In 2003, a consultant named Fred Reichheld published an article in Harvard Business Review with a provocative title: "The One Number You Need to Grow." The number was the NPS, the Net Promoter Score.
More than twenty years later, that metric is on the dashboards of the world's largest companies. Apple uses it. Netflix uses it. Amazon uses it. And thousands of small businesses do too, although most measure it in a way that's useless.
This article explains what NPS really is, what problem it tries to solve, why most people implement it wrong, and how to do it well without complicating your life.
The question that changed everything
NPS is based on a single question:
"On a scale of 0 to 10, how likely are you to recommend [company/product] to a friend or colleague?"
The answers are grouped into three categories:
- Promoters (9-10): enthusiastic customers who speak well of you without being asked
- Passives (7-8): satisfied but not loyal. They'll leave if they find something slightly better
- Detractors (0-6): dissatisfied customers. They can damage your reputation
The calculation is simple: the percentage of Promoters minus the percentage of Detractors. Passives are ignored.
The result is a number that ranges from -100 to +100. If everyone gives you a 9 or 10, your NPS is 100. If everyone gives you 0 to 6, it's -100.
Why this question works
The brilliant thing about the question isn't the number. It's what it asks.
Most satisfaction surveys ask "how was the experience?" People answer "fine" out of politeness, even if they never buy again. The answer predicts nothing.
The NPS question is different because it puts the customer in a social situation. They're not evaluating your product in the abstract, they're deciding whether to put their own reputation on the line to recommend you. That triggers a much higher level of honesty.
That's why it works as a predictor of behavior. Multiple studies have shown that companies with high NPS grow faster than their competitors in the same industry. An analysis of B2B SaaS estimates that a 25% increase in NPS can translate into a 10% to 15% increase in MRR (Dromo).
What a "good" NPS is
This is the most frequent question and the one that generates the most confusion.
The short answer: it depends on your industry.
An NPS of 30 can be excellent for an airline and mediocre for a clothing brand. Current benchmarks, according to 2026 data, look roughly like this (Sybill, Lorikeet):
- SaaS: 30-40 average
- Ecommerce: 45-55 average
- Healthcare: 38-58 average
- Financial services: 35-45
- Airlines: around 26 average
- Telecommunications: usually negative
The most universal general rules: above 0 means more promoters than detractors (the minimum acceptable), above 30 is good, above 50 is excellent, and above 70 puts you in the world's top tier.
The most common mistake is obsessing over the absolute number. What matters is the trend over time and your position relative to your competition. A business that went from 20 to 35 in six months is doing something very right. One that's been at 60 for three years could be in free fall without realizing it.
The three mistakes almost everyone makes
Mistake #1: Measuring without context
Asking for an NPS every 12 months "for the annual report" is useless. NPS is a living indicator; it should be measured frequently enough to detect changes and act on them.
For an ecommerce store, it makes sense to measure it after every purchase. For a SaaS, after specific milestones like completed onboarding or the first month. For a restaurant or clinic, after every visit. The key is to capture it in the hot moment, when the experience is fresh.
Mistake #2: Asking for the number and not the "why"
An NPS without a comment is almost useless.
If a customer gave you a 4, you need to know why. Without that information, you have a number that doesn't tell you what to improve. The best practice is to always follow the numeric question with an open one: "What would you tell us we could improve?"
The open answers are where all the valuable information lives. The number tells you how serious the problem is. The text tells you what to do.
Mistake #3: Measuring and not acting
The biggest waste we see. Companies that spend time and money collecting NPS, put it on a dashboard, and nothing. Nobody looks at it. Nobody responds to detractors. Nobody converts the passives.
Data from BrightLocal and other studies show that closing the loop with the customer (responding, telling them what you'll do, and then following through) is what turns NPS into growth. Without that step, you have a vanity metric.
The low response rate problem
Here's a fact few companies know: the average NPS measured by email has a response rate of between 6% and 25%, and anything below 5% suggests something is wrong, whether bad timing, low relevance, or respondent fatigue (Frill).
If only 5% of your customers respond, your NPS is highly biased. Those who respond tend to be the ones who are extremely happy or extremely angry. Passive people, who are the majority, rarely bother to answer.
Well-designed campaigns reach 30-40% response. How do they do it? Right timing (immediately after the experience), the right channel (a QR code on-site works better than email), low friction (a single question, 10 seconds max), and personalization.
Voice NPS: the method almost nobody uses
Here's something that radically changes the dynamic and that few companies are trying: combining NPS with voice feedback.
The classic problem with NPS is that the number without context is shallow. The "why" by text is usually short, dry, and uninformative. People write "ok," "normal," "fine," nothing actionable.
When the customer speaks instead of writing, something different happens. The answers are longer, more emotional, more specific. A customer can type "service was slow" in five seconds. But if you ask them to tell the story of their experience, they can talk for 45 seconds explaining exactly what went wrong, how it made them feel, and what they would have preferred.
That difference transforms NPS from a numeric metric into a source of qualitative insights. And if AI automatically classifies the topics mentioned, you have a real-time map of what's working and what isn't.
How to implement it in practice
If you want to start with NPS without complicating things, this is the simplest recipe that works:
Pick the moment. Identify the point in your customer's experience where the impression is freshest. For a restaurant, when paying the bill. For a clinic, when leaving the office. For a SaaS, when completing onboarding.
Design a survey of 2 questions max. The NPS question and "What would you tell us we could improve?" More than that is punishment and people give up.
Measure it frequently. Monthly at a minimum. In high-traffic businesses, weekly. The idea is to detect changes early.
Respond to everything you can. Thank the promoters. Respond to detractors personally, understand what happened, offer a solution. Ask the passives what would get them to give you a 9 or 10.
Segment. An aggregate NPS without segmentation is a dead number. Look at NPS by location, by product, by customer segment. That's where the real problems appear.
What NPS is not
Before closing, it's worth clarifying what NPS is not:
It's not a measure of momentary satisfaction (CSAT exists for that). It's not a replacement for qualitative feedback, it's a starting point, not a finish line. It's not useful if you measure it once a year and forget about it. And it's not the only number that matters, you also have to look at churn, CLV, CAC.
But used well, as part of a continuous system of listening to the customer, it's probably the most effective indicator there is for knowing whether your business is building real loyalty or just transactions.
The starting point
If you've never implemented NPS, don't start with a big project. Start with a small campaign: a single question, one channel (QR, email, in-app, pick one), one touchpoint, and one person responsible for reading the answers and responding.
In 30 days you'll have your first NPS. In 90 days you'll have enough data to start seeing trends. In 6 months, if you acted on the feedback, you should see the number rise.
And raising NPS is, in almost every sector, a preview of rising revenue.
With VozFeed you can implement NPS combined with voice feedback in 10 minutes. The customer scans a QR code, answers the question and leaves a comment by speaking or typing. AI classifies the topics automatically. Try it free for 14 days.