What Is a Good NPS Score? 2026 Benchmarks by Industry

What is a good NPS score?#

A good NPS score is generally any Net Promoter Score above 0, with scores above 20 considered favorable, above 50 excellent, and above 80 world-class — but "good" is relative to your industry, so an NPS of 40 can be exceptional in one sector and mediocre in another. Because the metric runs on a scale from -100 to +100, any positive number means you have more promoters than detractors; the real question is whether your NPS score beats your industry's average and improves over time.

That range comes straight from the framework's origin. NPS was introduced by Fred Reichheld of Bain & Company in the 2003 Harvard Business Review article "The One Number You Need to Grow", and Bain still publishes the tier thresholds most teams anchor to. If you want the underlying mechanics before comparing yourself to anyone, start with the full Net Promoter Score definition and formula and the 0-10 NPS scale explained.

NPS benchmarks by industry (2026)#

NPS benchmarks vary by more than 40 points across industries, which is why an industry-specific comparison is far more useful than a single global average. The all-industry average NPS sits around 32, while the median lands closer to 44 — but those aggregate figures hide enormous sector variation. The table below aggregates 2026 industry benchmark datasets into directional averages; treat them as a starting line, not a scoreboard, because methodology (relational vs. transactional, sample size, region) shifts every number.

Two 2026 movements are worth naming. Insurance saw one of the sharpest single-year drops of any category — driven by pricing increases and claims disputes — a reminder that a benchmark is a snapshot of a moving target, not a fixed bar. Healthcare also slid as access and cost pressures hit sentiment. If your score fell this year, the benchmark alone won't tell you whether you underperformed or whether your whole category moved; for that you need the reasons, which we get to below. For the wider metric landscape these numbers live inside, see the 8 customer experience metrics that matter in 2026.

What counts as world-class vs. poor NPS?#

World-class NPS is a score above 80, while anything below 0 is poor because it means detractors outnumber promoters. Between those poles, Bain & Company's widely cited tiers give a universal read that works across industries, published on the Bain Net Promoter System benchmarks page:

• Below 0 — Poor. More customers would actively discourage others than recommend you. This is a churn-and-reputation emergency.
• 0 to 20 — Needs improvement, but you have a net-positive base to build on.
• 20 to 50 — Good to great. You are outperforming most companies and generating word-of-mouth.
• 50 to 80 — Excellent. Strong loyalty and, typically, healthy organic growth.
• 80 to 100 — World-class. Reserved for the rare brands customers evangelize.

The catch: these universal tiers and the industry averages above can disagree. A B2B SaaS company with an NPS of 45 clears the "good to great" universal tier but only matches its sector average — so it is typical, not exceptional. Always read your number against both lenses. The good/average/world-class distinction matters most when you are setting a target, and the most defensible target is your own trailing score plus your sector's trajectory.

B2B vs. B2C: why the "good" number differs#

B2B NPS averages roughly 11 points lower than B2C NPS, so a "good" B2B score is a lower absolute number than a good B2C score. B2C companies average around 49 while B2B companies average around 38, largely because B2B purchases involve longer sales cycles, multiple stakeholders, and switching costs that dampen enthusiasm even among satisfied accounts. A B2B software firm posting an NPS of 40 is doing well; a consumer app with the same 40 is middling.

This gap is why copying a headline benchmark from a different business model is a trap. A customer satisfaction score (CSAT) or NPS number only means something against the right peer set. It also compounds with a measurement reality: response rates on relational NPS surveys are low, so a B2B account team hearing from a handful of champions can read a rosy score that hides quiet, at-risk accounts. Understanding customer retention and the signal surveys miss is often a better early-warning system than the NPS number itself.

Why NPS benchmarks mislead without context#

NPS benchmarks mislead when they are treated as a diagnosis rather than a starting line, because the score compresses every reason a customer might rate you into a single number that says what and who but never why. Three specific traps recur:

1. Methodology mismatch. Relational NPS (a periodic "how likely are you to recommend us?") and transactional NPS (fired after a specific interaction) produce different numbers for the same company. Comparing your transactional score to a benchmark built from relational surveys is apples to oranges.
2. Sampling and response bias. With typical NPS response rates in the single digits to low teens, the customers who bother to answer skew toward the delighted and the furious. The ambivalent middle — often where churn hides — goes unheard.
3. The reason gap. Two companies can post an identical NPS of 45 for opposite reasons: one is loved for its product and hated for its pricing, the other is the reverse. The benchmark treats them as equal. Only the open-ended "why" separates them, which is exactly what a static score discards. This is the same blind spot we cover in why traditional NPS surveys are not enough and why product teams are sunsetting NPS in 2026.

