NPS · Net Promoter Score
Net Promoter Score
What is NPS?
- 'Would you recommend us to a friend?' on a 0–10 scale
- Calculated as % promoters minus % detractors
- Loyalty summed up in one number
What NPS really changes is not the scoreboard but the language of the meeting. Walk in with a satisfaction score of 4.6/5 and the meeting ends with 'We're doing great.' Say 37% promoters and 18% detractors, and the follow-up question comes immediately: 'Who are those 18%?' Because it subtracts the share of unhappy customers outright rather than averaging them in, a single 0 drags the score down even when nine others give top marks. Grouping every 0–6 as a detractor may look harsh, but it rests on the premise that 'a lukewarm customer won't mention our name to anyone.' That one person's sign of leaving doesn't get buried in the average is the job this metric does in the field.
In B2B, you need to run two types separately. Relationship NPS, which asks about the overall business relationship, runs once a quarter or half-year, while transactional NPS asks at specific events, such as right after a training session or a delivery. A small sample is another way B2B differs from B2C. With 40 accounts, a single response moves the score by 2.5 points, so before looking at decimals, check whether the respondent is the decision-maker for that account. Accounts often fail to renew when the day-to-day user gives a 9 and the executive who holds the budget gives a 5. Industry benchmark figures circulate too, but your own trend line, measured with the same questions at the same points in time, is far safer than comparing against someone else's score from a different industry with different response habits. If CSAT asks 'Was this experience good?' and CES asks 'Was getting this done easy?', NPS asks whether you would put your own reputation on the line.
The most common failure happens the moment NPS becomes an individual performance metric. When account managers ask, 'I only get credit if you give me a 9 or 10,' the score climbs by about 20 points, churn stays the same, and from then on nobody trusts the metric. It is safer to use the score only for diagnosis and keep it separate from rewards. The second misconception is ignoring the passive range of 7–8. Because passives drop out of the formula, they drop out of management attention too, yet the group that switches on a single competitor proposal is these passives, not the detractors. The third is reading non-responses as indifference, when the accounts that have already checked out are the ones least likely to answer a survey. An organization that looks only at the scorecard and never reads the original comments will hear the same complaints again six months later.
John KimAuthor of “B2B Marketing Design” and CEO of Nanum Consulting
Sangmi LeeInstructor specializing in CS, leadership, and communicationIntroduced by Fred Reichheld (Bain & Company) in his 2003 Harvard Business Review article 'The One Number You Need to Grow.' The key finding was that a single question, 'likelihood to recommend (0–10)', best predicts repeat purchases and word of mouth.
Consider a hypothetical case. Company A, a manufacturer with 900 employees, ran the same leadership program for three years, through its 13th cohort, and completion satisfaction held steady at 4.6/5 every time. Yet only 2 additional department requests came in over the year, and the Q4 budget review asked, 'If satisfaction is so high, why isn't it spreading?'
- Kim, a manager on the HR development team, cut the 30 questions from the 14th cohort's post-course survey and replaced them with a single one, 'Would you recommend this program to a colleague? (0–10)', then added an open-ended question right below it asking why they gave that score.
- Of 84 respondents, 31 were promoters (37%), 38 were passives (45%), and 15 were detractors (18%), for an NPS of 19. When Kim reported this alongside the 4.6 satisfaction score, the team lead asked, 'Which departments are those 15 in?'
- Of the 15 detractor reasons, 12 were 'the examples don't fit our processes,' and Park, a manager on the production engineering team, wrote in a comment, 'All of the instructor's examples were about IT companies.'
- Kim, the manager in charge, met the training provider's PM and replaced the cases in three modules with Company A's equipment-shutdown case. To the 31 promoters, Kim sent an email asking for 'a one-line review, plus one colleague to attend with them.'
- The 15th cohort's NPS rose 22 points to 41, voluntary department sign-ups grew from 2 to 9 per half-year, and the top detractor reason shifted to 'the amount of pre-work', so the next improvement priority became obvious.