NPS benchmarks by industry: typical ranges across 12 sectors. What the median tells you, what it does not, and the better benchmark almost no team uses.
A 35 is average in B2B SaaS, top-quartile in airlines, and below median in consumer electronics. Industry benchmarks are useful for sanity checking and useless as targets — the same number means three different things in three different sectors. The benchmark that actually moves your program is your own prior wave on the same customers. Sopact tracks NPS as a trajectory per customer across waves, with every verbatim attached — so the score arrives with the reason and the comparison underneath.
By Unmesh Sheth · Founder & CEO, Sopact · Updated May 2026
An NPS benchmark is a reference NPS score, usually a median or quartile range, drawn from a published industry study or internal historical data. The benchmark answers where does this score sit relative to others. The benchmark does not answer what to do about it — that question lives in the verbatim that arrives with each response.
The typical NPS for companies in a given sector, published by sources like Bain, Satmetrix, Qualtrics, Forrester. Useful for sanity checking — is the program in the expected range for the sector. Misleading as a target.
Your own score from one quarter ago on the same customer base. The most useful comparison the team has — whether the program is improving on the people who actually use it.
A customer's score this wave compared to their own score last wave. The strongest single signal in NPS: a customer who scored 9 last wave and 4 this wave is the converted detractor every save-call program looks for.
Approximate medians and quartile ranges across twelve sectors, drawn from published benchmarking studies. Use the numbers as orientation, not as targets. The right column is the one that matters.
| Industry | Bottom quartile | Median | Top quartile | Leading |
|---|---|---|---|---|
| B2B SaaS / software | 10-20 | 30-40 | 50-65 | 70+ |
| Consumer SaaS / apps | 20-30 | 40-55 | 60-75 | 75+ |
| Retail (general) | 25-35 | 40-50 | 55-70 | 75+ |
| Consumer electronics | 25-35 | 45-60 | 65-75 | 75+ |
| Financial services (B2C) | 0-15 | 20-40 | 45-60 | 65+ |
| Insurance | 0-15 | 20-35 | 40-55 | 60+ |
| Healthcare (providers) | 5-20 | 20-40 | 45-60 | 65+ |
| Telecom / wireless | -15 to 0 | 0-15 | 25-40 | 45+ |
| Utilities / energy | -10 to 5 | 5-20 | 25-40 | 45+ |
| Airlines / hospitality | 0-15 | 15-40 | 45-60 | 70+ |
| Professional services | 25-35 | 40-55 | 60-75 | 80+ |
| Education / training | 15-25 | 30-50 | 55-70 | 75+ |
| eNPS (employee NPS) | -10 to 0 | 0-30 | 30-50 | 60+ |
Source orientation: medians and quartiles approximated from Bain & Company, Satmetrix (NICE), Qualtrics, and Forrester benchmarking studies. Methodology varies; the published numbers in any single study should be read as estimates with substantial confidence intervals. The ranges above are useful for sanity checking, not as targets.
For twenty years CX teams asked the same question after every NPS wave: how does our score compare to the industry. The answer felt important. It oriented the team. It made the conversation about external position rather than internal action.
The answer was useful and incomplete. A 35 in B2B SaaS is average. A 35 in airlines is exceptional. A 35 trending down on the same customer base across three quarters is a churn signal regardless of where the industry median sits. The industry comparison tells the team one thing: whether the score is roughly in the expected range for the sector. That is the entire informational value.
The comparison that tells the team what to do is different. It is the same customer's trajectory across waves — the customer who scored 9 last quarter and 4 this quarter is the converted detractor a save call could still reach. The industry benchmark cannot identify that customer; the per-contact trajectory does. One comparison orients the team. The other moves the program.
The industry median is useful for the first conversation after the wave lands — "are we roughly where we should be." After that, the conversation should move to the trajectory: which customers moved, which way, and what their verbatim says about it. The benchmark conversation that lasts longer than five minutes is a conversation the team is using to avoid the work.
This is the same locked argument that anchors /use-case/nps-analysis — expressed here through the benchmark frame. The pillar covers analysis broadly; this page resolves the benchmark question.
The industry benchmark answers one specific question well and dozens of others poorly. Knowing which is which is the difference between a useful benchmark conversation and a defensive one.
All useful. None of it tells the team what to do next.
