Delayed (cold) evaluation in vocational training: is it required for Qualiopi?

Many people setting up a training organisation hear about “delayed evaluation” (évaluation à froid) while preparing for their Qualiopi audit, without knowing whether the framework actually requires it. The answer is nuanced: no indicator formally requires it, but it remains one of the few tools that concretely proves a training course had a lasting effect — an argument both auditors and funders particularly appreciate.

What the National Quality Reference Framework actually says

The current framework (V9) never uses the term “delayed evaluation” (évaluation à froid). The closest indicator is indicator 11 on evaluating whether objectives are met: it requires measuring, at the end of the course, whether the operational objectives announced under indicator 5 have been achieved. By construction, this evaluation happens “hot”, right at the end of the course — a final quiz, a role-play exercise, a graded case study.

Nothing obliges you to go back to the learner weeks later. But two dynamics push providers to do it anyway in practice:

• Indicator 30, on collecting stakeholder feedback, doesn’t impose a precise timeframe either. A second, delayed collection naturally complements the first without replacing it.
• Funders (OPCO, France Travail, employers) increasingly ask, even outside the Qualiopi framework itself, for proof of real impact to justify their training budgets. A documented delayed evaluation then becomes as much a commercial argument as a piece of quality evidence.

Why the hot/cold distinction has a scientific basis

The distinction between evaluating a training course “hot” and measuring its effects “cold” doesn’t come out of nowhere: it extends academic work on learning transfer, meaning an individual’s actual ability to apply on the job what they learned in training. A study by researcher Rachel Chauvin, presented in 2018 at the Rencontres Jeunes Chercheurs en EIAH, proposes a scale measuring self-efficacy beliefs about learning transfer, built specifically to capture what an immediate end-of-course evaluation misses (see the study on Google Scholar). This work highlights a point central to any training provider: a high satisfaction score at the end of a session says nothing about whether the skills are actually used months later — which is exactly why it’s worth going beyond hot evaluation whenever possible.

Building a realistic delayed-evaluation process

A delayed-evaluation process doesn’t need to be heavy to hold up in an audit. A few principles are enough:

1. Pick a timeframe that matches the course. Three months for a short, operational course, up to six months for a longer programme or a deeper change in practice.
2. Target behavioural questions, not declarative ones. Rather than “were you satisfied with the training?”, prefer “have you applied this skill since the training?”, “how often?”, “what obstacles did you face?”.
3. Cross-check two viewpoints when possible. The learner’s own perception, and, when accessible, that of their manager or funder, to limit the overestimation bias that often comes with self-assessment.
4. Automate the send-out. A simple calendar reminder (spreadsheet, learner-relationship tool, or an LMS campaign feature) is enough to avoid depending on a manual step that gets forgotten.
5. Document even a low response rate. A 20-30% response rate remains usable in an audit if it’s acknowledged and analysed, as long as you don’t claim a representativeness it doesn’t have.

What an auditor looks at if you present one

A Qualiopi auditor never requires seeing a delayed evaluation — but if they find a trace of one in your file, they examine it closely because it goes beyond the strict regulatory minimum. They check that:

• the questionnaire clearly distinguishes transfer measurement from plain satisfaction, so it isn’t confused with a duplicate of indicator 30;
• results, even partial ones, are analysed rather than simply archived;
• there’s a link to indicator 32 on continuous improvement: a negative delayed result should be able to trigger a programme revision, a pedagogical adjustment, or additional support;
• the process is repeatable from one course to the next, not a one-off exercise done solely for the audit.

Delayed evaluation for VAE, skills assessments, and apprenticeships

The principle adapts depending on the type of service. For a skills assessment (bilan de compétences), delayed evaluation focuses on whether the professional project formalised in the summary document was actually implemented. For VAE (accreditation of prior experience), it can mean checking, several months after the panel review, how the certification obtained is actually used on the job or in a job search. For apprenticeships, it looks more like employment-outcome tracking, often already collected by the CFA for other purposes.

The most common mistakes

• Confusing delayed evaluation with a second satisfaction survey: the two measure different things and don’t substitute for one another.
• Promising a delayed-evaluation process in your training programme and never running it: an unfulfilled written commitment is more damaging in an audit than an honestly acknowledged absence of one.
• Sticking to a closed question like “did you use the training?”, which produces nothing analysable.
• Collecting no data across multiple courses, which prevents any comparison over time or between trainers.

Take action

The Complete Qualiopi Kit provides a ready-to-customise delayed-evaluation questionnaire template, along with tracking tools tied to indicators 11 and 32 (€297, 14-day guarantee). The “Create your training organisation in 30 days” ebook helps you build your evaluation process right from the design stage of your programmes, and the complete pack brings both resources together.