Net Promoter Score (NPS): The Complete Guide

NPS was introduced in 2003 by Fred Reichheld of Bain & Company, working with Satmetrix, in a Harvard Business Review article titled 'The One Number You Need to Grow.' The idea was deliberately simple: instead of long satisfaction questionnaires that few people finish, ask a single question that predicts whether customers will stick around and bring others with them. Recommending a company to a friend puts your own reputation on the line, so a high willingness to recommend is a strong signal of genuine loyalty.

That simplicity is why NPS spread so fast. One question is easy to answer, easy to benchmark and easy to explain to a board. It became a common language for loyalty across industries — but the simplicity is also its trap: the score alone tells you where you stand, not what to fix. The value comes from pairing it with a follow-up question and reading the reasons behind the number.

Every 0–10 answer falls into one of three groups. The buckets are fixed and non-negotiable — this is what makes NPS comparable everywhere:

Notice how uneven the buckets are: promoters take just two values, while detractors span seven. That is intentional. NPS treats a 6 as a warning sign, not a passing grade — a customer has to be genuinely enthusiastic (a 9 or 10) to count as a promoter. Passives count toward your total response base but do not add to or subtract from the score.

The formula is a subtraction of two percentages:

NPS = % Promoters − % Detractors

Passives are excluded from the subtraction but included when you calculate the percentages, because the percentages are always out of the total number of responses. The result is a whole number between −100 (everyone is a detractor) and +100 (everyone is a promoter). NPS is expressed as a number, not a percentage — a score of 40, never '40%'.

1. 1Count how many responses fall into each group: promoters (9–10), passives (7–8) and detractors (0–6).
2. 2Divide promoters by total responses to get % promoters. Divide detractors by total responses to get % detractors.
3. 3Subtract % detractors from % promoters. Round to the nearest whole number.

Worked example 1. You collect 200 responses: 120 promoters, 50 passives and 30 detractors. Promoters are 120 ÷ 200 = 60%. Detractors are 30 ÷ 200 = 15%. NPS = 60 − 15 = 45.

Worked example 2. You collect 500 responses: 150 promoters, 200 passives and 150 detractors. Promoters are 30%, detractors are 30%, so NPS = 30 − 30 = 0 — even though only 30% are unhappy, the large passive middle and equal detractors leave you at zero.

Worked example 3. A struggling product gets 80 responses: 10 promoters, 20 passives and 50 detractors. Promoters are 12.5%, detractors are 62.5%, so NPS = 12.5 − 62.5 = −50. The two passives-heavy and detractor-heavy examples show why the same 'number of unhappy people' can produce very different scores. Skip the arithmetic entirely with the NPS calculator — enter your counts and it returns the score and distribution.

The honest answer is: it depends. As a rough, universal rule of thumb, any positive score means you have more promoters than detractors, a score above 30 is generally considered good, above 50 is excellent, and above 70 is world-class. But those bands are only a starting point — the same number can be outstanding in one industry and mediocre in another.

• Poor — a negative score. Detractors outnumber promoters and word of mouth is working against you. This needs urgent attention.
• Average — roughly 0 to 30. You have more fans than critics, but plenty of room to grow. Most companies live here.
• Good — roughly 30 to 50. A solid, loyal base that is likely generating referrals.
• Excellent — 50 and above. Loyalty has become a real advantage; scores above 70 are world-class and rare.

Industry matters because customer expectations and switching costs differ. Sectors with strong emotional or brand relationships tend to post higher scores, while highly competitive or commoditized categories tend to run lower — not because those companies are worse, but because the baseline is different. Survey method matters too: a relationship survey sent by email to your whole base usually produces a different score than a transactional survey shown right after a support interaction.

For that reason the most useful benchmark is your own past score. A rising trend, measured the same way each time, tells you loyalty is improving regardless of where you sit against any published average. For a deeper look at ranges and why cross-company comparisons mislead, see the NPS benchmarks guide.

