NPS Benchmarks by Industry: Complete Guide

An NPS of 40 means nothing on its own. In one sector it puts you near the top quartile; in another it is unremarkable; in a third it may be structurally impossible to reach. The number only acquires meaning when it is placed against something, and choosing what to place it against is where most benchmarking goes wrong.

An NPS benchmark is that reference point: the score a defined group of companies, usually an industry or a region, typically reaches, offered as the line your own result should be read against. The complication is that published benchmarks are far less comparable than they appear. Different studies survey different populations, use relationship or transactional framing, sample at different points in the customer lifecycle, and pool respondents from countries with materially different rating habits. Two sources will give you two answers for the same industry, and neither is wrong.

This guide covers what the current benchmarks show, how to use them without over-reading them, and why your own trend line is a more reliable comparison than any published figure.

Key Takeaways

Here is a short summary of what this guide covers.

• What NPS measures, how it is calculated, and what counts as a good score in absolute terms.
• Current industry benchmark figures, with the methodological caveats that determine how much weight they deserve.
• How to benchmark properly, including which comparisons are valid and which are not.
• The factors that shift benchmarks: business model, region, survey timing, and channel.
• How to convert the score into action, how NPS compares with CSAT, and the mistakes that produce false confidence.

What is Net Promoter Score (NPS) and Why Does It Matter?

Net Promoter Score is a loyalty metric based on one question: how likely a customer is to recommend the organization to a friend or colleague, answered on a 0 to 10 scale. Developed by Fred Reichheld with Bain & Company and Satmetrix in 2003, it has become the most widely deployed customer metric in business and among the most frequently misapplied.

Its value lies in three properties. It is a single number, which makes it reportable to boards and comparable over time. It is widely used, which means external reference points exist. And it asks about advocacy rather than satisfaction, which sets a higher bar: recommending something to a colleague carries reputational risk in a way that rating an interaction highly does not.

Its limitations are equally structural. A single item cannot tell you why the score is what it is, which is why the follow-up question matters more than the number. It is volatile in small populations. And because it is relationship-level, applying it after every transaction produces noise rather than signal. NPS is best treated as one input among several rather than the metric that replaces the rest.

How is NPS Calculated?

• Ask the standard question. How likely are you to recommend us to a friend or colleague, on a 0 to 10 scale. Keep the wording consistent, because changes to it break comparability with your own history.
• Group the responses. Promoters score 9 to 10, passives 7 to 8, detractors 0 to 6. The asymmetry is deliberate: a 6 is treated as a detractor because passive tolerance is not advocacy.
• Convert each group to a percentage of total responses. Passives count in the denominator even though they do not appear in the calculation, which is the step most often got wrong.
• Subtract detractors from promoters. With 420 promoters, 300 passives, and 280 detractors out of 1,000 responses, that is 42 percent minus 28 percent, giving an NPS of +14.
• Report it as a number, not a percentage. The scale runs from -100 to +100. Writing “NPS of 14 percent” is a category error that appears in a surprising number of board decks.
• Record your sample size and response rate alongside it. A score without these is uninterpretable, particularly below a few hundred responses where normal variation looks like movement.
• Always pair it with an open follow-up. Asking the main reason for the score is what converts a number into something actionable, and a ready-made NPS survey includes it by default.

What is a Good NPS Score?

In absolute terms, the conventional thresholds trace back to Bain: anything above 0 means promoters outnumber detractors, above 30 is generally considered good, above 50 excellent, and above 70 world-class. These are useful orientation and poor targets.

The reason they are poor targets is that the same score sits in completely different percentiles depending on sector. Aggregated benchmark data puts the cross-industry average around 32 with a median closer to 44, and the gap between those two figures is itself informative: a handful of very low-scoring sectors pull the mean down, which means the average is not describing a typical company.

There is also a persistent structural gap between business models. B2B organizations tend to score differently from B2C ones, driven by relationship depth, account management, and who in the customer organization actually receives the survey. Comparing a B2B score against a B2C benchmark is one of the most common analytical errors in this area.

The honest answer to “what is a good NPS” is that a score improving consistently against your own prior periods, measured the same way each time, is better evidence of health than any single reading placed against a published average.

NPS Score Benchmarks by Industry

The figures below are drawn from Retently’s 2026 benchmark analysis, which aggregates responses across a large panel of organizations, alongside Bain’s NPS Prism research. Treat them as directional context rather than targets, and read the caveats underneath the table before using any of them in a goal.

Three caveats that matter more than the numbers. First, sources disagree substantially: another widely cited study places B2B software closer to 30, and the difference is methodological rather than a matter of one being wrong. Second, most published benchmarks pool relationship and transactional NPS, which are not the same measurement. Third, none of these studies sampled your customers, your segments, or your survey timing.

