In short: the Net Promoter Score (Reichheld 2003) is proposed as the "ultimate question" predicting company growth. Peer-reviewed research (Keiningham et al. 2007, Morgan & Rego 2006) finds weak correlations (R²=0.04-0.24) between NPS and revenue growth, lower than other metrics such as repeat purchase rate, share of wallet, customer satisfaction. NPS remains useful as relative tracking over time, not as an absolute predictor. It continues to be used out of C-suite inertia and ease of calculation, not for scientific validity.
What NPS is and where it comes from
The Net Promoter Score was proposed by Frederick Reichheld in "The One Number You Need to Grow" (Harvard Business Review 2003) and consolidated in "The Ultimate Question" (2006, updated 2011). The metric is simple: a single question — "On a scale of 0-10, how likely are you to recommend this company to a friend or colleague?" — and a formula: % Promoters (9-10) − % Detractors (0-6) = NPS, ranging from -100 to +100.
Reichheld's promise: NPS correlates with company growth better than any other customer satisfaction metric. It is "the one number" the CEO must monitor. The framework was adopted by Apple, Allianz, GE, and thousands of mid-market companies in 2005-2015.
The promise: ultimate question and growth
Reichheld's 2003 argument was based on a study of 14 companies in 6 sectors, where NPS showed the highest correlation with 1-3 year revenue growth. The operational framework is simplified: measure NPS, identify detractors, work to convert them, monitor the trend over time. Bain & Company (Reichheld's consulting firm) has built an industry of NPS certifications, software (Medallia, Qualtrics), transformation playbooks.
The success of NPS was amplified by three factors: simplicity (one question, immediate calculation), narrative (the "ultimate question"), Apple endorsement (Steve Jobs had adopted it). NPS entered boardrooms as a strategic KPI on a par with EBITDA.
What peer-reviewed research says
The most cited study is Keiningham, Cooil, Andreassen, Aksoy ("A Longitudinal Examination of Net Promoter and Firm Revenue Growth", Journal of Marketing 2007). The authors replicated Reichheld's analysis on 21 industries with rigorous longitudinal data. Result: the correlation between NPS and revenue growth is R² = 0.04-0.21, significantly lower than that of other metrics (American Customer Satisfaction Index, repeat purchase intent, recommend intent metric). NPS does not emerge as a superior metric.
Subsequent study Keiningham et al. (Journal of Consumer Research 2008): the "difference" NPS computes (Promoters − Detractors) introduces statistical noise compared to using the direct mean of the 0-10 score. The NPS transformation loses useful information.
Morgan & Rego ("The Value of Different Customer Satisfaction and Loyalty Metrics in Predicting Business Performance", Journal of Marketing 2006) compares various satisfaction and loyalty metrics across 200 companies. Conclusion: top-2-box satisfaction and share of wallet are superior predictors to NPS for revenue growth, market share, ROI.
A meta-analysis (van Doorn, Leeflang, Tijs, "Satisfaction as a Predictor of Future Performance", International Journal of Research in Marketing 2013) finds measurable but small effects: the variance in growth explained by customer satisfaction metrics (NPS included) is 5-15%, far from the "ultimate" suggested by Reichheld.
Why it is still used
(1) C-suite path dependency. NPS entered executive dashboards 15 years ago. Removing it requires changing routines, KPI bonuses, board reporting. The inertia is enormous.
(2) Operational simplicity. One question, one number. Easier to explain to non-technical stakeholders than "top-2-box satisfaction on a Likert scale weighted by recency".
(3) Consulting industry. Bain, Medallia, Qualtrics, Hotjar, dozens of other platforms have built business models on NPS. Their incentive is to keep the framework alive.
(4) Relative tracking is useful. Even if absolute NPS does not predict growth well, the same company's NPS over time (delta) is an indicator of customer-experience trend. For this function, NPS works.
The Apple and Tesla anomalies: coincidence or cause?
NPS supporters often cite Apple (NPS > 70) and Tesla (NPS > 90 in 2018) as evidence that high NPS = growth. But these are unreplicated anomalies. Apple has high NPS because it has highly self-selected customers (Apple fans are fans), not because NPS predicts growth. Same for Tesla pre-2020.
When you control for brand love (a confounding variable) and look at the cohort of normal customers, NPS does not emerge as a predictor. Apple grows because of product innovation, distribution, decade-long brand building. Tesla grew because of EV technological advantage. Their NPS is effect, not cause.
Alternatives: what to measure instead of (or alongside) NPS
(1) CSAT (Customer Satisfaction Score) — top-2-box: % of respondents giving 4 or 5 on a 1-5 scale. More predictive than NPS in many studies (Morgan & Rego 2006). Works better after individual interactions (post-checkout, post-call).
(2) CES (Customer Effort Score) — "How much effort did you have to put in to solve your problem?" On a 1-7 scale. Dixon, Freeman, Toman (HBR 2010) show that CES predicts loyalty and repeat purchase better than NPS in customer-service contexts.