The benchmark gives you context. The verbatim and the trajectory give you action.
A single customer named C-04812. Five quarterly NPS responses. The same persistent contact ID across every wave. This trajectory is more useful than any industry benchmark — it identifies the customers who moved, the direction they moved, and the verbatim that explains it.
The industry benchmark cannot see this trajectory. It sees an average for all customers in a wave. This customer's Q3 verbatim ("the integration we depend on broke") was the early warning. Two waves before churn. The per-contact comparison was the only one that named the risk while there was still time to act.
The shape of this trajectory does not appear in any industry benchmark report. Industry medians average across customers, sample broadly, and report a single number per sector per year. They cannot identify a converted detractor; they cannot point at the customer who dropped from 9 to 4; they cannot quote a Q3 verbatim. That work happens only at the per-contact level, on a persistent ID, with the verbatim attached.
A team running a quarterly NPS program with a persistent contact ID has access to this trajectory for every customer in every wave. Most teams do not use it — the wave is reported as a single number, compared to the industry median, and the trajectory data sits in the response sheet unread. That is the gap this page is trying to close.
Each is a pattern that shows up in NPS programs at least once a quarter. Each treats the benchmark as something it is not.
"Our goal is to reach industry median by Q4." The benchmark is an average across companies of varying maturity, customer mix, and survey methodology. Targeting the median is targeting average — and the team has no idea which specific customers would need to move to get there. Track trajectory instead.
"Apple's NPS is 70; ours is 35; we have work to do." Apple is a consumer-electronics monoculture with decades of brand investment; an enterprise SaaS at 35 is in the same range as its peers. The cross-industry comparison is a vanity exercise. Compare to your own sector and your own prior wave.
A benchmark study from 2019 reported a 38 median for B2B SaaS. The team set 38 as their bar in 2026. Industry NPS shifts with macro conditions — tightening budgets, shifts in switching costs, AI disruption. A six-year-old benchmark is a historical artifact. Use the most recent study you can find, and trust your own trend more than the published median.
The board sees "we are 8 points below industry median." The conversation moves to closing that 8-point gap. Nobody asks which customers moved, what they said, or whether the gap is even real given methodology differences. A program that reports only the benchmark gap is using the benchmark to avoid the work.
This page answers "where does my score sit." Three adjacent reads answer the questions the benchmark cannot: what to do about the score, how to handle the customers driving it, how to close the loop.
What NPS analysis means in 2026 — the methodology, the AI-era thesis, the broader treatment.
Read the pillar →Industry medians, the trajectory frame, the four ways teams misuse benchmarks.
This pageThe customers driving the score — four sub-types, follow-up workflow, anti-patterns. Where the action lives.
Read the workflow →After reading this page, the natural next stop is the detractor workflow — that is where the per-customer trajectory turns into action.
Same instrument, three different programs. None of them rely on industry medians as targets — all of them track per-customer trajectory across waves with the verbatim attached.
A B2B SaaS CS team reports two things to leadership each quarter: the headline NPS (with the industry-median note for context) and the per-customer trajectory chart for the top 50 accounts. The leadership conversation moves quickly from "where are we vs the industry" to "which of these 50 accounts moved, which way, and what did their verbatim say."
A training program tracks NPS at three moments per participant: end of module 1, end of module 4, six months post. The same participant ID carries the three scores plus their verbatim. The team reads each participant's arc rather than the cohort median, and the next cohort's curriculum is designed on the arcs that improved versus those that dropped.
A scholarship program runs NPS-style feedback at six months and one year. Each awardee's response sits on the same record as their original application essay. The team compares each awardee to their own starting point — the gap between what the application described and what the program delivered. Industry benchmarks for scholarship NPS do not really exist; the per-awardee benchmark is the only one available.
Your scores, your contacts, your verbatims. Sixty minutes. No demo accounts.
There is no universal good NPS. The same number means very different things in different industries: a 30 is excellent in airlines, average in SaaS, and poor in consumer technology. The more honest measure is whether the score is improving on the same customer base across waves — the trajectory, not the absolute number. Any single "good NPS" threshold is misleading unless it is benchmarked to a specific industry, sample size, and methodology.