How to Benchmark Your NPS Score?

• Establish your own baseline first. Three or four consistent periods of internal measurement is worth more than any external figure, because it is the only comparison where methodology is genuinely constant.
• Identify your true peer set. Not your broad sector but organizations with a similar business model, customer type, and price point. A regional B2B services firm has more in common with another regional B2B services firm than with its industry’s consumer giants.
• Check the methodology behind any benchmark you use. Relationship or transactional, sample size, geography, survey channel, and the point in the lifecycle at which it was asked. If the source does not state these, treat the figure as trivia.
• Match the framing. Compare relationship NPS against relationship benchmarks only. Transactional NPS runs on a different distribution entirely.
• Segment before comparing. Your enterprise and SMB populations may differ by 30 points, and a blended number matched against a blended benchmark tells you nothing about either.
• Consider a competitive benchmark study. Surveying your competitors’ customers with the same instrument is the only genuinely comparable external measurement available, and it is what the more rigorous benchmark providers actually do.
• Set targets from your own trajectory. A realistic goal is a defined improvement on your last period, not arrival at an industry average measured by someone else.

Key Factors That Affect Your NPS Score Benchmark

The same organization can produce meaningfully different scores depending on decisions that have nothing to do with customer loyalty. Understanding these is what separates benchmarking from numerology.

• Business model. B2B and B2C populations rate differently, and within B2B, transactional and account-managed relationships diverge again.
• Survey timing. Asked right after a successful onboarding, scores run high. Asked after a renewal invoice, they run lower. Same customers, different number.
• Relationship versus transactional framing. Relationship NPS asks about the overall relationship; transactional NPS asks after a specific interaction. They are separate metrics that share a question.
• Channel. Email, in-app, and phone-administered surveys produce different distributions, with interviewer-present modes typically scoring higher.
• Geography and culture. Rating conventions vary substantially by country, enough to swing a multinational’s score without any change in experience.
• Sample size. Below roughly 100 responses, a handful of answers moves the score several points.
• Who receives it. In B2B, surveying the day-to-day user and surveying the economic buyer produce different scores from the same account.
• Response rate and non-response bias. Customers who are leaving frequently do not answer, which inflates every reading.
• Whether the score is a target. Any metric attached to compensation gets managed, including by choosing who to survey and when.

Why NPS Benchmarks Differ Across Industries?

Sector differences are usually structural rather than a straightforward reflection of service quality. Understanding the mechanism helps you judge whether your gap to a benchmark is a problem or a category feature.

Switching cost is the largest factor. Where changing provider is difficult, as in enterprise software or banking, customers stay through irritations they would not tolerate elsewhere, and relationship depth tends to produce higher advocacy scores. Where switching is trivial, as in consumer subscriptions, scores are more volatile and generally lower.

Emotional valence matters too. Categories people enjoy, such as leisure and premium consumer goods, start from a warmer baseline than categories associated with obligation, such as utilities, insurance claims, or internet service. Nobody enthusiastically recommends their internet provider, which puts a ceiling on the entire sector.

Regulation and constrained choice compress scores in another direction. Where pricing, features, and interactions are constrained by rules, providers have less room to differentiate, and customers rate against an experience they had no ability to shape.

Finally, contact frequency shapes exposure. A provider a customer interacts with daily has many more opportunities to disappoint than one they engage with annually, which is part of why some low-touch B2B categories post scores that would be implausible in retail.

How to Turn Your NPS Score Into Action

The score itself changes nothing. What follows it does, and there are two distinct workstreams that need to run at once: recovering individual detractors, and fixing the pattern that produced them.

• Route detractor responses within 48 hours. Any 0 to 6 with an open comment becomes an assigned, tracked follow-up. Closed-loop workflows and ticketing are what stop this depending on someone reading a dashboard.
• Theme the open-text responses. Group the stated reasons into recurring themes and count them. The theme mentioned most often frequently differs from the lowest-scoring touchpoint, and both matter.
• Segment the score before diagnosing. By tier, tenure, product, region, and channel. A flat company NPS routinely conceals two populations moving in opposite directions.
• Correlate themes with behavior. Which themes appear disproportionately in accounts that churned or downgraded? That is the priority list, not the volume ranking. Customer analytics is where the two views meet.
• Locate the problem in the journey. NPS tells you the relationship is weakening without telling you where, which is what journey-level measurement supplies.
• Act on promoters too. They are the referral, testimonial, and case study pipeline, and most organizations collect the score and never ask.
• Commit to two or three changes with owners and dates, then report progress including where an action failed.

