I'mBoard API Docs

Net Promoter Score — % of survey respondents who are promoters (score 9–10) minus % detractors (0–6), passives (7–8) excluded. Per the original NPS methodology (Reichheld, Bain & Company, 2003). The score ranges from −100 to +100. The board reads NPS as one read on product-market fit and word-of-mouth potential, not as a precise customer-loyalty measurement — the methodology is well-known for being sensitive to sample bias, response rate, and survey timing. Common pitfall: comparing NPS across companies without normalizing for industry — B2B SaaS NPS distributions sit much higher than consumer-app NPS, and the absolute number means little without a peer cohort. — Customers KPI anchored to Retently NPS Benchmarks 2025.

Rogue ID: customers.nps_score Type: Number Domain: Customers

Definition

Formula

NPS = (% promoters, score 9–10) − (% detractors, score 0–6). Passives (7–8) are excluded from both. Range: −100 to +100. Per Bain & Company / Reichheld NPS methodology (HBR 2003, "The One Number You Need to Grow").

Why it matters

A coarse-grained directional read on customer affection and word-of-mouth potential. Sustained movement (especially regressions) is the signal the board should focus on, not absolute values — the methodology is too noisy for fine comparisons across companies.

How to interpret

Per Retently NPS Benchmarks 2025, B2B SaaS NPS medians by industry cluster around the +30 to +50 band, with top-quartile +50 to +70. Translate scores to categories: −100 to 0 = needs work, 0–30 = good, 30–70 = great, 70–100 = excellent — these category bands are widely circulated industry folk-wisdom (Bain does not publish strict thresholds). Always pair the score with sample size and response rate; an NPS based on <50 responses or <10% response rate should be flagged as low-confidence.

Calculation policy

How an AI agent should compute this KPI from messy company data. Free-text rules consumed at reasoning time — not a deterministic DSL. The most common ways to get this wrong are listed under Common miscomputations.

Inclusion rules

NPS = (% promoters, score 9–10) − (% detractors, score 0–6), per the Reichheld / Bain methodology (HBR 2003).
Compute the promoter and detractor percentages over the completed responses (customers.nps_responses).
Result ranges −100 to +100; point-in-time as of the survey window.

Exclusion rules

Passives (score 7–8) from the numerator — excluded from both promoter and detractor counts, but they REMAIN in the response base / denominator.
Partial / abandoned surveys with no usable score.
Cross-company comparison without an industry-normalized peer cohort — B2B SaaS distributions sit far above consumer-app NPS.

Required inputs

Counts of promoters (9–10), passives (7–8), and detractors (0–6).
Total completed responses (customers.nps_responses).
Survey window + response rate (the confidence qualifier).

Data-source priority

Survey platform with the raw 0–10 score distribution.
Manual tally of the 0–10 distribution as a fallback.

Edge cases

Small sample (< ~50 responses) or low response rate (< ~10%): report but flag low-confidence rather than trending.
Survey-timing / sample-bias swings: a movement driven by who was surveyed, not by sentiment — note the survey context.
Passives must stay in the denominator even though they net to zero — dropping them inflates the magnitude.

Validation checks

NPS = %promoters − %detractors — recompute against any pre-computed value; bounded −100 to +100.
%promoters + %passives + %detractors = 100%.
Always present with nps_responses and the response rate.

Common miscomputations

Excluding passives from the denominator (computing the spread over promoters + detractors only) — overstates the magnitude.
Reporting the score without sample size — a low-n swing mis-reads as a real movement.
Comparing the absolute score across industries without a peer cohort — B2B SaaS NPS is not comparable to consumer-app NPS.
Averaging the raw 0–10 scores instead of applying the promoter-minus-detractor formula.

customers.nps_trend
customers.retention_insights
customers.churn_risks
customers.key_initiatives

Source

Retently NPS Benchmarks 2025 · section: NPS Benchmarks — published 2025-01-01.

Why does this cite Retently NPS Benchmarks 2025? Read the ontology methodology for the published vs editorial tier system, attribution rules, and dispute process.