NPS: the metric that measures what your customers won’t tell you

CES, CSAT, NPS: three metrics, three dimensions of customer experience

Sources: Corporate Executive Board / Gartner 2010-2024 · Harvard Business Review · Armatis

Why NPS has become a global standard

Its strength lies in its simplicity. A single question: “On a scale of 0 to 10, how likely are you to recommend our company to a friend or colleague?” This standardised wording, defined by Bain & Company and Satmetrix, makes it possible to compare results over time, across teams, and against competitors in the same sector.

But its popularity isn’t only down to how easy it is to implement. Fred Reichheld showed, in his original research, that a 5% increase in customer retention can generate between 25% and 95% growth in customer lifetime value. NPS, as a predictive indicator of loyalty, is directly correlated with this dynamic. According to Satmetrix, a high NPS is statistically linked to greater market share and faster growth than competitors.

This is why NPS is now built into BPO contracts, executive dashboards, and the quarterly targets of many CX leadership teams. It’s no longer a standalone measurement tool: it’s a steering signal for the customer relationship over time.

How to calculate NPS: the formula and the three customer categories

Data is collected by asking customers to rate their likelihood of recommending you on a scale of 0 to 10. Their responses fall into three categories:

• Promoters (scores 9 and 10): loyal, enthusiastic customers. They actively recommend your company, buy more, stay longer, and are your best ambassadors.
• Passives (scores 7 and 8): satisfied but not engaged customers. They don’t cause harm, but they don’t drive growth either. They’re vulnerable to competitor offers.
• Detractors (scores 0 to 6): dissatisfied customers. Studies show that an unhappy customer tells around ten people in their circle on average. They represent an active risk to brand reputation.

Formula:

NPS = % Promoters − % Detractors

Example: out of 200 respondents, 90 give a score of 9 or 10 (45%), 30 give a score of 0 to 6 (15%). NPS = 45 − 15 = +30.

Passives aren’t included in the calculation, but they shouldn’t be ignored either. They’re the easiest to convert into promoters with the right targeted actions.

How to interpret your NPS score

NPS ranges from -100 (all customers are detractors) to +100 (all customers are promoters). In practice, here are the broad reading zones generally used by Bain & Company and the Satmetrix benchmarks:

A score above 0 is positive, and a score above 50 is generally considered excellent. But these absolute thresholds aren’t enough on their own to assess performance. An NPS of +20 is very satisfactory in the telecommunications sector, and disappointing in luxury. This is why sector comparison is essential.

2026 NPS benchmarks by sector

The data below comes from publications by Bain & Company, Satmetrix and Init Marketing (France data, 2026). It reflects observed median scores and serves as market reference points, not absolute standards.

Two important points on reading these benchmarks. First, US scores are structurally higher than French scores: French consumers tend to rate more harshly, regardless of sector. Second, the trajectory over time is often more informative than the absolute level. An NPS moving from +15 to +25 over six months following a corrective action is a strong signal, even if it remains below the sector average.

What NPS doesn’t measure (and how to close the gap)

NPS measures the overall strength of the bond between a customer and a brand. It doesn’t measure satisfaction after a specific interaction (that’s CSAT’s role), nor the effort a customer had to make to resolve their issue (that’s CES’s role). These three metrics are complementary and reinforce one another.

Excessive effort, as measured by CES, eventually degrades transactional satisfaction (CSAT), which in turn erodes the overall relationship (NPS). Conversely, a smooth, effective experience feeds immediate satisfaction and the likelihood of recommendation. This is the virtuous circle that cross-monitoring all three metrics helps build.

Post-complaint NPS deserves particular attention. According to field data compiled by Armatis, it’s often 20 to 30 points lower than overall NPS. Friction points in the customer journey (complaints, refunds, service failures) are precisely where the relationship is won or lost. Companies that handle these moments methodically turn potential detractors into lasting promoters.

How to improve your NPS in practice

Improving your NPS isn’t about optimising the score itself. It’s about identifying and addressing the root causes of dissatisfaction. A few principles that have proven effective:

Read the verbatims, not just the scores. The open-ended question that follows the NPS rating (“Why did you give this score?”) is often more valuable than the score itself. Customers’ words reveal the real irritants, where the numbers only show their intensity.

Target passives as a priority. Detractors are visible and well known. Passives (scores 7 and 8) are often overlooked. Yet they’re the most accessible: a targeted action, a bit of extra attention, a stronger value proposition can be enough to move them into the promoter category.

Measure by journey, not just globally. A global NPS gives you a trend. An NPS at key journey moments (first order, complaint, renewal) gives you precise levers for action. The gaps between journeys are often far more revealing than the overall score.

Act fast on detractors. Handling negative feedback within 24 to 48 hours significantly reduces churn risk and can reverse a customer’s perception. Speed of handling is often more important than its nature.

Measuring customer satisfaction and turning it into an operational lever are two different things. If you’d like to set up an NPS, CSAT and CES management framework as part of an outsourced customer relationship, see our complete guide to NPS, CSAT and CES and how to manage the performance of your outsourced contact centre.

FAQ: everything you need to know about NPS

What’s the difference between NPS and CSAT?

NPS measures how likely a customer is to recommend your company: it’s a metric of the overall relationship and long-term loyalty. CSAT measures satisfaction after a specific interaction: it’s a transactional, immediate metric. The two are complementary. A customer can be satisfied with an interaction (high CSAT) but not recommend the brand (low NPS), particularly if other aspects of the relationship are problematic.

How often should NPS be measured?

A quarterly or half-yearly measurement is recommended for overall NPS, which assesses the relationship over time. For post-interaction NPS (after a complaint or a renewal), measurement should be triggered by the event, not by a fixed schedule. Consistency matters more than frequency: steady tracking over time is worth more than a single highly detailed measurement.

Does a negative NPS mean you’re in a bad situation?

A negative NPS indicates that detractors outnumber promoters, which is a warning sign. But it should be read in its sector context: in telecommunications or insurance, slightly negative NPS scores can be in line with the market average. What matters most is the trajectory: a negative NPS that improves steadily after corrective action is far more encouraging than a stable NPS of +10.

Should you communicate your NPS externally?

That’s a strategic decision. If your NPS is above +60, communicating it on your sales materials and website strengthens credibility. Below +30, communicating it can backfire. In all cases, NPS is primarily an internal management tool: its value lies in the decisions it enables, not in how it’s displayed.

Is NPS relevant in B2B?

Yes, and it’s actually widely used in B2B, particularly in SaaS, professional services and BPO. In B2B, the question can be adapted to target operational contacts rather than decision-makers, and NPS is often supplemented with qualitative interviews to dig deeper into the reasons behind the scores given.