Approximate industry medians published by Bain, Satmetrix, and the major benchmarking studies: SaaS and B2B software 30-40, retail 40-50, financial services 20-40 with wide variation, telecom and utilities 0-15, airlines and hospitality 15-40, healthcare 20-40, consumer electronics 40-60. These are medians. The top quartile in any industry is roughly 20-30 points higher; the bottom quartile is roughly 20-30 points lower. Use the medians as a sanity check, not a target.
B2B SaaS NPS medians typically run 30-40, with top-quartile programs at 50-65 and leading companies above 70. Consumer SaaS skews higher, often 40-60 median. Both are wide ranges. A SaaS NPS of 35 is average; a SaaS NPS of 35 trending downward on the same customer base is a churn signal regardless of where the industry sits.
Healthcare NPS medians typically run 20-40, with substantial variation by sub-sector (provider organizations, payers, devices, pharma). Top-quartile healthcare programs reach 50-65. Healthcare also has unique measurement challenges — patients are not pure customers, and the rating instrument was not designed for clinical contexts. The verbatim matters more in healthcare than in most other sectors.
eNPS (employee NPS) benchmarks differ from customer NPS. Typical eNPS medians run 0-30; top-quartile employers reach 30-50; the upper extreme reaches 60+. Negative eNPS is more common than negative customer NPS — employees can be dissatisfied without intent to leave. The verbatim that arrives with each eNPS response is the part HR teams reliably underuse; the resignation that arrives six months later was usually named in the eNPS comment two quarters earlier.
Three reasons. Industry differences — airlines and telecom have structurally lower scores than SaaS for reasons unrelated to program quality. Methodology differences — some benchmarks use 0-10 scales, some convert from 1-5, some use different definitions of detractor cutoffs. Sample differences — the customer base sampled changes whether the median represents satisfied current customers or broader market sentiment. Treat any single benchmark with caution; trust the trajectory of your own program more than the absolute industry median.
Your own prior wave, on the same customer base. Industry medians tell you where you stand; your prior wave tells you whether you are moving and which customers moved. The first is interesting; the second is actionable. The best NPS programs report two numbers — the current wave's score and the per-customer trajectory across the last four waves — rather than the score and a benchmark comparison.
Major benchmarking studies (Bain, Satmetrix, Qualtrics, Forrester) typically survey a sample of customers across companies in each industry, calculate NPS for each company, and publish industry medians and quartile ranges. Methodology varies. The benchmark numbers in any published study are estimates with material confidence intervals — useful for orientation, misleading if treated as targets.
The theoretical maximum is +100 (every respondent a promoter). Real programs rarely reach 80; the leading customer-experience programs cited in the literature — Tesla, Costco, Trader Joe's, certain Apple product lines — have reported scores in the 70-90 range, though methodology varies. Any single "highest NPS" claim should be read with sample size, recipient set, and methodology in mind.
Use the comparison as a sanity check, not as a target. The right comparison is your prior wave on the same customers. The industry benchmark tells you whether your program is roughly in the expected range for your sector; it does not tell you what to do next, which customers to call, or what verbatim to read. The trajectory of your own program does.
An NPS benchmark report is a published study (usually annual) that reports NPS medians by industry, by company size, by region, and sometimes by segment. Common sources: Bain & Company, Satmetrix (now NICE), Qualtrics, Forrester, ClearlyRated for professional services. Treat any single report as one data point — methodology varies enough that two reports on the same industry can produce very different numbers.
Benchmarks tell you where the score lands. The cluster covers what to do about it. NPS analysis is the methodology pillar. NPS detractor is the workflow that handles the customers driving the score. NPS feedback covers the closed-loop workflow that turns the score into action regardless of where the benchmark sits.
The benchmark is orientation. The cluster covers what comes after the orientation conversation.
The pillar above this page — the methodology, the AI-era thesis, the broader treatment of what NPS analysis means in 2026.
The customers driving the score — four sub-types, follow-up workflow, anti-patterns. Where the per-customer trajectory becomes action.
When the score lands below zero — what it means, what it does not mean, and the four moves before the next leadership conversation.
Your last four NPS waves, your contacts, your verbatims. Sixty minutes. We pull every customer's per-wave score and verbatim onto one record, walk through who moved which way, and identify the converted detractors and converted promoters the benchmark conversation never named. No demo accounts. No slideware. Your own data, read live.
No slideware. No demo accounts. Your own data, read live.
