Irena Swanson

What makes a good talk and what makes a bad talk

I wrote the notes below around year 2000; I sound very judgemental. Since then I have seen many excellent projector talks, and technology has improved as well. I am leaving the notes as they were, however. Read with a few grains of salt, or better, do not read at all.

Content: First of all, you need to want to say something. Decide what you want to present and emphasize, and decide what you want the audience to gain from your talk. Plan the talk accordingly. Let the interests and the background of your audience guide your choice of material and emphasis. The best talks give a general idea of what the research is about and do not trouble the audience with distracting details. You may not be able to present every proposition in your paper, but hopefully you will communicate why this research is interesting. This is more beneficial to the research community. Those who are interested to know more can then go read your paper. Your job in the talk is to get them interested.

Preparation: Prepare, prepare, prepare your talk. Practice aloud. Yes, this does take time. If you think preparation is a waste of time, compare it to the waste of time for each member of your audience if they have to listen to a bad talk! A speaker should respect the audience.

Timing: Don't go overtime. Overtime may be a sign of insufficient preparation and of lack of respect. Prolonging your talk even three minutes could be a waste of 60 minutes for a small audience, and a greater waste for a larger audience. Think if presenting your last proposition is worth taking 3 extra minutes from your precious audience. If not, the audience will be grateful for the timely finish.

Level: It can happen that the talk you planned was not quite at the right level. Adjust your talk. If the audience has questions, answer them. It makes no sense to force your planned lecture when the audience cannot comprehend it. Again, respect your audience, adjust. Of course, to adjust, you need to really understand your material. Well, that's why you were invited to give the talk.

Physical presence: Face the audience as much as possible. Don't talk to the blackboard, don't read your talk, don't only look at your three best friends.

For ears and eyes: Speak loudly and clearly. Make any visual aids large enough. Make sure that your audience can hear you and can see the expected. If the audience has to strain their eyes and ears, they will sooner or later tune you out. You do not want to lose your audience this one time because this may be your only chance.

Blackboard versus technology: Technology can be wonderful for presenting tables, formulas, long results, etc. However, in general, I prefer blackboard talks. Here are some reasons:
  1. Machines buzz, hum, and are subject to poor or unpleasant light conditions. I am surely not the only person who gets headaches after a long day of overhead projector talks.
  2. I find the speakers and the audience more passive during projector talks. I prefer those talks where everybody is engaged in learning.
  3. Projector speakers tend to go fast: they get to present more material, but the audience does not necessarily absorb more. The definition of ABC is presented on page 1, then used on pages 3 and 7, but the audience is too polite (and passive) to ask for a reminder, so that much of the talk is lost on them. In contrast, a blackboard talk goes more slowly, the definition of ABC stays in plain sight longer, so the audience is better able to follow the talk.
  4. If you do use overhead projectors, put at most 100 words per page, better still, fewer than 50. Step aside. Talk about motivations, history... without reading the page directly. Pause when you are done with a page: the audience is still absorbing it. Do not fuss with the technical issues during your talk, fix those ahead of time.
  5. Typographical errors happen at the blackboard and on projected screens, but the latter are more awkward to correct. I have seen overhead projector speakers laugh them off, but it is no laughing matter for the person who copied them and wasted time trying to understand the misprinted result.
If you do use overhead projectors, use large enough letters for legibility. (In TeX, magnify five fold!) Don't block the screen with your body. Step aside, look at the screen, look at the audience. Don't read line by line from the screen -- that is really boring! You want to engage the audience into thinking, not into reading! Do not present on the screen your paper, but present your ideas, main results, motivation for the research.

If you use the blackboard, write legibly, face the audience, erase judiciously, speak while you are writing and erasing.

I've seen many good talks, and I tried to summarize above what I thought made them good. Good talks are energizing for the audience and the speaker. Please give good talks.

Here are some of my worst experiences from talks:
  1. Speaker leaves his cell phone on, and when his wife calls, he answers.
  2. Speakers gave the impression that they spent more time sightseeing and scouting for the best restaurant than preparing for the talk.
  3. Speaker started his talk to a math department as though they were all experts in the field. Audience asked questions, the speaker got annoyed. For the remaining 40 minutes, the speaker talked to himself, audience sat politely and passively.
  4. Speaker refused to answer background questions, saying: "I won't be able to finish my material." Some members of the audience got up and left.
  5. Speaker does not prepare, rambles on in no order, even takes extra time.
  6. Speaker gives the impression of not understanding more than the statements on the screen.
  7. Speaker talked about his papers' rejections and acceptances rather than the content of those papers.
  8. The speaker defined many three-letter acronyms, which made it very hard to follow the talk.
  9. A speaker at a summer camp for bright high school students gave a vague overview of a subject. A student interjected with essentially a proof of a main result, but the speaker was unable to appreciate the student's insight. The speaker displayed disrespect for his audience: he did not prepare the material deeply, he was very vague, and he did not adjust to the bright audience.