# Net Promoter Score The net promoter score is promoters minus detractors, used as a proxy for loyalty and for the likelihood of a recommendation. **The question.** How likely are our customers to recommend us, and what is behind the low scores? Also known as: Usually written NPS. The wider program built around the score is called the Net Promoter System. ## When to use it - A client wants a repeatable, low-effort read on customer sentiment that can be tracked by segment, branch, or product over time. - You need the open-text reasons behind low scores to build a service-recovery or product-fix list. - Several business units need one comparable question, and a long satisfaction survey will not get answered. ## When to avoid it - The sponsor expects the score alone to forecast revenue growth. An independent 2007 study did not replicate the claim that it beats other loyalty questions as a growth predictor. - Customers cannot choose to stay, leave, or recommend, as with a regulated utility or a captive internal service. The question asks about something they never do. - The score will set individual bonuses or branch rankings. Staff learn to coach customers toward 9 and 10, and the survey starts measuring the coaching. - You need to know what to fix and have only the number. Interviews or a driver analysis will tell you more. ## Overview Net Promoter Score reduces customer feedback to one question: how likely are you to recommend us to a friend or colleague, scored from 0 to 10. The share of 9 and 10 answers, minus the share of 0 to 6 answers, gives a number from minus 100 to plus 100. It supports a choice about where to look first: which segment, branch, or product has customers who would not vouch for you, and what they say is wrong. Two groups can share a score and need opposite action, and a score built on small samples moves on a handful of answers. A 2007 independent study did not replicate the claim that it predicts growth better than other loyalty questions, so use it to track sentiment and to find customers to call, and read the comments. ## The three groups ### Promoters Respondents who answer 9 or 10. They are the customers most likely to say a good word and to stay, and their comments show what the business does well enough to repeat. ### Passives Respondents who answer 7 or 8. They count in the total, and they count in neither group. They are satisfied but would not go out of their way, and a competitor offer can move them. ### Detractors Respondents who answer 0 to 6, which is seven of the eleven points on the scale. Some of them are mildly disappointed rather than angry. Their comments are the working list for service recovery and product fixes. ## What it is not - A satisfaction score. It asks about advocacy, and a customer can be content and still not recommend. - A percentage. It is a difference of two percentages, reported as a whole number between minus 100 and plus 100. - A figure you can compare directly with another company's. Industry, survey channel, and country change the typical level. ## The steps 1. **Fix the question and the timing.** Use the standard 0 to 10 recommendation question, followed by an open question asking why. Decide whether you are measuring the overall relationship, surveyed on a regular schedule, or a specific interaction, surveyed soon after it. Do not mix the two in one trend line. 2. **Define who is surveyed.** Survey the people who use the service, not only the buyer who signed. Set a minimum number of responses per segment. With 30 responses, one customer is 3.3 percent of the sample, and moving a single customer from detractor to promoter changes the score by about 7 points. 3. **Classify and compute.** Sort scores into promoters (9 and 10), passives (7 and 8), and detractors (0 to 6). Convert each group to a percentage of all responses, subtract the detractor share from the promoter share, and report the counts next to the score. 4. **Code the comments.** Group the open-text reasons by theme, separately for promoters and detractors, and by segment. A theme that shows up among detractors in one segment and nowhere else is a specific fix. A theme in every segment is a company-wide one. 5. **Close the loop and track.