Instructional design for French training courses: the steps, with a worked example
Editorial lead: The Kit Certif team
Ingénierie pédagogique, the French term for instructional design, is the process of designing the learning inside a training course: starting from the analysed need, it sets what participants will be able to do at the end, chooses the methods and activities that will teach them, plans the assessments that will check it, and organises everything into timed sequences. If you run an organisme de formation in France, it comes down to seven steps: analyse the request and the need, define the audience and prerequisites, write the objectives, align objectives, activities and assessments, choose the methods, build the session plan, then write the programme and revise it after each session. This article details each step, with a fictional example followed from start to finish and the mistakes that come up most often.

Training design and instructional design: two levels
The two French expressions are often confused. The distinction most practitioners share is this one.
- Ingénierie de formation (training design) designs the scheme: which skills problem to solve, for which population, with which pathway (one course, several modules, work-study, a certification), over what duration and with what funding.
- Ingénierie pédagogique (instructional design) designs the learning within that scheme: objectives, methods, activities, materials, order of sequences, assessments.
The French Labour Code uses neither expression, but it sets the frame: a training action « se définit comme un parcours pédagogique permettant d’atteindre un objectif professionnel » (is defined as a learning pathway leading to a professional objective), and it may be delivered wholly or partly at a distance or in a work situation (article L. 6313-2). The Qualiopi certification framework devotes its criterion 2 to it, defined in article R. 6316-1 as the precise identification of the objectives of the services offered and their adaptation to the beneficiaries, at design stage.
ADDIE, SAM: what the models are worth
Two names come up in every train-the-trainer course. ADDIE (analysis, design, development, implementation, evaluation) is a generic framework from English-language instructional systems design; it has no single author or reference version, and its effectiveness has not been demonstrated by comparative studies. It usefully reminds you to analyse before you design. SAM (Successive Approximation Model), put forward by the American designer Michael Allen, replaces linear stages with short cycles of prototype, test and correction; it, too, is a practitioner’s method.
No auditor requires you to name a model. The method below borrows from both: a complete analysis at the start, then quick trials of the trickiest sequences before the first session.
Step 1: analyse the request and the need
The request is what the client says; the need is the gap between what people do and what they should do. The commanditaire (sponsor) asks and often pays; the bénéficiaire (beneficiary) is trained. Indicator 4 of the framework requires the provider to analyse « le besoin du bénéficiaire en lien avec l’entreprise et/ou le financeur concerné(s) » (the beneficiary’s need in relation to the company and/or funder concerned).
In practice:
- hold a needs interview with the sponsor: context, specific problem situations, expected result (« what will be different in three months? »), constraints;
- ask the future participants, even briefly;
- look at the real work situations (documents, procedures, observation);
- sort out what training can fix and what it cannot (a faulty tool or an unclear instruction is not solved in a classroom);
- send a written summary that the sponsor validates.
A legal limit frames this step: information requested from a candidate may only serve to assess their ability to follow the course and must have « un lien direct et nécessaire avec l’action de formation » (a direct and necessary link with the training action) (article L. 6353-9). On disability, ask about adaptation needs, never about a diagnosis. The method and the evidence auditors expect are detailed in our article on training needs analysis.
Example (fictional). A services company, which we will call company A, asks for « a complaint-handling course » for eight customer advisers who answer the phone, over two days. The analysis shows two training needs (handling a call from an angry customer, applying the goodwill-gesture procedure) and one non-training need: that procedure is recent, has never been explained, and must first be clarified by the department manager. Three advisers joined less than a year ago.
Step 2: define the audience and prerequisites
The target audience describes the people the course is designed for. Prerequisites are the entry conditions without which they will not benefit from the course: they must be necessary, verifiable and honest. A contract with an individual must state, on pain of nullity, « le niveau de connaissances préalables requis » (the required prior knowledge) (article L. 6353-4), and prerequisites are part of the public information required by indicator 1.
