Giving a Talk

How to give a talk

So you've been asked to give a talk in front of a seminar--or possibly in front of a much larger audience. Or maybe you've been giving lots of talks, but you wonder about how you can make your talks more effective?

The purpose of this page is to present some ideas about presentation style. These are all surface issues and do not address the actual content of your talk (because it is, of course, spectacular). Every single point can be argued, so a justification for the point is given. If you disagree with a point and want to state a reason why you disagree, or have additional tips to share, please send a message to me and your comments will be included if they're reasonable.)

I would like to thank James Allen. I used to give lectures and harangues containing this material at Cornell, and James had the idea of writing it down (and making it "somewhat humorous" as some authorities on the Web dryly characterize this page). I can't promise it will be useful for you, but I can say that my students, who followed this advice, have gotten great jobs after graduating as you can see here.

What I tell my students about giving talks and writing papers is: Most students, initially, have little idea how to do either. At least I have an idea about this. Try my way for five years. One possibility is it works great for you, and you can use it. The other is it doesn't, and then you can react to it and develop your own style in response. But you can't start with no idea and "react to" that and develop your own style: if you multiply by zero you get zero. By the way, no one in my lab has ever rejected these ideas; instead they have refined them and improved them. You can too.

These tips were created in the days of overhead transparencies, but, surprisingly still people find them useful even in the PowerPoint era and beyond. Nevertheless, I have tried to update this with some comments on PowerPoint and other computer projection methods, without repeating the usual advice on "how not to use PowerPoint" (I have my own advice on how not to use it :-) And also how to use it.

  Q. Will the audience be annoyed if you sit instead of stand?
A. Yes

One of the most important things you can do is watch other speakers. Figure out what you like and what you don't like about what they do and then try to do or not do those things. Ever been annoyed at a speaker who puts of an overhead with such tiny print you can't read it at all? Then be sure you use a larger font. How about those people who constantly block the screen so you can't see it? Maybe you should try not to do the same thing.

Great speakers sometimes codify their techniques. You should begin by reading Patrick Winston's Lecturing Heuristics, and video here, called "How to speak":

Hints for a good presentation

Speak clearly. It shouldn't be too much of a shock, but people can't hear you if you mumble or talk really quietly. Most audiences are afraid of sitting too near the front of a class, either because they're worried about being called on, or because their third grade teacher spit during lectures. Remember the ones in the back (who are thinking of sneaking out early) and speak up and speak distinctly (so they'll stay).

Use large fonts. Anything smaller than 24 point is probably a mistake. If you photocopy a paper from a book and project that, you deserve severe punishment. The only exception is if you are trying to impress the audience with the density of something, or otherwise make a point that specifically requires dense and unreadable text.
  Use figures instead of words!  

Use lots of figures. A picture is worth a thousand words. If your work is very mathematical, try to develop a talk that is entirely in pictures. Then go back and add one or two words per slide.

Point to the projection (screen), not the source. You want to point out part of a picture or a bullet item on a slide to make it clearer what you mean. Walk up to the screen and point at the bullet or picture. Do not point to the transparency on the projector itself. There are several reasons for doing this:

There are occasions when you cannot reach the projection to point at it directly. Put your hand into the light and make shadow pictures: use the shadow of your hand to point at the part you want to deal with. You probably do not want to use a pointer.

Do not use a pointer. A pointer seems particularly useful if you cannot reach the projection. Those laser pointer things seem totally cool, too, don't they? Well, they're annoying and should be outlawed. Why?
Pointers are guaranteed to annoy at least 35% of your audience.  

Do not adjust the slide unless it's falling off. Ever watch someone adjust each overhead over and over again? Ever want to slap them and tell them to stop? It's pointless. Who cares if it's 10 degrees off vertical? The little jiggering of the slide doesn't make it easier for the audience to read it. And it makes you look really nervous. Get away from the projector and point at the screen. You won't be blocking the view of your audience and you won't look as nervous. Of course, if the slide's about to fall off the projector....

Be sure the projection is on the screen. How many times have you watched a speaker talk and talk and talk without ever noticing that the projection is somewhere to the left of the screen and you can't read it? You want to yell but are afraid you'll annoy people. So you should be sure it's pointing the right place. Of course, if you walk up to the screen and point at the projection, you're addressing this problem at the same time, aren't you? (Amazing how multi-purpose these tips can be.) Using large margins is helpful for this one, too, since there is less text to spill off the sides.

Be sure the text is projected at the top of the screen. This is related to the previous point, but refers more to where the text is than to where the projection is. Position the slide so that the first line of text is as far toward the top of the screen as possible. That means that people in the back can see what's on the screen even though some big-headed person is partially blocking their view. Having trouble figuring out where the slide should be lined up? Point to the screen and you'll clear up this problem, too.

