Definition
Net Promoter Score (NPS) is a customer-loyalty metric calculated from a single question — 'How likely are you to recommend [company] to a friend or colleague?' — scored 0–10, with the result expressed as % promoters minus % detractors.
NPS gets criticised as a single-question vanity metric, and the criticism has merit. But it's also stupidly comparable across industries, time periods and companies — and that comparability is why it stuck. Used as a directional indicator (is loyalty trending up or down?), NPS is useful. Used as a target ('hit NPS 50 by Q4'), it gets gamed.
The follow-up open-ended question — 'why did you give that score?' — is where the real value lives. The scores are noisy; the verbatim feedback is gold. Most teams that get value from NPS are running it monthly, mining the verbatims, and ignoring the headline number.
Origin
Developed by Fred Reichheld at Bain & Company and published in Harvard Business Review in 2003 ('The One Number You Need to Grow'). Adopted across SaaS, retail and services through the 2010s.
How it works
- Send the question 'How likely are you to recommend us, 0-10?' to customers.
- Send a follow-up: 'What's the main reason for your score?'
- Bucket: Detractors (0-6), Passives (7-8), Promoters (9-10).
- NPS = % Promoters - % Detractors. Range: -100 to +100.
- Track over time and segment by cohort, plan tier, and acquisition channel.
- Mine the verbatims for themes — that's the actionable output.
When to use it
Use when
- Quarterly across the customer base, with verbatim follow-ups.
- Post-purchase or post-onboarding to catch early signals.
- After a major product change, to detect impact on loyalty.
Skip when
- As a vanity metric. NPS-as-target invites gaming.
- Without segmentation. A 35 average can mask 50 from enterprise and -10 from SMB.
Key metrics
- NPS score (% Promoters - % Detractors).
- Response rate (target: 25%+ for the survey to be representative).
- Verbatim themes (counted manually or via NLP).
- NPS by cohort and segment.
Examples
- NPS climbed from 22 to 41 after we shipped the new onboarding flow.
- The verbatim themes told us pricing was the friction, not the product.
- Don't optimise for NPS. Optimise for the things that drive NPS.
In practice at Makreate
Makreate doesn't run product but our brand and UX work is judged in part on NPS shifts. A recent SaaS client had NPS at -8 (more detractors than promoters). Verbatims pointed to onboarding confusion, not product quality. We redesigned the first-week experience: a clearer signup flow, a checklist of three quick wins, an onboarding email sequence. NPS climbed to 31 in five months, churn dropped, and the verbatim shift from 'I don't know how to use this' to 'this saves me hours' was the actual leading indicator.
UX Design →Common mistakes
- Treating NPS as a vanity target. It can be gamed; don't.
- Ignoring the verbatims. The score is noise; the comments are signal.
- Sending NPS too often. Customers tire of the question.
- Comparing your NPS to other industries. Benchmarks are category-specific.
Frequently asked
What's a good NPS score?
SaaS averages around 30. Enterprise software 40+. Telcos and airlines often live below 20. Compare yourself to your category, not the absolute number.
How often should I run NPS?
Quarterly is plenty. Monthly creates survey fatigue. Pulse surveys after key moments (signup, first invoice, 90-day mark) are more useful than constant NPS.
Is NPS a good predictor of growth?
Loosely. Reichheld's original claim was overblown. NPS correlates with retention and word-of-mouth but doesn't predict revenue growth on its own.