What Is Net Promoter Score (NPS)? Definition, Formula, and the Why Behind the Score

What is Net Promoter Score (NPS)?#

Net Promoter Score (NPS) is a customer loyalty metric that measures how likely customers are to recommend a company, product, or service to a friend or colleague, calculated as the percentage of promoters minus the percentage of detractors on a single 0–10 survey question. It produces one number between -100 and +100 that leaders use as a shorthand for customer loyalty and a leading indicator of growth.

The net promoter score was introduced by Bain & Company consultant Fred Reichheld in a 2003 Harvard Business Review article, "The One Number You Need to Grow." Reichheld's research, conducted with Satmetrix, found that in 11 of 14 industries studied the "likelihood to recommend" question was the strongest survey-based predictor of a company's future growth. Two decades on, NPS is used by roughly two-thirds of the Fortune 1000, which makes it the most widely deployed loyalty metric in business — and, increasingly, the most argued-about.

This guide covers what NPS is, the formula and the exact math behind it, the difference between transactional and relational NPS, and the one thing the score structurally cannot tell you: why customers answered the way they did.

How does the NPS formula work?#

The NPS formula works by subtracting the percentage of detractors from the percentage of promoters: NPS = % Promoters − % Detractors. Passives are counted in the total response base but are deliberately excluded from the subtraction, which is why the score can range from -100 (every respondent is a detractor) to +100 (every respondent is a promoter).

Every respondent is sorted into one of three segments based on their 0–10 answer:

Here is a worked example. Say 100 customers respond: 50 give a 9 or 10, 30 give a 7 or 8, and 20 give a 0 through 6. Your promoters are 50% and your detractors are 20%, so your NPS is 50 − 20 = 30. Note that the 30 passives moved the denominator but never entered the calculation — two companies with identical NPS scores can have very different numbers of passives sitting quietly in the middle. For the step-by-step math, edge cases, and the errors that quietly inflate scores, see our companion guide on how to calculate your NPS score.

The 0–10 rating itself carries more logic than it looks — the reasons Bain chose an 11-point scale rather than 1–5 or 1–7, and how the three buckets are drawn, are worth understanding before you trust the output. We break that down in the NPS scale explained.

Is NPS a percentage?#

No — NPS is not a percentage, even though it is built from percentages. The score is an absolute number between -100 and +100, so a company should say its NPS "is 30," not "is 30%." This distinction matters because leaders sometimes compare their NPS against percentage-based metrics like customer satisfaction score (CSAT) and draw false conclusions from numbers that are not on the same scale.

What is the difference between transactional and relational NPS?#

The difference is timing and intent: relational NPS measures overall loyalty to your brand at a regular cadence, while transactional NPS measures satisfaction with one specific interaction right after it happens. Both use the same 0–10 question and the same formula, but they answer different questions and belong in different parts of your program.

Relational NPS is your board-level trend line — the number you report over time and benchmark against your industry. Transactional NPS is your operational feedback loop, firing after an onboarding call, a support ticket, or a renewal so you can tie the score to a moment you can actually fix. Bain & Company, which maintains the Net Promoter System framework, recommends running both: the relational survey for strategy and the transactional survey for closing the loop with individual customers. Where each touchpoint sits and which metric belongs at each one is the core of customer lifecycle management.

What is a good NPS score?#

A good NPS score is highly context-dependent: 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 these thresholds shift dramatically by industry. A 30 might be exceptional in a low-loyalty category and mediocre in a high-loyalty one, which is why comparing your number to a global average is close to meaningless.

Because benchmarks are so industry-specific, we keep a dedicated, regularly updated breakdown in what is a good NPS score, with 2026 benchmarks by industry. The short version: treat any benchmark as a starting line, not a diagnosis. Two companies at the same NPS can be on opposite trajectories, and the number alone will never tell you which one you are.

What NPS does not tell you: the "why" gap#

NPS tells you who your promoters and detractors are, but it never tells you why they feel that way — and the "why" is the only part you can act on. A score is a lagging summary of sentiment; it compresses a customer's entire experience into a single digit and discards every reason behind it. You can watch your NPS slide from 40 to 32 and have no idea whether the cause is a pricing change, a shipped bug, a support regression, or a competitor's new feature.

This is the structural limit of any single-number metric, and it applies equally to NPS, CSAT, and every other score in your customer experience metrics stack. Reichheld himself acknowledged the gap in his 2021 HBR follow-up, "Net Promoter 3.0," arguing that the score had been widely gamed and needed to be paired with harder evidence of customer behavior rather than treated as an end in itself.

Three practical problems flow from the "why" gap:

• Low, biased response rates. Email NPS surveys typically return 15–25% response rates, and many programs see 5–15%, which means your score often reflects your most extreme customers, not your median one. B2B relational surveys average around 12% response.
• The open-text box is thin. The follow-up "Why did you give that score?" field gets skipped, gets one-line answers, and never asks a follow-up question when a customer says something vague like "it's fine" or "the pricing."
• No probing. A static survey cannot ask "what specifically felt too expensive?" the way a human interviewer would. The most valuable moment — the messy, half-formed reason — is exactly where a form gives up.