A benchmark tells you whether you are behind. It never tells you what to fix. That distinction is the whole game, and it is where most CX programs stall — they chase a number without a mechanism to move it. Customer experience analytics dashboards have the same limitation: they show which line moved, not the human reason it moved.

How to move your NPS score (act on the "why")#

You move an NPS score by understanding the reasons behind detractor and passive ratings and closing the loop on them — not by chasing the benchmark itself. The number is a lagging summary; the levers are the specific, nameable frustrations and delights underneath it. A practical sequence:

• Segment before you average. A single company-wide NPS hides the segments dragging it down. Break the score out by plan tier, tenure, industry, and lifecycle stage. Mapping metrics to the stages of the customer lifecycle shows you where loyalty leaks.
• Read the verbatims, not just the number. The comment box after the 0-10 question is where the actionable signal lives — yet a text field gets short, vague answers ("too expensive," "works fine") that rarely explain the decision.
• Follow up on the "why." The single highest-leverage move is asking a real follow-up question and probing the vague answer. "Too expensive" means nothing until you learn relative to what and for which feature. This is the open-ended follow-up that actually matters on an NPS survey.
• Close the loop and re-measure. Ship the fix, tell the detractors who raised it, and watch the segment recover. Loyalty economics — the core of Reichheld and Markey's The Ultimate Question 2.0 — ties recovered detractors directly to customer lifetime value.

Here is the shift worth making: a web form or a static survey can capture the score, but it cannot capture the reason at any depth. This is where AI-moderated conversations change the economics. Perspective AI runs hundreds of AI-led interviews at once that ask the recommendation question and then follow up in the customer's own words — probing "too expensive" until it becomes "I couldn't justify the Pro tier without SSO." That is the difference between a benchmark you can report and a reason you can act on. It is the same argument behind the broader shift from survey-based CX measurement to conversational voice-of-customer, and it is why teams comparing options increasingly evaluate NPS alternatives that capture the why alongside the score itself.

Understanding the "why" is also the fastest way to read your broader customer experience metrics and your customer sentiment — because every one of them shares the same blind spot the raw NPS number does.

Frequently Asked Questions#

What is considered a good NPS score in 2026?#

A good NPS score in 2026 is any positive number, with 20+ considered favorable, 50+ excellent, and 80+ world-class, per Bain & Company's tiers. In practice, "good" is relative to your industry: an NPS of 40 is strong for B2B SaaS but only average for consumer retail. The most reliable target is beating your sector's average and improving on your own prior score.

What is the average NPS score across industries?#

The average NPS score across all industries is roughly 32, with a median closer to 44. That spread reflects how much sector matters — consulting and hospitality often average 50 to 60, while telecommunications and B2B SaaS cluster in the low-to-mid 30s. Because averages hide 40-plus points of variation, compare yourself only to your own industry, not to the global figure.

Is a higher NPS always better than a lower one?#

Not necessarily, because NPS is only comparable within the same industry, business model, and survey methodology. A B2B company with an NPS of 38 may be outperforming its peers, while a B2C brand with 45 may be lagging its own sector. A higher raw number from a different context tells you nothing actionable — trajectory against your own baseline is the better signal.

Why is my NPS score high but customers still churn?#

A high NPS with ongoing churn usually means your survey is hearing from promoters while at-risk customers stay silent, since NPS response rates run in the single digits to low teens. The score reflects your most engaged customers, not the ambivalent middle where churn hides. Closing that gap requires reaching non-responders and capturing the reasons behind passive and detractor ratings, not just the score.

Can you compare NPS across different industries?#

You should not compare NPS across different industries as a performance judgment, because benchmark averages differ by more than 40 points between sectors. A financial-services NPS is not directly comparable to a telecom or healthcare NPS. Cross-industry figures are useful only for understanding the range; for evaluating your own performance, use your industry benchmark and your historical trend.

Conclusion#

A good NPS score is one that beats your industry benchmark and climbs against your own history — not a universal number you can lift from a chart. The 2026 benchmarks by industry give you the starting line: know whether you are ahead of consulting's 60, insurance's 45, or B2B SaaS's 36, and read your number against both the sector average and Bain's universal tiers. But a benchmark is a diagnosis you cannot act on. You move an NPS score by understanding the reasons behind it — the specific, nameable "why" that a 0-10 number and a short comment box will never surface.

That is the work worth doing, and it is why the fastest-moving CX teams are pairing the score with conversation. If you want to go past the benchmark to the reasons, start a conversational study with Perspective AI and let AI-moderated interviews probe the "why" behind every rating — or explore how CX and product teams are rebuilding their measurement stack around understanding, not just scoring.