How to evaluate a training course: the three moments that matter
A satisfaction sheet measures how the day felt, not what was learned. Here is how to evaluate training before, during and after: and what each moment actually tells you.
Most training is evaluated once: a sheet handed out in the last ten minutes, filled in by people already thinking about their train home. That document measures something real (the mood of the room) but it says nothing about what was learned, and less still about what will survive the month.
Evaluating a course seriously means measuring three moments rather than one. Each answers a different question, and none replaces the others.
Before: establish the starting point
Without a baseline, a good score at the end proves nothing. A group answering 80% of questions correctly on the last day may well have answered 78% correctly on the first. The progress is the information; the final level alone is not.
A short questionnaire beforehand (ten questions, five minutes) is enough. It has a second effect that is often underrated: it shows the trainer what the group already knows, and saves an hour spent on something everybody has understood.
During: see what lands, while you can still go back
This is the most useful moment and the least used. An evaluation that arrives only at the end teaches the trainer what they should have done differently: too late for the group in front of them.
Asking a question after each block, and reading the room's answer immediately, changes what the exercise is: it stops being marking and becomes steering. If 40% of the group has a concept wrong, it gets retaught now, not at the next course.
Mid-session evaluation is not a test. It is the only feedback that arrives early enough to change anything.
After: measure what stayed
Delayed evaluation happens weeks after the course. It measures the only thing the client who paid actually cares about: what survived the return to the job.
It is also the one almost nobody runs, because it means bringing back a group that has dispersed. Response rates collapse the moment somebody has to create an account, install something or recover a password. A link that opens and is done in two minutes is the difference between 60% and 8%.
What each moment tells you
- Before: the group's real level, and what there is no point labouring.
- During: what just landed, and what to reteach within the hour.
- After: what held, and what evaporated on the way back to the desk.
Together these give a curve rather than a mark. A curve can be shown to whoever paid; a satisfaction score of 4.2 out of 5 can be shown to nobody.
Per competency, not on average
One overall mark hides exactly what you are looking for. A group at 70% may have understood three ideas perfectly and the fourth not at all: and it is the fourth that will cost something.
Tagging each question with a competency lets you read progress competency by competency, and know not that "it went well" but that "the regulatory part landed, the negotiation part did not". That is the level of detail at which an evaluation becomes something you can act on. We go into it in measuring what was actually learned.
In practice
SkillLens runs all three moments in one flow: a quiz beforehand, steps released one at a time during the session, and, on the paid plans, an individual follow-up sent weeks later with no account for attendees to create. You see it, participant by participant, beside their answers from the session.