What the feedback form at the end of training does not tell you

Most training ends the same way. A form appears, people rate the session out of five, the average comes back somewhere around four, and that number goes into a report as evidence the training worked.

It is evidence of something. It is not evidence of that.

What the form is actually measuring

Three things are working against it, and none of them are the form's fault.

It is collected at the worst possible moment. People rate the session in the last five minutes, while the trainer is still in the room, before anyone has tried to use any of it. At that point the only thing available to judge is how the session felt.

Feeling fluent is not the same as being able to do it. A confident presenter working through worked examples produces a room that follows every step and concludes the material is straightforward. Following someone else's reasoning is genuinely easier than generating your own, and the gap between the two does not show up until somebody is alone with the task a week later.

Nobody wants to give a person a two. The rating is attached to an individual who has just spent a day trying to help, and most people round up. That is a social fact, not a measurement problem you can fix with better wording.

So the number is real, it is just answering a different question: did this feel like a reasonable use of a day. Worth knowing. Not the same as whether anyone can now do something they could not do before.

1. Ask what they will do differently, in their own words

Replace one rating with one open question: what will you do differently next week because of this.

The answers sort themselves. Specific answers ("stop writing the ticket before I have talked to the requester") mean something landed. Vague ones ("communicate better") mean the session produced a feeling rather than a change, which is exactly what the five-out-of-five was hiding.

It is also the cheapest question on this list. It costs one box and it is worth the other five.

2. Test recall a week later, not at the end

A check at the end of the session measures whether it is still in working memory, which it always is. That is why end-of-session quizzes flatter training.

Ask the same handful of questions a week later and you learn something else entirely: what survived. The drop between the two is not a failure of the training, it is the normal shape of forgetting, and seeing its size tells you whether one session was ever going to be enough.

3. Ask the manager, not the attendee

The person who took the training is the worst-placed observer of whether their own behaviour changed. The person who works next to them is the best-placed one.

A single question to a manager a month later, phrased about behaviour rather than opinion, beats any number of self-reported ratings. "Have you noticed anyone doing X differently" gets a real answer. "Was the training valuable" gets a polite one.

4. Put the check inside the session, where you can still act on it

Everything above happens after the room has emptied, which means it improves the next session rather than this one.

Checking understanding during the session is the only feedback that can change the outcome for the people in front of you. Three questions after each module tells you which part needs a second pass while a second pass is still possible. This is the one that matters most and the one that gets skipped, because it takes time out of the material.

5. Count what changed, not what was delivered

Attendance, completion rates and satisfaction scores are all measures of delivery. They are easy to collect and they are what most training reports contain.

Pick one thing the training was supposed to change and measure that instead, even crudely. Fewer tickets of a particular kind. Fewer escalations. Faster onboarding to first commit. One imperfect number about the work beats five precise numbers about the session.

What to change next time

Do not redesign your evaluation. Add one open question about what people will do differently, and ask the same three recall questions a week later.

If the answers to the first are vague and the recall has collapsed, you have learned more from two small changes than the last two years of averages told you.

In defence of the form

The feedback form is not useless and it should not be thrown away. It reliably catches the things it is actually good at: a room that was too cold, a pace that lost half the audience, material pitched at the wrong level, and a trainer who genuinely should not be in front of a room.

Those are worth catching. Just stop filing the average as evidence that anyone learned anything, because that is the one thing it was never able to tell you.