The rating question gives you the number; the follow-up gives you the reason. A single open-text follow-up placed right after the score is the highest-value question in the entire survey, because it turns an abstract number into specific, actionable feedback.

• The universal follow-up — 'What is the main reason for your score?' This one question works for everyone and should be in every NPS survey.
• For promoters — 'What do you love most about us?' and 'What would make you recommend us even more strongly?' to identify and amplify your strengths.
• For passives — 'What would it take to earn a higher score from you?' to surface the specific gap keeping them from becoming fans.
• For detractors — 'What went wrong, and what would fix it?' and 'What is the one thing we could improve?' to catch problems and reduce churn.

The most effective NPS surveys route the follow-up based on the score, so a promoter and a detractor see different, relevant questions. That conditional logic keeps the survey short while making the answers far richer. For a complete bank of wording by group, see NPS survey questions, and for follow-up strategy specifically, NPS follow-up questions.

A great NPS survey is short — often just the score plus one follow-up. Here is a complete, ready-to-use structure with light conditional logic:

1. 1How likely are you to recommend us to a friend or colleague? (0–10 scale, required) — the standard NPS question.
2. 2What is the main reason for your score? (open text) — shown to everyone; the single most valuable follow-up.
3. 3If detractor (0–6): What is the one thing we could do to improve your experience? (open text)
4. 4If passive (7–8): What would it take to earn a higher score from you? (open text)
5. 5If promoter (9–10): What do you value most about us? (open text)
6. 6May we follow up with you about your feedback? (yes/no, optional)

You do not have to build an NPS survey from scratch. A good starting template already includes the correctly-worded 0–10 question, the open-text follow-up and sensible defaults, so you can launch in minutes and customize from there.

• Start with the ready-made NPS survey template — the standard question plus follow-up, editable end to end.
• For a relationship-level loyalty program, adapt the customer loyalty survey, which goes beyond a single score.
• Browse the full template library for industry-specific NPS variants tailored to retail, SaaS, hospitality and more.

• Treating NPS as a target instead of a signal — gaming the score (begging for 9s and 10s) corrupts the data and hides real problems.
• Skipping the follow-up question — a score with no 'why' tells you where you stand but nothing about what to fix.
• Changing how you collect the survey and then comparing scores — a transactional and a relationship NPS are not the same measurement.
• Comparing your number to an external benchmark collected with a different method or in a different industry.
• Surveying customers too frequently — NPS fatigue lowers response rates and skews who answers.
• Reporting only the headline number and ignoring the distribution — a score of 30 built from many promoters and many detractors is very different from one built from mostly passives.
• Never closing the loop with detractors — the biggest retention wins come from following up individually, not from charting the score.
• Ignoring the open-text answers because they are harder to process than the number (AI removes this excuse — see below).
• Rounding or expressing NPS as a percentage — it is a whole number from −100 to +100.

The hardest parts of an NPS program have never been the arithmetic — they are writing a survey that actually gets answered and reading hundreds or thousands of open-text comments. Both are exactly what AI is good at.

• AI survey builder — describe your business and goal and get a complete NPS survey with the correct question, follow-ups and conditional logic in seconds.
• Sentiment analysis — the emotional tone of every written comment scored automatically, so you can triage at volume instead of reading one by one.
• Theme detection — recurring topics surfaced across all the open-text answers, turning scattered comments into a ranked list of what promoters love and what detractors want fixed.
• Detractor alerts — flagging low scores and negative comments quickly so your team can close the loop before an at-risk customer churns.

Together these turn NPS from a number on a dashboard into a working feedback loop: the score tells you the trend, and AI tells you the reasons and the actions. To go deeper, see how to improve NPS and detractor analysis.

Ready to start? Create your NPS survey with AI, calculate a score from existing data with the NPS calculator, or read the difference between loyalty and effort metrics in customer satisfaction vs NPS.