NPS in training: calculating the Net Promoter Score and using it for Qualiopi

A single question — “Would you recommend this course to a colleague?” — and a score that fits in one number: the Net Promoter Score appeals to many training organisations because of its simplicity. But an NPS that is miscalculated, confused with a satisfaction rate, or published without context on a home page does more harm than good. Here is how to calculate it exactly, what it measures — and does not measure —, and how to use it meaningfully in a Qualiopi quality system.

The NPS: one question, three categories, one subtraction

The Net Promoter Score is based on a single recommendation question, scored from 0 to 10: “Would you recommend this course to a colleague or a friend?”. Respondents are then sorted into three categories:

• Promoters: scores of 9 or 10 — enthusiastic trainees, likely to recommend actively;
• Passives: scores of 7 or 8 — satisfied but unenthusiastic, they count neither positively nor negatively;
• Detractors: scores of 0 to 6 — disappointed or critical, potentially negative word-of-mouth.

The score is obtained by subtraction: NPS = % promoters − % detractors. The result is not a percentage but a score between −100 (all detractors) and +100 (all promoters). That is the first mistake to avoid in your publications: writing “NPS of 45%” betrays a methodological confusion.

A worked example (fictitious, for teaching purposes)

Take an invented session to illustrate the mechanics, with 20 respondents to the end-of-course questionnaire:

Scores given	Number of respondents	Category	Share
9 or 10	11	Promoters	55%
7 or 8	5	Passives	25%
0 to 6	4	Detractors	20%

NPS = 55 − 20 = +35. Note that the five passives vanish from the calculation: two sessions with the same NPS can hide very different realities depending on the share of passives — one more reason to keep the detailed answers, not just the final score.

NPS and satisfaction rate: two indicators not to confuse

The NPS measures an intention to recommend; the satisfaction rate measures a judgement on the service received, usually calculated as the share of respondents scoring above a threshold (8/10 or 4/5, for instance). The two formulas, scales and questions differ — so the two figures are never interchangeable in your publications.

This is a direct point of vigilance for indicator 2, which requires dated, sourced and verifiable results indicators: displaying “92% recommendation” when the figure is actually an NPS of +92, or presenting an NPS as a satisfaction rate, means publishing an indicator whose calculation basis does not match its label. In an audit as in a sales meeting, every figure must carry its exact label: “NPS of +35, 20 respondents, first-half 2026 sessions”.

Adding it to the end-of-course questionnaire without weighing it down

The NPS requires no extra questionnaire: it is one more question — just one — in your existing end-of-course evaluation. Three precautions are enough:

• A stable wording from one session to the next, otherwise scores can no longer be compared over time;
• The full 0-to-10 scale displayed, with explicit anchors (“0 = not at all likely, 10 = extremely likely”);
• An open follow-up question — “what explains your score?” — which turns a mute number into usable verbatim feedback, especially for low scores.

The recommendation question complements the thematic dimensions of your satisfaction survey (content, delivery, practical conditions); it does not replace them. A questionnaire reduced to the NPS question alone would lose precisely what the auditor wants to see: detailed feedback from which actions can be drawn.

Using the score: trend, detractors, continuous improvement

An NPS calculated and then filed away brings nothing. Three uses give it value:

Track the trend session after session. Recorded in your dashboard with its respondent base, the NPS becomes a simple warning signal: a score dropping over two or three consecutive sessions warrants analysis, even if each score taken in isolation still looks “acceptable”.

Treat each detractor as an individual signal. A score of 0 to 6 paired with a critical comment is closer to a complaint than to a statistic: contact the trainee again, understand the grievance, record the response given — this is exactly the complaint-handling logic expected under indicator 31, and a detractor’s verbatim comment is often worth more than ten points of score.

Feed the continuous improvement loop. A documented sequence — “score dropping → analysis → corrective action → effect measured on the next session” — is direct evidence for indicator 32 and the continuous improvement process. An NPS measured weeks or months after the course usefully complements the end-of-course score: the intention to recommend evolves once the skills acquired meet the realities of the job.

The limits of the NPS: what research says

The NPS was popularised by Frederick F. Reichheld in a 2003 Harvard Business Review article, “The One Number You Need to Grow”, which presented the recommendation question as strongly linked to company growth (see the article on Google Scholar). The initial enthusiasm was later tempered by academic research: the longitudinal study by Timothy L. Keiningham, Bruce Cooil, Tor Wallin Andreassen and Lerzan Aksoy, published in 2007 in the Journal of Marketing under the title “A Longitudinal Examination of Net Promoter and Firm Revenue Growth”, did not find the claimed superiority of the NPS over other satisfaction measures for predicting growth (see the study on Google Scholar). The practical conclusion for a training organisation: the NPS is one useful indicator among others, not a miracle predictor that would replace the rest of the evaluation system.

Two operational limits add to these substantive reservations:

• Sensitivity to small samples: with 10 respondents, each answer weighs 10 points; a single detractor turns an NPS of +60 into +40. On small cohorts, combine sessions over a period before commenting on the score.
• Publication without context: an NPS displayed alone, with no period, respondent count or method, has no evidential value — the rule “one figure = one period + one base” applies to the NPS as to any results indicator.

What the Qualiopi auditor actually assesses

No indicator of the framework requires the NPS — or favours it. What the auditor checks under indicator 30 is that the organisation collects feedback from its stakeholders and uses it: questionnaires sent out, results consolidated, analyses recorded, actions that follow. Whether the measurement uses an NPS, a satisfaction rate out of 5, or both, is entirely the organisation’s choice — provided the method is defined, stable and documented. A rigorous NPS with its exploitation loop is worth more than a battery of decorative indicators; the reverse is just as true.

Take action

The Complete Kit Certif (€297, 14-day guarantee) includes end-of-course and follow-up evaluation questionnaires, indicator calculation templates and tracking dashboards to document indicators 30 to 32 — recommendation question included. Launching your organisation and building your measurement system from day one? The ebook “Create your training organisation in 30 days” (€67) guides you step by step, and the full pack (€347) combines both resources.