Net Promoter Score (NPS): Methodology, Benchmarks

NPS is the most cited and most abused SaaS metric. Correctly implemented, it's a directional signal for customer sentiment and expansion propensity.

Net Promoter Score: What It Measures, What It Doesn't, and How to Use It Honestly

NPS asks one question: on a 0-10 scale, how likely are you to recommend us to a friend or colleague? Promoters (9-10) minus Detractors (0-6) gives a score from -100 to +100. The metric is beloved by executives because it produces a single number that fits on a slide, and mocked by researchers because its methodology is statistically weak and its predictive power varies wildly by industry. Both views are correct. NPS is useful as a trend line and diagnostic conversation-starter; it's dangerous when treated as a definitive measure of customer health or as an OKR that teams optimize toward.

Correct methodology

Survey at defined moments (post-onboarding, quarterly, post-support interaction — different moments produce different scores, so don't blend them). Include the follow-up 'why did you give that score?' — the free-text response is often more valuable than the number. Survey a representative sample; don't cherry-pick happy users. Aim for 20-40% response rate; below that, response bias distorts the score heavily. Compute score = %Promoters - %Detractors, ignoring Passives (7-8). Track trend over time on a consistent survey population.

Benchmarks that hold up

SaaS median NPS is around 30-40. Top-quartile B2B SaaS: 50-70. Top-decile: 70+. Consumer NPS runs wider — some category leaders hit 70-80, most consumer products land at 20-40. Industry matters enormously — insurance and telecom NPS averages hover around -10 to +10, so a fintech disruptor at +40 looks amazing relative to peers. Compare against your own trend and against direct comparables, not absolute cross-industry benchmarks.

Where NPS misleads

(1) Small samples make the number swing wildly month-to-month — never make decisions off a single quarter. (2) Response bias favors happy and very unhappy customers; the vast middle rarely responds. (3) NPS doesn't predict churn well by itself — churn is driven by many factors, of which sentiment is only one. (4) Optimizing for NPS often incentivizes gaming (surveying only happy users, timing surveys after positive experiences). (5) A high aggregate NPS can hide a small but growing segment of detractors that will churn — segment the score before drawing conclusions.

What makes NPS useful

The score itself is directional; the verbatim comments are gold. A thoughtful NPS program surfaces: recurring product gaps mentioned by Detractors, specific features Promoters cite as loved (candidates for marketing emphasis), moments of friction (onboarding, billing, support handoffs) that generate poor scores. Route Detractor comments to CS immediately for outreach. Route Promoter comments to marketing for case-study candidates. NPS as a listening system beats NPS as a scoreboard.

Better complements

Pair NPS with: (1) product engagement metrics (weekly active users, feature adoption depth) — sentiment plus behavior beats sentiment alone; (2) actual retention cohorts — a company with rising NPS and declining retention has a signal problem; (3) CSAT for transactional moments (post-support, post-onboarding); (4) customer effort score (CES) for friction diagnosis. NPS in isolation is theater; NPS embedded in a broader customer-health system drives real decisions.

Frequently asked questions

How often should we survey?
Relationship NPS: quarterly for most B2B, biannually for enterprise (survey fatigue is real). Transactional NPS: after specific events (onboarding complete, support ticket closed). Never survey the same account more than 4 times a year total across all surveys.
Should we tie compensation to NPS?
Rarely. Comp-linked NPS almost always creates gaming — sales teams coach customers on how to score, CS teams delay surveys until issues are resolved, product teams focus on Promoter delight over Detractor pain. Use NPS as a leading indicator in reviews, not a comp lever.
What's a 'good' NPS to report to investors?
Investors care more about trajectory and segmentation than absolute number. NPS of 45 with a clean upward trend and strong scores in your target segment beats NPS of 60 that's flat and hides segment weakness.

Related fundraising guides (40)

Investor directory · Fundraising library · Articles A–Z · Company funding database