** Assign an owner to each recurring detractor theme, and contact detractors who agreed to be contacted. Track the score by segment over time, with counts and the sample size, so a change can be separated from noise. ## The formula NPS = percent promoters minus percent detractors The share of respondents scoring 9 or 10, minus the share scoring 0 to 6, reported as a whole number from minus 100 to plus 100. Passives sit in the denominator and nowhere else. - Promoters: respondents scoring 9 or 10 on the 0 to 10 recommendation question. - Passives: respondents scoring 7 or 8. They lower both percentages by adding to the total. - Detractors: respondents scoring 0 through 6. - Result: a score of +40 means promoters outnumber detractors by 40 points of the sample, not that 40 percent of customers are promoters. ## In practice The quarterly report for a managed-services client shows an NPS of +40 for two customer segments, and the account director reads that as parity. The counts say otherwise. Each segment returned 100 responses. Segment X had 60 promoters, 20 passives, and 20 detractors, so 60 percent minus 20 percent gives +40. Segment Y had 40 promoters, 60 passives, and 0 detractors, so 40 percent minus 0 percent also gives +40. X is polarized: one customer in five is unhappy, and those 20 comments are where to start. Y has nobody unhappy and little enthusiasm, so the question there is what would move passives to a 9 or 10. Same number, two different work plans. ## What you leave with - NPS by segment, with the counts of promoters, passives, and detractors shown beside each score. - A coded list of the reasons given by detractors and by promoters. - A named owner and a recovery or fix action for each recurring detractor theme. - A sampling and timing rule, so the next wave can be compared with this one. ## Failure points ### The score becomes a target When pay or rankings depend on it, front-line staff ask customers for a 9 or 10, or steer unhappy customers away from the survey. The score rises and the information in it falls. ### Small samples read as trends A branch with 25 responses has one respondent for every 4 percent of the sample. Moving one customer from detractor to promoter shifts its score by 8 points, so a quarterly swing can be one person. ### Comparing countries without adjustment People in different countries use a 0 to 10 scale differently, and response styles are a known issue in cross-country survey work. A gap between two markets can reflect the scale as much as the service. ### Reading the number and skipping the comments The score says where to look. The reasons carry what to do about it. A dashboard that shows only the score gets reviewed every quarter and changes nothing. ### Using it as a growth forecast The original claim was that the score predicts growth better than other loyalty measures. A 2007 Journal of Marketing study using Norwegian data from 21 firms did not replicate that. Treat it as a tracking measure and say so to the sponsor. The failures above are usually [Commercial Acumen](https://usefulconsultingframeworks.com/your-edge/#commercial): the work was not tied to money, for the client and/or for you. ## Origin Fred Reichheld, then a partner at Bain & Company, introduced the metric in "The One Number You Need to Grow," Harvard Business Review, December 2003. Satmetrix, a survey software firm, worked with Bain on the method, and Reichheld extended it in later books. In practice the score is almost always paired with an open question asking why the customer answered as they did. Net Promoter and NPS are registered trademarks, and Net Promoter Score and Net Promoter System are service marks, of Bain & Company, NICE Systems, and Fred Reichheld. Older Bain material names Satmetrix instead of NICE. Bain's published terms require attribution, and a license for commercial use. The description here is of the arithmetic and the practice only. ## Further reading - [The One Number You Need to Grow](https://hbr.org/2003/12/the-one-number-you-need-to-grow). Fred Reichheld, Harvard Business Review, December 2003. The original article. Metered paywall. - [A Longitudinal Examination of Net Promoter and Firm Revenue Growth](https://doi.org/10.1509/jmkg.71.3.039). Keiningham, Cooil, Andreassen, and Aksoy, Journal of Marketing, 2007. The most cited academic challenge to the growth claim. The abstract is free, and the full text is behind the publisher's paywall. - [Trademarks and Licensing](https://www.netpromotersystem.com/resources/trademarks-and-licensing). Bain's own statement of who owns the marks and when a license is needed. Read it before you put the name in a client deliverable. ## Related frameworks - [Marketing Funnel](https://usefulconsultingframeworks.com/frameworks/marketing-funnel/): Shows where prospects drop out before they become customers. NPS only hears from people who already bought, so pick the funnel for acquisition questions. - [Brand Equity](https://usefulconsultingframeworks.com/frameworks/brand-equity/): Treats loyalty as one component of what a brand is worth. Pick it when the question is why customers pay more or stay, not how many would recommend. - [Customer Lifetime Value](https://usefulconsultingframeworks.com/frameworks/customer-lifetime-value/): Puts a dollar value on keeping a customer. Pick it when the question is what a promoter or a detractor is worth, not how many of each you have. Last updated October 19, 2026.