NPS (Net Promoter Score) is a customer loyalty metric based on one question: "On a scale of 0–10, how likely are you to recommend [company] to a friend or colleague?" Respondents are classified as Promoters (9–10), Passives (7–8), or Detractors (0–6). NPS = % Promoters − % Detractors. Scores range from −100 to +100. Unlike CSAT (which measures transaction satisfaction), NPS measures overall relationship loyalty.
How NPS is calculated
The formula is: NPS = % Promoters − % Detractors. Passives are excluded from the calculation — they count toward the denominator (total respondents) but contribute zero to the NPS score. Example: 100 respondents — 60 Promoters, 25 Passives, 15 Detractors → NPS = 60% − 15% = 45.
| Score | Classification | Implication |
|---|---|---|
| 9–10 | Promoter | Likely to recommend; high retention probability |
| 7–8 | Passive | Satisfied but not loyal; vulnerable to competitor offers |
| 0–6 | Detractor | Dissatisfied; risk of churn and negative word-of-mouth |
NPS ranges from −100 (every respondent is a Detractor) to +100 (every respondent is a Promoter). Most organizations operate somewhere between −20 and +60 depending on industry.
What NPS scores mean in context
NPS scores vary significantly by industry — a score that signals strong performance in telecom may be average in hospitality or professional services. The baseline varies because different industries have fundamentally different customer relationship types, switching costs, and interaction frequencies. Comparing your NPS to a company in a different vertical produces noise rather than insight.
Trend matters more than the absolute number. A rising NPS over 6–12 months — accompanied by falling detractor rates — is a more meaningful signal of genuine improvement than reaching a specific score target. Organizations that optimize for the number rather than the underlying drivers of loyalty tend to produce inflated scores that don't predict actual retention behavior.
Methodology also matters: differences in question wording, scale anchoring, survey timing, and respondent selection produce different baselines. Do not compare NPS figures across surveys that used different methodologies, even within the same organization. Establish a consistent measurement approach before drawing trend conclusions.
NPS vs CSAT
CSAT measures satisfaction with a specific transaction — a post-call survey asking "how satisfied were you with this interaction?" on a 1–5 scale. NPS measures relationship loyalty — "would you recommend us?" CSAT is a leading indicator for individual interaction quality; NPS is a lagging indicator of accumulated relationship perception. Both are useful; they answer different questions.
High CSAT on individual interactions does not automatically produce promoters. A customer can have a well-handled support call (high CSAT) but still score 6 on NPS because of a poor onboarding experience, a billing dispute from three months ago, or dissatisfaction with the product itself. NPS reflects the entire relationship, not any single touchpoint. This is why contact center teams use both — CSAT to measure what they can directly control, NPS to understand their contribution to overall loyalty. See CSAT measurement.
Relationship NPS vs transactional NPS
Relationship NPS is sent periodically — quarterly or annually — to the full customer base or a representative sample. It measures overall loyalty: how does the customer feel about the company as a whole, right now? This is the "classic" NPS use case, and it is best for tracking macro loyalty trends over time.
Transactional NPS is sent after specific interactions — a support call, a purchase, an onboarding session. It measures the loyalty impact of that particular touchpoint, not the overall relationship. Contact centers often use transactional NPS after service interactions to isolate how support quality affects the promoter/detractor distribution. Both are valid; they answer different questions and should not be directly compared.
NPS in contact center operations
Post-call NPS surveys after support interactions give contact center operations a direct line to how service quality maps to loyalty outcomes. Segmenting transactional NPS by call type, agent, and queue identifies which operational dimensions drive detractors — enabling targeted coaching and process improvement rather than general CX initiatives.
Linking NPS data to operational metrics (AHT, FCR, abandonment rate, wait time) reveals which operational factors correlate with NPS movement. A queue with high FCR and low wait time will typically produce a higher Promoter rate than one with the same FCR but long wait times — the operational data explains the NPS signal. See contact center analytics for how to build that linkage in a reporting environment.
Detractor follow-up (closed-loop programs) contact low-score respondents within 24–48 hours of the survey response to acknowledge the issue and attempt resolution. Studies show that successfully resolving a detractor's issue can convert a portion to passives or promoters over subsequent survey cycles — and that the follow-up itself, independent of outcome, signals responsiveness that has loyalty value.
Limitations of NPS
Seven documented limitations should inform how NPS is used and interpreted:
- Response bias. Customers who feel strongly — very satisfied or very dissatisfied — are more likely to respond. Passives (the middle) are systematically underrepresented, which means the score skews toward extremes and may overstate both promoter and detractor rates.
- Not diagnostic. NPS tells you whether loyalty is high or low; it does not tell you why. A follow-up open-text question ("what is the primary reason for your score?") is required to make NPS results actionable.
- Industry baseline variation. Scores are not comparable across industries or business models. A NPS of 30 in B2B professional services signals different performance than a 30 in consumer retail.
- Slow to change. NPS moves on a months-to-quarters timescale. It is not a metric for measuring the impact of a two-week process improvement. Use CSAT for short-feedback-loop measurement; NPS for long-term trend tracking.
- Single-question limitations. The standard NPS question measures a single dimension (likelihood to recommend) and misses nuance about what is driving that score. Always pair with follow-up qualitative collection.
- Survey timing effects. An NPS survey sent immediately after a billing dispute will produce different results than one sent mid-cycle. Survey timing relative to recent interactions significantly affects scores, making methodological consistency critical.
- Inflation and gaming risk. When agents know NPS is tracked and linked to performance, coaching callers toward high scores ("I hope you can give us a 9 or 10 today") becomes a risk. Score inflation removes the diagnostic signal from the metric and should be monitored as a data-quality indicator.
Responsible NPS use
Eight practices that improve the reliability and actionability of NPS programs:
- Always pair the NPS question with a follow-up open-text question asking why.
- Segment results by contact type, call queue, and customer segment — do not rely on overall scores alone.
- Track trend over time rather than optimizing for a snapshot number.
- Operate a closed-loop program to follow up with detractors within 24–48 hours.
- Do not coach agents toward a specific score — monitor for coaching behavior as a data quality issue.
- Validate that survey methodology is consistent across measurement periods before drawing trend conclusions.
- Correlate NPS with FCR and repeat contact rate to understand what operational factors drive loyalty movement.
- Align NPS measurement timing to the interaction type being assessed — relationship NPS and post-service transactional NPS require different timing protocols.