Example (fictional). Audience: customer advisers handling inbound calls. Prerequisites: working in a phone-based customer role and using the company’s ticketing tool. The three newcomers receive a sheet on the clarified procedure one week before.
Step 3: write the objectives
There are three levels: the training goal (what the person will do at work), the learning objectives (what they will be able to do at the end of the course) and sequence objectives. Indicator 5 requires that the provider defines « les objectifs opérationnels et évaluables de la prestation » (operational and assessable objectives).
A solid objective fits in one sentence: an observable action verb, conditions (from what, with what), and a success criterion. « Comprendre », « connaître », « maîtriser » or « être sensibilisé à » (understand, know, master, be made aware of) cannot be observed and should be banned. The revised Bloom’s taxonomy (Anderson and Krathwohl, 2001) helps you pick the verb at the right level: remember, understand, apply, analyse, evaluate, create. Our article on writing learning objectives gives verbs and corrected examples, and the indicator 5 page explains what the audit checks.
Example (fictional). Training goal: handle phone complaints through to a solution that complies with the procedure, without needless escalation to the manager. Learning objectives: conduct a complaint call following a four-step script, in a role play, completing all four steps; choose the goodwill gesture that applies to a case and justify it, with no error on at least four cases out of five; respond to an aggressive customer by reframing the exchange, validated on an observation grid; write a complaint summary a colleague can act on, rated on a three-criterion grid.

Step 4: align objectives, activities and assessments
This is the step that makes the difference. Each objective must be practised through an activity that exercises its verb, and checked by an assessment that tests the same verb. The researcher John Biggs formalised this principle of « constructive alignment » in « Enhancing teaching through constructive alignment », published in 1996 in Higher Education (https://doi.org/10.1007/BF00138871): objectives are expressed as verbs, then teaching activities and assessment tasks are chosen to engage them. The article is about higher education, but it is a rule of design consistency that applies to any course.
The simplest tool is a matrix, one row per objective, filled in before the session plan:
| Objective (fictional example) | Activity | Assessment | Criterion |
|---|---|---|---|
| Conduct a call in four steps | Demonstration, then observed role plays in threes | Final role play on a new case | Four steps completed |
| Choose the goodwill gesture | Ten cases studied in pairs | Five individual cases | Four out of five justified |
| Respond to an aggressive customer | Phrase workshop, short drills | Observation in the final role play | Grid validated |
| Write a usable summary | Analysis of summaries, writing, peer review | Summary of the final case | Three criteria met |
In the example, the matrix reveals that a morning planned on « the psychology of the angry customer » served no objective: it is cut to twenty minutes in favour of role plays. Entry assessment and assessment of learning are designed from the same objectives; see our articles on entry assessment and on assessing learning outcomes.
Step 5: choose the methods
Three families complement each other: expository (the trainer explains), demonstrative (the trainer shows, the participant reproduces), active (the participant searches, produces, practises). A short input is still useful to set a rule, but an objective at « apply » level or above is not reached without practice. On this point, the meta-analysis by Freeman and colleagues, published in 2014 in the Proceedings of the National Academy of Sciences (https://doi.org/10.1073/pnas.1319030111), concludes from 225 studies of undergraduate science, engineering and mathematics courses that active learning improves exam performance and lowers failure rates compared with traditional lecturing. That population is not adult vocational training: the transfer is reasonable, but it remains an extrapolation.
Two regulatory points. For distance learning, delivery includes technical and pedagogical support, information on the activities to complete and their average duration, and assessments (article D. 6313-3-1); from 1 November 2026, indicator 19 also requires the provider to check « l’effectivité de leur suivi par les apprenants » (that learners actually follow the modules). Training in a work situation (AFEST) has its own rules (article D. 6313-3-2), detailed in our article on AFEST. For how adults learn, see also our article on andragogy.