Watch the time. Try not to go over your given time. Even if you start late, it's a courtesy to the audience to end as close to on time as possible. A good lecture room will have a clock positioned so that you can see it. (A spectacular lecture room will not have one positioned where the audience can see it, so they're less likely to fidget.) Pay attention to it. If you're running behind, skip a slide, or gloss over one, or talk a bit faster, or don't accept questions. Yes, your work is exciting and interesting, but your audience has other appointments, too. If not, they'll talk to you afterward.

Walk in front of the projection occasionally. This one seems kind of silly, but it serves two purposes. First, it gets you to the other side of the room so that the people on that side will have you in the way of the projection (only sometimes since you'll usually be up near the screen); it is only fair to share the discomfort. Second, the sudden bright flash of light reflecting back to the audience as you break the projection beam will wake a few people up. Seriously.

Talk to the audience, not the screen. This sounds simple, but it's amazing how many people look at the screen and talk at it rather than at their audience. If you have to face the screen, speak a bit louder while you're facing it so that your voice will reflect from it and back to the audience. Better: don't talk to the screen. Contort your body, or point at the screen and then turn around.

Do not cover up parts of the slide. The "overhead striptease" act is one of the most common and most annoying features. What in the world do you think you're accomplishing by feeding the words on the slide to the audience one line at a time? It's infuriating. It makes it harder to pay attention to the speaker, too: the audience keeps having to read a line, look back at you and listen, watch you fiddle with the slide, read another line, turn back to you, and so on and so forth. Tiresome. Why not let the audience skim the slide and then talk about it all at once? Are you afraid they'll be so busy reading that they won't hear you talk? Then make your talking more interesting. (The term "overhead striptease" is alleged to have been coined by Tufte.) Consider using an overlay transparency if you need to keep something in suspense. They're sometimes a bit hard to get lined up, but not too bad. However, beginners should use this technique sparingly, until you practice a lot and get the multiple-overlay technique to be fast, slick, and good-looking.

Modern machine-driven slide display (e.g., from Powerpoint) make it really easy to do these sort of "build up" or "multiple overlay" talks. These are a bit easier, and sometimes even quite effective.

Summary: Never cover up your slides! avoid the striptease! Overlays are often useful to build up a palimpsest of information gradually -- much better than putting up one dense hairy slide to annoy the audience. Audiences tend to like overlays pretty well.

The only thing worse than the "overhead striptease" is leaving part of the slide covered and never revealing what's under it. You will be convincing the audience that something embarrassing is under there (a naked person?). Bad move. Who cares if it's an old slide that's no longer quite appropriate; just don't talk about the extra stuff. Adds a bit of mystery to your talk, but in a nice way.

Do not read your slides to the audience. Why would I want to come to your talk to hear you read your slides? Unless you're a famous poet or novelist reading your own work, what is the point? (Not even sure there's a point then.) The slides should be an outline of the talk to help the audience follow what you're saying. Or complex equations or pictures or something that you can't convey easily with words. A simple trick is to leave out all of the articles and connectives--e.g., "simple trick: omit articles, connectives". Then if you have no better imagination, you can read it back to the audience with the articles and connections put back in. At least your presence serves a purpose then.
  Props good, fire bad.  

Use props. Talks are about show and tell and keeping your audience amused, so you can inform them painlessly about what you are doing. Whenever possible, bring and use props: videotapes, robots, pieces of robots, models of molecules, a gear your algorithm machined, circuit boards implementing your algorithm in silicon, etc. However, if you use videotapes, be sure to have them cued up beforehand and practice turning them on and off so it goes smoothly.

Use color. It used to be that you could use LaTeX and make black & white slides for a talk. This worked, because LaTeX typeset things nicely, and no one had color printers. Now we do have color printers (and copiers, and 35mm slides) and black LaTeX slides look (a) all the same and (b) boring. Monochrome slides give the impression you are not colorful either. These days, there is no excuse for a monochrome talk. Use colored pens if you are making your talk by hand. If you're using the computer, use color LaTeX or Powerpoint, or Adobe Illustrator, and print out your slides on a color printer. For better or worse, audiences these days expect color; it's easy to use, and you can convey more information with it.

Case Study: How to Commit "Talk Suicide"

Sometimes you'll be presenting results from a paper of yours, or by someone else. Here is example of how not to do it. In particular, here is an example of a terrible talk, which violates almost all of the rules above. (Actually, it is only the PowerPoint slides for a terrible talk, but the talk using the slides was equally terrible). It looks as if the student simply scanned in paragraphs from the paper and stuck them into PowerPoint. During the presentation he simply read the text and symbols. So the student behaved more like a parser, than a lecturer. Not only that, the scanned-in images are fuzzy and ugly! The entire talk is black-and-and white (no color), there are almost no figures (other than a few black-and-white line drawings scanned in from paper), there is no attempt to teach the audience something or explain the results -- the talk is merely a garbled recitation of (putative) verbatim sections of the paper. Needless to say, this is terrible: don't ever do this!

    Two slides from a truly horrible talk. The student simply scanned in paragraphs from the paper and stuck them into PowerPoint. During the presentation he simply read the text and symbols. This talk violates almost every rule on this page. It was a miserable, brain-frying experience. Don't ever do this!