experience, behaviour and sample.

Developments of the model

NPS became common in consumer and employee programmes because it is easy to communicate. Employee recommendation questions need separate validation and should not be treated as interchangeable with customer NPS.

Research reported by Hayes (2008) questioned whether likelihood to recommend predicts growth better than satisfaction or repurchase intention. Small samples are especially unstable. If only 100 customers respond and most choose 7 or 8 out of 10, a movement of one or two responses among promoter and detractor categories can change the score visibly.

Use NPS as one indicator within a wider system. Ask about reasons, satisfaction, repurchase, effort and observed retention. Preserve the full response distribution and confidence interval.

How to use it

A historical case describes Molson Coors’s UK distribution division. In 2011 it sought to become customers’ first choice and asked how likely they were to recommend the company after a recent transaction. Short telephone interviews, verbatim feedback, audio excerpts and quarterly reviews connected the score with action.

The case reported movement from +25 to +60 alongside improvements in account-management and credit-control interactions, lower churn and broader purchasing. These are historical claims from the case, not proof that the score caused the outcomes. The programme’s useful feature was closing the loop; linking regional-director bonuses directly to NPS also created gaming risk and should be balanced with quality and anti-coaching controls.

Some things to think about

• Retention can create value, but NPS does not by itself measure retention or causality.

• The inherited guidance recommends at least 50 and ideally 100 responses because many respondents choose 7 or 8 out of 10. Adequate sample size actually depends on population, variability, segmentation and precision; calculate it for the decision.

• Add diagnostic questions and ask “Why?” Use results to improve systems, not pressure customers or employees.

Top practical tip

Track NPS with the full distribution, uncertainty, response rate and customer comments. Assign improvement owners to recurring themes and verify whether behaviour changes.

Top pitfall

The score is easy to game and easy to overinterpret. Do not coach for the top 10-point category, compare unlike samples, or claim that a rating on its 10-point scale is the single truth about loyalty.

Further reading

• Reichheld, F.F. (two thousand and three). “The One Number You Need to Grow.” Harvard Business Review.
• Keiningham, T.L. et al. (two thousand and seven). “A Longitudinal Examination of Net Promoter and Firm Revenue Growth.” Journal of Marketing.