The number is useful because it is simple, portable across industries, and cheap to collect, but it is only honest when teams pair it with a follow-up question and a closed-loop process that reaches the customers who answered.
NPS, or Net Promoter Score, is a loyalty metric built on one question: how likely a customer is to recommend a company to a colleague, scored 0 to 10. Respondents are grouped into Promoters (9-10), Passives (7-8), and Detractors (0-6). The score equals the percent of Promoters minus the percent of Detractors, giving a result from negative one hundred to positive one hundred.
The six things to understand before anyone on the team quotes an NPS number in a board deck or a QBR.
The classic wording is "How likely are you to recommend us to a colleague or friend?" on a 0-10 scale, where 0 is not at all likely and 10 is extremely likely. One question, one scale. The simplicity is the point: it survives translation, works across industries, and takes seconds to answer.
Scores of 9 and 10 are Promoters, loyal enough to actively recommend you. Scores of 7 and 8 are Passives, satisfied but unenthusiastic and easy for a competitor to peel off. Scores of 0 through 6 are Detractors, unhappy customers who can damage the brand through word of mouth or churn.
Add up Promoters, divide by total responses, express as a percent. Do the same for Detractors. Subtract. The result lands from negative one hundred to positive one hundred. Passives are intentionally ignored in the math, which is why a company with mostly 7s and 8s can still post a modest score.
A positive score means more Promoters than Detractors, a decent bar on its own. Industry benchmarks put software and SaaS averages around thirty, retail banking around forty, and travel categories like airlines spread widely. A score only means something against your own history and your category.
NPS breaks down when teams treat it as a target to game instead of a reading to learn from. Reps nag happy customers for 10s. Surveys fire right after onboarding and miss churn-risk accounts. Cultural response bias skews scores by region. The number drifts from the experience it is supposed to measure.
A follow-up question ("What is the main reason for your score?") turns a number into coachable text. A closed-loop process routes Detractors to a human within forty-eight hours, Passives into nurture, Promoters into advocacy. The score is diagnostic. The follow-up and the response are the medicine.
The math fits on a napkin. The mechanics around the math (what question goes to whom, when, through what channel, and what happens after they answer) are what separate a program that earns a slide in the board deck from a program that gets quietly archived.
The standard wording is "How likely are you to recommend our company to a colleague or friend?" on a 0-10 scale. The subject matters. "Our product," "our service," and "our company" produce different scores, so a program picks one phrasing and holds it steady across waves.
Scores split into three groups. Nine and ten are Promoters. Seven and eight are Passives, satisfied but not committed. Zero through six are Detractors. The bands are deliberately asymmetric: the Detractor band is wide because a 6 is a warning sign, not a mild compliment.
If one hundred responses come in and sixty are Promoters, twenty Passives, twenty Detractors, the score is sixty minus twenty, which equals forty. Passives count in the denominator but not the subtraction, so a shift from 7 to 8 does not move the score.
Every serious program adds an open-text question right after the score. "What is the main reason for your rating?" turns a rating into coachable language. The verbatim is where product finds real feature requests, support finds recurring pain, and sales finds the words customers actually use.
A survey sent to every customer at month-end produces one kind of signal. A survey sent to a new buyer three days after onboarding produces a different kind. Programs decide the sample frame in advance and document it, so quarter-over-quarter comparisons stay honest.
A rating with no follow-through is a measurement exercise, not a loyalty program. The best systems create a case the moment a Detractor responds, route it to an owner, start an SLA clock, and track whether that customer later moved up a tier. The score is a leading indicator only when action follows.
Leadership wants a single comparison: are we above or below average? That has an answer, but the answer carries footnotes. Benchmarks shift by industry, region, channel, and sample skew. A number is meaningful against your own baseline and against companies that measure the same way you do.
A score above zero means more Promoters than Detractors. In many categories that alone puts you ahead of the median. Below zero means Detractors outnumber Promoters, which is a serious signal regardless of industry. Celebrating a score of three is premature. Treating a score of negative five as a crisis is reasonable.
Published research puts the typical software or SaaS score in the high twenties to low thirties. The best cloud companies post scores in the fifties and above. If your SaaS score is in the single digits, the gap between you and the category leaders is informational, not just competitive.
Retail banking benchmarks tend to run around forty on average, with top-tier regional and digital-first banks posting higher. Industries with embedded, repeated use (banks, utilities, telecom) often see higher raw scores than infrequent-purchase categories, which affects interpretation.
Airline NPS swings dramatically between carriers in the same market and between seasons at the same carrier. Weather, loyalty-program design, and a single viral incident move a quarterly score more than a service improvement can. In high-variance industries, trend matters more than any snapshot.
Response patterns vary by region. Some cultures reserve 10s for extraordinary experiences, which depresses raw scores. Others score generously by default. A global company comparing a Nordic subsidiary to a North American one by raw NPS alone will reach the wrong conclusion. Normalizing by region is standard.
The most useful comparison is you, last quarter, measured the same way. Industry benchmarks are directional. Vendor-reported scores from case studies are marketing. Your own trend, with the same question, the same sample frame, and the same channel, tells you whether the thing you changed worked.
NPS has fans and detractors in roughly the proportions it measures. Taking the critiques seriously is what separates a program that improves the business from one that produces a slide and nothing else. Most objections are legitimate and most are solvable with better survey design, not by abandoning the metric.
