How to Get Honest Feedback on Your Speaking

Why the feedback you get now is useless

After a talk, three or four people come up and say it was great. They are the people who liked it — the ones who did not, left. So the sample is selected against exactly the information you need, and it arrives while you are still full of adrenaline and least able to hear it.

Conference organizers send a form days later, when memory has smoothed into a general impression. Colleagues who owe you a good working relationship tomorrow tell you it went well. None of this is dishonesty; it is what happens when the cost of a critical answer is higher than the cost of a kind one.

So the problem is not that people are polite. It is that nothing in the normal setup makes an honest answer cheap to give.

Ask while they are still in the room

The last two minutes of the session are the only moment when the audience remembers specifics and is all still present. A form sent on Thursday reaches the fraction who bother and asks them to recall Tuesday.

So build the ask into the talk. Last slide, one link or QR code, thirty seconds of silence while people answer. Say plainly that it is anonymous and that you read all of it. Response rates from a room you are standing in front of are far better than anything you will get by email afterwards, and the answers are specific because the talk just happened.

Two questions is the right number. Three is where people start abandoning halfway.

Ask questions that cannot be answered with praise

"Any feedback?" gets you "great talk." "What did you think?" gets you the same. The question has to make a specific, non-flattering answer the easiest one to give.

Questions that work:

"Where did you stop following?" This presupposes that it happened, which removes the implication that saying so is a criticism. Answers point at a section, and sections are fixable.

"What would you cut?" Everyone has an answer, and nobody experiences giving it as harsh. When the same section comes back from six people, that is your edit.

"What's the one thing you'll actually do differently?" This measures whether anything landed. A wall of vague answers means the talk was pleasant and changed nothing.

"What did you expect that I didn't cover?" Catches the mismatch between your abstract and your content, which is the most common reason a competent talk disappoints.

Ratings out of ten are worth one slot at most. A 7.4 average tells you nothing about what to change; free text does.

Anonymity is the mechanism, not a nicety

Attach a name to a feedback form and you are asking people to be critical of someone who might be their manager, their host, or a person they will see at lunch. Most will decline, quietly, by writing something positive.

Anonymous collection is what makes the honest answer the cheap one. Any tool that does not require respondents to sign in works — rifts.to takes answers with no account and no app, so the friction that would cost you responses is not there either. Two questions on a closing slide, answers arriving while people pack up.

The trade is that you cannot follow up on a comment you do not understand. That is worth it. Vague honest feedback beats specific polite feedback.

Read it later, and discount the extremes

Do not read the answers on your phone in the corridor. You will over-weight the harshest one for a week. Read them the next morning, all at once.

Then look for repetition rather than intensity. One person saying the middle dragged is a preference. Nine people saying it is a fact about the talk. The single furious comment and the single glowing one are both outliers, and neither should move a slide.

Keep the answers between talks. Feedback from one session tells you about one session; the same complaint showing up across four tells you about how you speak, which is the thing you were trying to find out.