Step 6: build the session plan sequence by sequence
The session plan (déroulé pédagogique) is the trainer’s working document: one row per sequence, with time slot, objective served, method, materials and check. It is built from the matrix: design the main activity of each sequence first, then the minimum input needed. The right granularity runs from about twenty minutes to about an hour and a half. Each half-day opens with an activation and closes with a summary, and the total must match the duration sold exactly.
Example (fictional), day 1 at company A: 9:00 opening and entry assessment on a mock call; 9:45 demonstration of the four-step script; 10:30 role plays in threes with an observer; 13:30 presentation of the clarified procedure; 14:15 ten goodwill-gesture cases; 16:15 summary. Day 2 covers the aggressive customer and the written summary, then the final assessment. Before the session, the trainer tests the trickiest role plays with two colleagues and corrects the instructions. The standard template is detailed in our article on the session plan.
Step 7: write the programme, then revise
The programme presents the course to the public and the client. No article of the Labour Code in force speaks of a « programme préétabli » any more: article L. 6353-1 now deals with the training agreement. But several texts require its content: the agreement for a funded course (title, objective, content, means, duration, period, share delivered at a distance, arrangements for delivery, monitoring and sanction, under article D. 6353-1), the contract with an individual (L. 6353-4), and the information given before final enrolment (article L. 6353-8). From 1 November 2026, indicator 1 adds the type of recognition the course leads to and the teaching and funding arrangements, and prohibits any statement likely to mislead the public. The headings are detailed in our article on the training programme.
The programme also states accessibility for people with disabilities: possible adaptations and the contact of the disability officer, rather than a bare « contact us ».
Finally, after each session, compare results per objective, timing gaps and feedback, decide on adjustments and record them with a new dated version. From 1 November 2026, indicator 32 refers to a « démarche d’amélioration continue » (continuous improvement process) together with a risk analysis on the quality of the training delivered.
Example (fictional). After the first session, two participants do not reach the written-summary objective and the matching sequence overran. The trainer adds a commented example before the writing exercise and moves the session plan to version 1.1, without changing the programme.
The most common mistakes
- Starting with content (« everything we need to say ») instead of the need and the objectives: you end up with a generic programme.
- Unobservable objectives: « comprendre », « maîtriser », « être sensibilisé à » make assessment impossible.
- Assessment disconnected from objectives: a recall quiz to check an objective such as « conduct an interview ».
- A lecture-only session plan, or one whose total does not match the duration sold.
- Promising teaching « adapted to each person’s learning style »: the benefit of such matching is not established by research; adapt to the content, the prerequisites and the real constraints instead.
- A programme that does not keep up: content changed without a new version, or a programme that differs from the agreement and the attendance sheets.
- Accessibility reduced to a single line with no adaptation and no named contact.
Frequently asked questions
+What is the difference between ingénierie de formation and ingénierie pédagogique?
Ingénierie de formation (training design) designs the scheme: which skills problem to solve, for whom, with which pathway, duration and funding. Ingénierie pédagogique (instructional design) designs the learning inside that scheme: objectives, methods, activities, sequences and assessments. The two levels fit together, and an independent trainer often does both.
+Is the ADDIE model mandatory?
No. No French text and no indicator of the national quality framework requires a design model. ADDIE (analysis, design, development, implementation, evaluation) is a generic framework from professional practice, useful to organise the work, but it is neither an obligation nor a model whose effectiveness has been demonstrated. The auditor checks that design work took place and shows in your documents.
+What does the Qualiopi audit check on instructional design?
Criterion 2 of the framework covers the precise identification of objectives and the adaptation of services to their audiences at design stage. The auditor looks at the needs analysis (indicator 4), operational and assessable objectives (indicator 5), adapted content and delivery (indicator 6), alignment with the target certification where relevant (indicator 7) and entry assessment (indicator 8).
+Should teaching be matched to each trainee's learning style?
Research does not support the idea that matching instruction to a style (visual, auditory, kinaesthetic…) improves learning. It is more useful to adapt the method to the content being learned, to the participants' prerequisites and to their real constraints, such as a disability.