Reducing a customer relationship to one 0-10 answer loses detail that renewal rate, expansion revenue, support volume, and feature adoption would surface. A smart program treats NPS as one gauge, not the dashboard. The score answers "how customers feel" and leaves "why" and "what next" to adjacent signals.
In some countries a 7 is enthusiastic praise. In others it is damning. Response-style bias is documented, and comparing a global average to an industry benchmark without segmenting by region produces misleading conclusions. Programs that operate across markets segment or normalize before publishing.
Send the survey after a successful onboarding call and the score skews high. Send it after a billing dispute and the score skews low. Programs that do not control for timing end up measuring their communications calendar more than their customer relationships.
When NPS is tied to compensation, there is pressure to ask only happy customers, prime with hints, or chase 10s. The number goes up and the signal degrades. Governance (sample selection removed from the measured team, verbatim audits, response-rate monitoring) keeps data honest.
A moving NPS line is interesting. It is not actionable until the open-text follow-up, segment breakouts, and journey-stage breakouts are overlaid. The score is a smoke alarm. The follow-up questions and segment data are the fire investigation.
Studies on the correlation between NPS and growth are mixed. In some industries the link is strong, in others noisy. A company that leans on NPS as the single north star without validating the link against its own retention and expansion data is making a bet, not reading a measurement.
There are two main shapes of NPS program, and mature teams run both. The infrastructure to send the survey is cheap. The infrastructure to act on it (route Detractors to humans, nurture Passives, mobilize Promoters) is where a CRM earns its place.
A tNPS survey fires right after a defined interaction: onboarding completion, a support resolution, a renewal, a product release. The score reflects how that interaction landed. Great for operational tuning, because the number maps to a specific moment the team can change.
A rNPS survey goes to the broad customer base on a cadence (quarterly, biannually, annually). The score reflects the overall relationship, not a single moment. Great for leadership reporting and benchmark comparisons, because the sample frame is stable across waves.
A single company-level score hides more than it reveals. Breakouts by plan tier, industry, size, region, product line, and account owner show which segments drive the overall number. Where Detractors cluster is where the next intervention goes.
A good program routes any Detractor response to an owner within forty-eight hours, with the customer's record, verbatim, and prior history in one view. The call is a listening motion, not a sales motion, logged against the account, with the outcome tied back to whether the next score moved.
Promoters are already saying nice things. The job of the program is making it easy to say those things publicly: ask for a review, invite them to a reference call, nominate them for a case study, offer a referral path. Promoter outreach pays for the rest of the program.
Strkr surveys attach to the account record. Detractor responses auto-create a case and route to the account owner with an SLA. Passives drop into a nurture campaign. Promoters get an advocacy tag that feeds the referral workflow. Score, verbatim, and follow-through live on the timeline the sales team already reads.
Strkr sends NPS surveys, scores responses, routes Detractors to the account owner with an SLA, drops Passives into nurture, and flags Promoters for advocacy. Score, verbatim, and follow-through sit on the timeline the revenue team already works from.
NPS stands for Net Promoter Score. It is a customer loyalty metric built on one survey question, measured on a 0-10 scale, with respondents grouped into Promoters, Passives, and Detractors. The score itself is calculated by subtracting the percent of Detractors from the percent of Promoters.
Take the share of respondents who scored 9 or 10 (Promoters) as a percent of total responses. Take the share who scored 0 through 6 (Detractors) as a percent of the same total. Subtract Detractor percent from Promoter percent. Passives (7 or 8) count in the denominator but not the subtraction. The result lands between negative one hundred and positive one hundred.
Any score above zero means more Promoters than Detractors, which is a basic bar to clear. Beyond that, "good" depends on industry. Published benchmarks put software and SaaS around thirty, retail banking around forty, and industries like airlines across a wider spread. The most useful comparison is against your own baseline from prior waves, measured the same way.
NPS measures loyalty with a likelihood-to-recommend question, usually at the relationship level. CSAT (customer satisfaction) measures satisfaction with a specific transaction, usually with a "how satisfied were you" prompt. NPS is a broader loyalty signal. CSAT is a tighter operational signal. Mature programs use both: CSAT for day-to-day tuning, NPS for relationship-level reads.
Transactional NPS (tNPS) fires after a specific event like onboarding, a support resolution, or a renewal, and reflects how that event landed. Relational NPS (rNPS) goes out on a scheduled cadence (quarterly, biannually, annually) and reflects the overall relationship. Mature programs run both: tNPS for operational feedback, rNPS for leadership reads.
Yes, when used as one gauge instead of the only gauge. The common critiques (single-metric reduction, cultural bias, timing sensitivity, incentive gaming) are real, but each has a mitigation: pair the score with a follow-up, segment by region, standardize the sample frame, keep measurement independent of the measured team. The score misleads only when a program ignores those practices.
A CRM makes NPS actionable instead of just measurable. Responses attach to the customer record, so the account team sees score and verbatim alongside deals and activities. Detractors auto-create cases with SLAs. Passives drop into nurture. Promoters get flagged for advocacy. The score, the follow-up, and the human response all live on one timeline.
Relational NPS is typically quarterly or biannual, which gives enough signal to track trends without survey fatigue. Transactional NPS fires on each defined event, so cadence matches event volume. Response rates fall fast when customers get the same survey too often, so most programs stagger samples and cap how many surveys any one customer receives.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