All In Credit Union applies structured member feedback across branches and transaction types, sustaining renewal above 97 percent, which is the behavioral confirmation that a loyalty score is supposed to predict.

Ready to turn your score into something you can act on? Request a demo → and see how SogoCX handles NPS collection, segmentation, and closed-loop follow-up.

How to Compare Your NPS Score With Industry Benchmarks?

• Confirm the benchmark’s framing matches yours. Relationship against relationship, transactional against transactional. Mixing them invalidates the comparison before you start.
• Check the sample composition. Business model, company size, geography, and customer type. A benchmark pooled across 18 countries is not comparable to a single-market score.
• Match the survey channel where possible. Mode effects are real and can account for several points.
• Compare like segments, not company averages. Your enterprise segment against enterprise benchmarks, if such a figure exists.
• Look at the distribution, not just the average. Where a benchmark reports quartiles, your position in the range is more informative than the gap to the mean.
• Use multiple sources and expect them to disagree. Where two credible studies differ by 15 points for the same sector, that spread is the honest confidence interval.
• State the caveats when you present it. A benchmark comparison offered without methodology notes will be quoted back as fact for years.

How to Interpret Your NPS Benchmark?

Interpretation is mostly about resisting two temptations: treating a gap as a verdict, and treating a match as a pass.

A score below the benchmark is a prompt to investigate, not a conclusion. Before accepting that you underperform, check whether the benchmark measured the same framing, the same customer type, and the same geography, and whether your own sample is large enough to be stable. A substantial share of apparent gaps dissolve on inspection.

A score at or above the benchmark is equally unsafe to rest on. Industry averages include organizations losing customers steadily, so matching one is not evidence of health. More useful questions are whether your score is improving, whether the spread between your best and worst segments is narrowing, and whether your detractors are being recovered.

The most informative reading combines three things: your trend over the last several periods, the variation between your own segments, and whether the score tracks actual behavior. If your promoters and detractors renew at similar rates, the metric is not measuring what you think it is, and no benchmark comparison will fix that. Our CX Experience Index 2026 covers the wider expectation shifts worth reading any score against.

How to Improve Your NPS Score?

• Fix the driver, not the score. Managing to the number produces a better number and the same underlying problems.
• Reduce effort before adding delight. Removing friction reliably outperforms exceeding expectations, particularly after a service failure.
• Close the loop on every detractor. It is one of the few interventions with a measurable effect on an individual relationship.
• Attack repeat contacts. Every additional contact about the same issue compounds dissatisfaction and cost simultaneously.
• Fix the worst touchpoint rather than the average. A single bad stage drags the whole relationship score regardless of how good the rest is.
• Convert passives deliberately. They are closer to promoters than detractors are, and moving them is usually the cheapest available gain.
• Send causes upstream. Support volume is generally created by product, billing, or documentation decisions made elsewhere.
• Ask promoters for something. Referrals, reviews, and testimonials convert a score into pipeline, and public reviews compound. Virginia Physicians for Women found that asking at the right moment produced a sixfold increase in positive reviews and a rating lift from 3.5 to 4.5 stars.
• Tell customers what changed. Visible follow-through improves both the next score and the next response rate.

NPS vs. CSAT: Which Metric Should You Benchmark?

For benchmarking specifically, NPS has the advantage of volume: more organizations publish it, so more external reference points exist. For improvement work, CSAT is the better instrument because it attaches to something a specific team can change this quarter.

Most mature programs run both, plus an effort measure, and are clear about which does which job. The failure mode is running NPS alone and expecting it to explain why the number moved.

Common NPS Score Benchmarking Mistakes to Avoid

• Comparing across industries. The single most common error, and it produces confident wrong conclusions in both directions.
• Mixing relationship and transactional scores. Different measurements sharing a question.
• Ignoring sample size. Celebrating a five-point move built on 60 responses.
• Reporting NPS as a percentage. It is an index, and the error signals that nobody in the chain understands the metric.
• Comparing across geographies without adjustment. Rating conventions differ enough to swing a score by tens of points.
• Using a single source as truth. Credible studies disagree substantially, and picking the flattering one is a decision rather than an analysis.
• Benchmarking the company average. It conceals the segment where the problem actually lives.
• Treating an industry average as a target. Averages include declining organizations.
• Changing the question wording or scale. This resets your own trend, which was the more reliable benchmark.
• Tying the score to compensation. Guarantees it will be managed rather than improved.
• Benchmarking without acting. The comparison is not the work.