(3) Repeat Purchase Rate / Share of Wallet — the most predictive data: what customers actually do, not what they say. Requires CRM tracking, but eliminates self-report bias.
(4) Direct Recommend Intent — "Have you recommended us in the past?" or "Who have you talked to about our product in the last month?" Measures stated behaviour, not abstract intent.
(5) Brand Health Tracker (Romaniuk) — periodic survey of 200-500 target consumers with questions on mental availability, distinctive brand assets, CEP coverage. Richer in insight than a single NPS.
When NPS makes sense (relative, not absolute)
NPS works as a relative tracking metric over time for the same company. If your NPS drops from 45 to 30 in 3 quarters, it is a credible signal of customer-experience erosion. If it is 50 vs sector average 35, it is a signal of relative advantage.
NPS does NOT work as an absolute cross-industry benchmark, single predictor of future growth, sole strategic KPI, top-management compensation lever. In these uses it introduces decision distortions.
Metric comparison table
| Metric | Growth predictiveness | Cost | Use case |
|---|---|---|---|
| NPS | Low (R² 0.04-0.24) | Low | Relative CX tracking |
| CSAT | Medium | Low | Post-interaction |
| CES | Medium-high (CS) | Low | Customer service |
| Repeat purchase rate | High | Medium (CRM) | Real loyalty |
| Share of wallet | High | High (panel) | Competitive category |
| Brand health tracker | High | Medium-high | Brand strategy |
How to "stop" relying on NPS without causing friction
Step 1. Don't remove NPS from dashboards immediately. Add predictive secondary metrics (repeat rate, CSAT, share of wallet).
Step 2. Explain the evidence-based rationale to management. Cite Keiningham et al., Morgan & Rego. Position NPS as "relative CX tracking", not as "predictor of growth".
Step 3. Gradually shift compensation KPIs from NPS to outcome metrics (revenue, retention, share of wallet).
Step 4. Publish internally the "metric stack" framework (NPS + CSAT + CES + repeat rate) as a substitute for the "ultimate question".
FAQ
Is NPS "pseudoscience"?
More accurately: NPS is a metric with weak scientific basis when used as an absolute predictor of growth. Reichheld's 2003 original promise is not confirmed by peer-reviewed replication. As relative tracking over time it is a legitimate tool. Distinguishing fragile use (absolute strategic predictor) from valid use (CX trend tracking) is critical.
Can I remove NPS from the board dashboard?
Difficult in many contexts. Better strategy: keep it as one of the tracking metrics, but pair it with more predictive metrics (repeat rate, share of wallet, CSAT) for strategic decisions. Shift the decision weight onto outcome metrics.
If NPS is weak, why does Bain & Company keep defending it?
Structural conflict of interest: Bain has built a huge consulting practice on NPS. It does not mean they are in bad faith; it means peer-reviewed criticism struggles to cut through promotional narrative. Critical readers should read Reichheld with caution and Keiningham for counter-evidence.
Does B2B NPS work?
It works worse than B2C, because in B2B the number of respondents is low (few client companies), statistical noise is high, and "recommendation" does not translate into direct growth as in B2C word-of-mouth. For B2B, repeat contract rate, expansion revenue, retention are more solid predictors.
Are there more rigorous adaptations of NPS?
Yes. NPS 2.0 (Reichheld & Markey, 2011) introduces "earned growth" as a complement. Net Easy Score (Forrester) replaces recommendation with effort. Customer Effort Score 2.0 (Dixon 2017) refines CES. All reduce the problem of the Promoter-Detractor transformation. Peer-reviewed validation is still limited.
How can I measure customer experience without NPS?
A stack of three metrics: (1) CES post-interaction (customer service, checkout); (2) periodic CSAT (monthly, sample 200+ customers); (3) repeat purchase rate / share of wallet from CRM (monthly, behavioural data). For brand health, add a six-monthly brand tracker (200-500 target consumers).
Sources and references
- Reichheld, F. — "The One Number You Need to Grow" (Harvard Business Review, December 2003)
- Reichheld, F. — "The Ultimate Question 2.0" (Harvard Business Review Press, 2011)
- Keiningham, T., Cooil, B., Andreassen, T., Aksoy, L. — "A Longitudinal Examination of Net Promoter and Firm Revenue Growth" (Journal of Marketing, July 2007)
- Keiningham, T. et al. — "The Value of Different Customer Satisfaction and Loyalty Metrics" (Journal of Consumer Research, 2008)
- Morgan, N., Rego, L. — "The Value of Different Customer Satisfaction and Loyalty Metrics in Predicting Business Performance" (Journal of Marketing, 2006)
- van Doorn, J., Leeflang, P., Tijs, M. — "Satisfaction as a Predictor of Future Performance" (International Journal of Research in Marketing, 2013)
- Dixon, M., Freeman, K., Toman, N. — "Stop Trying to Delight Your Customers" (Harvard Business Review, July-August 2010)
- Romaniuk, J. — "Better Brand Health: Measures and Metrics for a How Brands Grow World" (2023)
- Bain & Company NPS resources: bain.com/insights/topics/customer-loyalty