The result is a program that is excellent at telling you the number moved and useless at telling you what to do about it. That is why so many teams are re-examining the metric entirely; we cover the shift in why traditional NPS surveys are not enough and in why product teams are sunsetting NPS in 2026.

How to capture the reason behind the score#

You capture the reason behind the score by replacing the static open-text box with a short conversation that follows up on each answer in the customer's own words. The score still anchors the interaction — someone gives you a 6 — but instead of a dead-end comment field, an AI interviewer asks "what would have made that a 9?" and keeps probing until the actual driver surfaces.

This is the wedge between measuring loyalty and understanding it. A survey flattens a customer into a dropdown; a conversation lets them explain the "it depends" and the "well, actually" that contain the roadmap. Running this at the scale of a relational NPS program — hundreds or thousands of respondents — used to require a research team you don't have. AI-moderated interviews change the economics: Perspective AI runs the follow-up conversation automatically for every respondent, so the detractor who scored you a 3 gets asked why, and their answer comes back as a themed reason you can route to product, support, or pricing.

Teams that make this switch stop treating the number as the deliverable and start treating the reasons as the deliverable. If you want the full method — question design, the single follow-up that matters, and how to run it at scale — see the conversational NPS survey method that captures the why and our practical guide to the NPS survey in 2026. For CX and product leaders standing up this loop, it's built for CX teams.

NPS vs. CSAT, CES, and CLV: how the metrics compare#

NPS measures loyalty and likelihood to recommend, while CSAT measures satisfaction with a specific interaction, CES measures how much effort a task required, and CLV measures the total revenue a customer generates — each answers a different question and none replaces the others. Choosing the right one depends on what decision you are trying to make.

The pattern across the table is the recurring theme of this cluster: every metric is a strong quantitative signal and a weak explanation. NPS pairs naturally with customer sentiment as an emotional read and with customer lifetime value (CLV) as a financial one, and all three feed customer retention. For the full picture of how the score fits into the broader discipline, start with our pillar on what customer experience (CX) is. And if you're weighing the model shift directly, survey-based CX measurement vs. conversational VoC makes the case in full.

Frequently Asked Questions#

What is a Net Promoter Score in simple terms?#

A Net Promoter Score is a single number, from -100 to +100, that tells you whether more of your customers would recommend you than would warn people away. You ask everyone one question — "how likely are you to recommend us, 0 to 10?" — then subtract the percentage of low scorers (0–6) from the percentage of high scorers (9–10). A higher number means more loyal customers.

How is NPS calculated?#

NPS is calculated as the percentage of promoters minus the percentage of detractors, ignoring passives in the subtraction. Promoters rate you 9–10, passives rate 7–8, and detractors rate 0–6. If 60% of respondents are promoters and 15% are detractors, your NPS is 45. The passives still count toward your total response base but do not add to or subtract from the score itself.

What is a good NPS score?#

A good NPS score is generally above 30, with above 50 considered excellent and above 70 world-class, but the right benchmark depends heavily on your industry. Some categories run structurally low and others structurally high, so comparing against a peer benchmark matters far more than any global average. Any positive score means promoters outnumber detractors.

What's the difference between NPS and CSAT?#

NPS measures long-term loyalty and likelihood to recommend your brand overall, while CSAT measures satisfaction with a single, specific interaction. NPS uses a 0–10 scale and produces a number from -100 to +100; CSAT usually uses a 1–5 scale and is reported as a percentage of satisfied respondents. Most mature programs track both, using NPS for strategy and CSAT for operational quality.

Why is NPS criticized?#

NPS is criticized because a single number hides the reasons behind it, response rates are low and skewed toward extreme opinions, and the metric is easily gamed when it becomes a target tied to bonuses. Its creator, Fred Reichheld, addressed several of these issues in "Net Promoter 3.0," urging companies to validate the score against real customer behavior rather than treating the number as the goal.

How often should you measure NPS?#

You should measure relational NPS quarterly or twice a year to track your loyalty trend, and fire transactional NPS immediately after key interactions such as onboarding, support resolution, or renewal. Surveying too frequently causes fatigue and depresses response rates; surveying too rarely means you spot problems long after they cost you customers.

Turning your Net Promoter Score into action#

Net Promoter Score earned its place as the default loyalty metric for a reason: one honest question, one comparable number, a proven link to growth. But the score was always meant to be the start of a conversation, not the end of one. It tells you precisely who is loyal and who is at risk — and nothing about why, which is the only lever you can actually pull.

The teams getting the most out of NPS in 2026 keep the number for its comparability and add a layer that recovers the reasons: a short, AI-moderated conversation that follows up on every score and comes back with themed, act-on-able drivers instead of a thin comment field. If your NPS program produces a chart nobody knows how to move, that missing "why" is the gap to close. You can start a conversational study against your own customers, or see how it fits your stack on our pricing page. The score tells you the temperature; the conversation tells you what's causing the fever.