NPS Benchmarking Best Practices

• Keep the question wording, scale, and survey timing identical across every period.
• Record sample size, response rate, and framing alongside every score you report.
• Treat your own trend as the primary benchmark and external figures as context.
• Segment by tier, tenure, region, and product, and report segments rather than only the aggregate.
• Always pair the score with an open follow-up asking the main reason.
• Verify the score predicts behavior in your own data before relying on it.
• Use multiple benchmark sources and present the spread rather than a single figure.
• Set improvement targets from your own trajectory, not from an industry average.
• Route detractors for follow-up within 48 hours and measure adherence to that standard.
• Review the measurement design annually, since touchpoints and channels change faster than survey programs.
• Pair NPS with a transactional and an effort metric so you can explain movement rather than only observe it.

A structured Voice of the Customer program is what holds this together, since the benchmarking discipline is worth little if the collection and follow-up mechanics are inconsistent.

FAQs About NPS Benchmarking

Can NPS benchmarks vary by business model?

Substantially. B2B and B2C populations rate differently, driven by relationship depth, account management, and switching cost, and published averages for the two differ by roughly ten points. Within B2B, account-managed relationships score differently from self-serve or transactional ones. Subscription businesses differ from one-time purchase businesses because the renewal decision changes how customers evaluate the relationship. Always match the benchmark’s business model to yours before drawing any conclusion.

Can NPS benchmarks vary by customer segment?

Yes, and internal segment variation frequently exceeds the gap between your company and an industry average. Enterprise and SMB customers, new and tenured customers, and customers on different products or price tiers can differ by 20 to 30 points within the same organization. This is why a blended company score compared against a blended benchmark is close to uninformative: it hides the segment that is actually driving churn.

Should NPS benchmarks be based on competitors or industry averages?

Competitor benchmarks are more useful when you can get them, because they compare you against the alternatives your customers actually consider rather than against a pooled sector average that includes organizations serving entirely different markets. The practical obstacle is that genuine competitive NPS requires surveying your competitors’ customers with the same instrument, which means a dedicated study. Industry averages are the fallback, and they are best used as a plausibility check rather than a target.

How can you find reliable NPS benchmark data?

Judge a source by what it discloses rather than by the specificity of its numbers. A reliable benchmark states its sample size, the composition of that sample, the geography, the survey channel, whether the scores are relationship or transactional, and the collection period. Sources that publish a clean table of industry figures with no methodology section are aggregating other people’s data, often across incompatible studies. Where a source cannot tell you how the number was produced, treat it as background rather than evidence.

Where can you find industry NPS data?

The main categories are research firms such as Bain and its NPS Prism benchmarking product, survey platform providers who publish aggregated data from their own customer base, syndicated studies including satisfaction indices that report adjacent metrics, and industry associations that occasionally run sector-specific research. Analyst firms publish periodic CX benchmarking as well. Each carries the bias of its sample: platform-published benchmarks reflect that platform’s client mix, which tends to skew toward organizations already investing in CX measurement.

How reliable are published NPS benchmarks?

Less reliable than their presentation suggests. Credible sources routinely disagree by 10 to 15 points for the same sector, because they sample different populations, mix relationship and transactional framing, pool across countries with different rating conventions, and draw from self-selected client bases. This does not make them useless, but it does mean a single published figure should never be treated as the standard you are measured against. Where several sources are available, the spread between them is a more honest representation than any one of them.

How often do NPS industry benchmarks change?

Most published benchmarks update annually, and year-over-year movement at industry level is usually modest, in the range of a few points. Larger swings do occur, particularly in sectors affected by a shock or by rapid change in customer expectations. The more important point is that a change in a published benchmark often reflects a change in the study’s sample composition rather than a real shift in customer sentiment, which is another reason to weight your own consistent trend more heavily.

How can historical NPS data be used for benchmarking?

It is the most reliable benchmark available to you, because methodology, customer base, and survey timing are constant in a way no external comparison can match. Use it three ways: compare the current period against the same period last year to remove seasonality, track the trend across at least four periods so you can distinguish signal from normal variation, and monitor the spread between your highest and lowest scoring segments, since a narrowing spread is often better evidence of progress than a rising average. The prerequisite is that the question wording, scale, and timing have not changed.

Should you use internal or external NPS benchmarks?

Both, for different purposes. Internal benchmarks tell you whether you are improving, which is the question that determines whether your work is producing anything. External benchmarks tell you whether you are competitive, which matters for positioning and for setting expectations with a board. When the two conflict, weight the internal one, since it is the comparison where the methodology is genuinely constant. An organization improving steadily against its own history while sitting below a published industry average is in better shape than one matching the average while declining.