What is NPS in B2B — and is it still useful in 2027?

Direct Answer

NPS — Net Promoter Score — is the percentage of your customers who rate you 9 or 10 on "how likely are you to recommend us" minus the percentage who rate you 0 through 6. Created by Fred Reichheld at Bain in 2003. In 2027, the honest take is that NPS is half-dead.

The number itself is mostly vanity, especially when compared across vendors. But the delta in NPS over time within your same cohort, plus the verbatim comments behind the scores, plus the operational use of identifying promoters for referrals — that part still earns its keep.

TL;DR

• NPS = % Promoters (9-10) minus % Detractors (0-6); Passives (7-8) don't count toward the math but matter directionally.
• 2027 B2B SaaS median NPS is 33 per CustomerGauge; world-class is 50+; below 20 signals a real problem.
• Public benchmarks worth knowing: Snowflake 65, Atlassian 49, Slack 47, HubSpot 41, Salesforce 36.
• The score is 5% of the value; the verbatim comments and the delta over time inside your own cohort are the other 95%.
• Stop treating NPS as a health score, stop surveying only annually, and stop ignoring the open-ended feedback — those are the three failure modes that turn it into vanity.

Real 2027 NPS Benchmarks

The numbers below come from CustomerGauge's 2024 B2B NPS benchmark report cross-referenced with publicly disclosed investor-day metrics and analyst tear-downs. Treat them as orientation, not a leaderboard — sampling and methodology differences make exact cross-comparison meaningless, which is itself part of the lesson.

Two things to notice. First, the spread within a segment is wider than the spread between segments — your competitor in the same category can be 20 points away from you and it tells you almost nothing about real loyalty. Second, the top-quartile threshold sits roughly 15-20 points above the median in every row, which is a more useful internal goal than chasing a public number.

What NPS Tells You (and What It Doesn't)

NPS does three things well. First, it tracks trend over time within your own customer base. If your relationship-NPS drops 8 points quarter over quarter among the same cohort, something real happened — a pricing change, an outage, a support degradation.

That signal is real because the methodology is held constant. Second, the verbatim comments are gold. A $35M ARR SaaS we know runs quarterly relationship-NPS plus transactional-NPS after key events like onboarding completion and renewal; the open-ended comments drove six product roadmap changes in one year while the absolute number drifted by three points.

Third, NPS gives you an operational list of Promoters. The 9s and 10s are your referral engine, your case-study pool, your G2 and Gartner review army — that list alone justifies the program for many companies.

NPS does not tell you three other things reliably. It does not measure absolute customer health — a single question cannot capture product adoption, support ticket volume, executive sponsor stability, or contract value at risk. It is a fragment, not a dashboard.

It does not predict churn well — some Detractors complain bitterly for five years and never leave because switching costs are too high, while some 9-rated Promoters churn when their champion gets a new job. And it does not support cross-vendor comparison. Two vendors with "47 NPS" may have sampled wildly different cohorts at wildly different moments using slightly different question framing in different languages.

Marketing teams love to put NPS in pitch decks; sophisticated buyers learned to discount it years ago.

The 3 NPS Failure Modes That Turn It Into Vanity

The first failure mode is treating NPS as your health score. It is not. A health score is a composite — product usage, support ticket sentiment, executive engagement, billing health, NPS, and renewal proximity rolled together.

When a CS team says "our health score is NPS," they mean they don't have a health score, they have a fragment that happens to produce a number. Build the composite in Gainsight, Catalyst, or even a thoughtful spreadsheet.

The second failure mode is surveying once a year. Annual NPS is a lagging indicator of a lagging indicator. By the time the score drops, the customer has already had three bad quarters.

The fix is two-cadence: a quarterly relationship-NPS that goes to a representative sample, plus transactional-NPS triggered by specific events — post-onboarding at day 30, post-support-ticket close, post-renewal, post-major-feature-release. The transactional version surfaces problems within days instead of months.

The third failure mode is ignoring the comments. Teams obsess over the number, build dashboards around the number, set OKRs against the number — and never read the verbatims. The score is 5% of the value.

The comments are 95%. A good NPS program reads every single open-ended response, tags them by theme (pricing, product gap, support quality, account team), and routes themes to the function that can fix them. Tools like Qualtrics XM at $30-100K per year do this with AI-driven theme detection; Delighted at $350-2,000 per month handles it for mid-market; Wootric and SurveyMonkey serve SMB; Gainsight CS Index is the B2B-specific option with workflow built around survey response.

Frequently Asked Questions

How often should we survey? Quarterly for relationship-NPS to a representative sample, transactional immediately after key events. Annual is too lagging; monthly creates survey fatigue and pollutes the data.

NPS vs CSAT vs CES — which one? Use all three for different jobs. CSAT measures specific interactions (support ticket, training session). CES measures effort at a friction point (renewal, onboarding). NPS measures overall loyalty and surfaces promoters. They answer different questions, not the same question three ways.

What's a good response rate? B2B benchmarks land around 15-25% for relationship-NPS and 30-40% for transactional. Below 10% means your sample is biased and the score is unreliable; chase the rate before you chase the score.

Sources

• Reichheld, Fred. "The One Number You Need to Grow." Harvard Business Review, December 2003 — the original NPS paper.
• Bain & Company. "Net Promoter System — The Loyalty Acid Test." Bain NPS practice publications, 2023-2024.
• CustomerGauge. "2024 B2B NPS and CX Benchmarks Report" — segment-level median and top-quartile data.
• Qualtrics XM Institute. "2024 Global CX Trends and NPS Benchmarks."
• Gainsight. "2024 Customer Success Index — NPS in B2B SaaS."
• Forrester Research. "The State of CX Measurement, 2024" — critique of NPS-as-health-score practice.
• HubSpot, Salesforce, Snowflake, Atlassian — publicly disclosed NPS figures from FY2023-2024 investor materials and annual reports.
• Delighted, Wootric, SurveyMonkey product documentation — tooling pricing and methodology references.