NPS (Net Promoter Score) is a single-question survey metric. The question is fixed: “How likely are you to recommend [brand] to a friend or colleague?” on a 0–10 scale. Respondents bucket into detractors (0–6), passives (7–8), and promoters (9–10). The headline number is the percentage of promoters minus the percentage of detractors, a value between -100 and +100. Passives count toward the response base but never enter the subtraction — the part operators most often misremember.
How the math works
Net Promoter Score
NPS = % Promoters − % Detractors
Take 100 respondents split 45 promoters (9–10), 35 passives (7–8), 20 detractors (0–6). NPS is 45% − 20% = +25. The 35 passives sit out of the subtraction entirely, which is why a brand with mostly 7s and 8s and a thin layer of 9s posts a modest positive number rather than the strong score the raw distribution suggests.
Where it came from, and how to read it
Fred Reichheld introduced NPS in a 2003 HBR article (“The One Number You Need to Grow”), and Bain & Co popularized it as a single-number proxy for customer loyalty. It became standard because it’s cheap to collect and easy to compare year over year — not because it predicts retention or LTV better than the alternatives. Academic work has pushed back on the original growth-prediction claim for years; the metric still earns its keep, but as a tracking signal, not as the forecast it was originally pitched as.
Three uses for DTC operators
Post-purchase NPS as a customer-experience canary. Fulfillment issues, packaging damage, and product regressions land here before they reach returns or churn data.
Cohort-level NPS as a brand-strength signal. Tracked across acquisition cohorts (month, channel, campaign), it’s harder to game than aggregated review averages and reads how each cohort feels a few weeks in.
NPS by segment to find who’s net-promoting. Sliced by channel, SKU, tenure, or acquisition source, NPS surfaces which customer types love the brand and which tolerate it — actionable input for product mix and acquisition spend.
How NPS reads vs. CSAT
CSAT is transactional (how was this specific interaction); NPS is relational (how do you feel about the brand overall). The two are complements, not substitutes — a brand can post strong CSAT on individual touchpoints while NPS drifts down, or vice versa, and each pattern points to a different problem.
Where operators misuse it
Three patterns recur. Chasing NPS as a board KPI rather than reading it as a signal turns the survey into a target and corrupts the data. Comparing across categories where benchmarks differ wildly flattens information the score was never meant to carry. And treating a small-sample monthly movement as real change when the confidence interval swamps the delta is the most common reporting mistake. State the period, state the sample size, state the segment.