I wrote this document a while back and it used to be sent out to all the GDC speakers each year before the conference. I figured I'd post it here; it's still the technique I use to prepare, especially the focus on rehearsals in front of people. One year in response to the mailing, Chris Crawford, an excellent public speaker, sent me some notes on his technique for preparing lectures which is different in some imporant ways, so I've included that below as well for comparison.
I'm also working on a related page on my presentation style, which has been changing over the years.
It occurs to me, after seeing these documents sitting under the title How to Give a Good Presentation, that they're more about how to rehearse and prepare yourself to deliver the presentation, not how to make one in the first place, which has more to do with what content you choose, and how you arrange it. I'll write about that at some point as well.
Some Tips on Giving a Good GDC Talk
by Chris Hecker
Below are some suggestions for making your GDC talk better, and hopefully as good as the people in the audience want it to be! The GDC Advisory Board is dedicated to increasing the quality of the conference, and that means increasing the quailty of the lectures. Please read this short document, even if you're an experienced speaker! I guarantee it will improve your talk.
Expectations & Takeaway
Make sure you reread what you promised to the attendees when you signed up. The abstracts and takeaways are on http://www.gdconf.com. We're trying very hard to narrow the gap between the title/abstract and the actual talk. Attendees are in your lecture because they read your abstract and your takeaways. You should deliver on those promises. If you said it was an advanced lecture, make it an advanced lecture. If you said you'll give technical details, give them. You should cover the pros and cons of issues...what didn't work is as important as what did. Cover prerequisites quickly and clearly, and get to the meat of your talk. Have the rehearsal audience (see below) read the abstract and takeaways and have them help you make sure your talks delivers on its promises. If you spoke in previous years, look at the feedback you got and what people said about your talk and what they expected.
How to Rehearse
Rehearsing is the most important thing you can do to give a good talk.
The most important thing about rehearsing is that it has to be in real-time (meaning a 1 hour talk takes 1 hour to rehearse), and it has to be in front of an audience of actual people (even just one person). It is helpful to get audience members who are your target for the actual talk, but any technical peer will do in a pinch. Your audience should keep track of the time and take notes on paper (tell them to be brutally honest; note taking should refer to the slide title if appropriate so during the post mortem you can find where the issue was). They should NOT interrupt you when you're talking to discuss meta-issues with the talk, they should use the notes to remind themselves of what to say during the post mortem. If you have something you want to remember, tell them quickly to write down a word that will jog your memory, and then keep going. If you plan on encouraging questions during the talk, they should ask some. If you plan on leaving questions until the end, say that!
Trying to emulate the real lecture timing and focus as much as possible is key to getting the most out of a rehearsal. You should use the same audio-visual materials that you plan to use in the real lecture...set up your laptop pointing towards the people, or move your desktop so its monitor can point towards the audience. If you plan to demo something during the real lecture, demo it during rehearsal, so you can know how long it really takes and be prepared. YOU SHOULD STAND UP TO REHEARSE (unless you plan on sitting at your talk, which doesn't work, so stand up anyway! :). Yes, you look sort of dorky standing up talking loudly to a single sitting person, but it better prepares you for the real lecture. You should not just look at the one or two people, you should look out at the rest of the "real audience" that you'll have in the actual talk. Occasionally make eye contact with the real people, of course. :)
I find one of the most useful things about rehearsing is that it lets you figure out and practice the verbal phrases you will use to describe a concept when you're up there for real. I often know exactly what I want to say in my head, but trip up on the words, and in front of an audience that is distracting and a waste of their time at best, and makes you choke up and start panicking at worst. Simply saying the talk out loud is a huge help in this respect. During rehearsal you'll often screw up mightily and just say "I'm starting this slide over" and pretend you just flipped to the slide, and it'll come out smooth the second time. Better to do that in front of 2 or 3 people than 300.
After you're done, and you check the time, you get out your notebook. Then you have each person in the group talk to you about their notes, discuss the issues they bring up, and problems with individual slides and demos. During this time you take good notes about what they're saying. Then you go change your slides (sometimes little touchups, sometimes complete rewrites) and invite them over again for anothe rehearsal. Wash, rinse, repeat. If you make a major change, or if you sucked, rehearse again. Rehearse again anyway. If you make a minor change, and you're happy with your last rehearsal, just say the couple slides you changed to get the words out.
I have done rehearsals of talks that I was convinced were done, the rehearsals ended up completely and utterly sucking, I rewrote the talks, rehearsed again, and then went on on to give some of the highest rated talks at the conference. I sound like a salesman for Rehearsal Corp. International, but it's true.
If you live at the North Pole and simply cannot find people to sit in front of you for an hour and then talk about their notes with you for another hour (and then do it again when you change things) then call friends on the phone. Email them your slides first, and then say "next" and both of you stay in sync. I've done rehearsals over the phone. It isn't as good as in person, but it's still 100% better than not rehearsing. Yes, you should still stand up on your end of the phone. If you don't have any friends and don't have a phone, a mirror is okay, but if you can possibly manage to get a rehearsal audience, even if it's your mom, wife, girlfriend, or roommate, do it.
One more thing: rehearse!
Chris, I read with interest your pointers on rehearsing a lecture. I think they’re good — rehearsal is always the most important part of good lecturing. However, I use a different approach that I thought you might find interesting.
I never rehearse in front of people. It’s a big hassle getting people together and finding a decent venue. Besides, you can only do it once or twice. I prefer many rehearsals given alone. I wander through the house, talking out the lecture. My best lectures were the result of up to thirty such rehearsals. My worst (such as the ones I did last year at GDC — with a lecture and an all-day tutorial, there was simply too much to rehearse) got only about three or four rehearsals.
This approach permits me to prepare the lecture mentally. I don’t need to write anything down, and it’s all in my head, so I don’t need lecture notes. Lecture notes destroy a good lecture — they tie you down, making it impossible to modify it to suit the mood of the audience. Moreover, my approach allows me to develop half a dozen different ways to make each point, so that I end up picking the best.
I also disdain the use of audiovisual aids. Sometimes, you just gotta show a picture, but for the most part, I avoid them because they always break the stride of the lecture. A really good lecture flows, and you just can’t maintain the flow when you have pictures to show.
Ultimately, people are not at all rational; they “hear” the emotional content of the lecture more clearly than they “hear” its intellectual content. Therefore, the emotional content establishes the context in which they receive the intellectual content. If the emotional context is solid, then they are receptive to the intellectual content, they understand it better, and they take away more. If the emotional context is weak, then they just don’t pay enough attention to really absorb the intellectual content.
There is one drawback with this approach: it yields bimodal evaluations. By pushing emotional buttons, you run the risk of pushing the wrong buttons for a few people. They’ll be enraged, spend the entire lecture rejecting your content, and give you the worst possible ratings. On the other hand, the great majority of the audience will be pleased, will learn more, and will give you very positive evaluations. Since most speaker evaluations report on the mean of the distribution, not the mode, you tend to get somewhat lower net evaluations, but if you understand the dynamics of the evaluation system, you can brush this off with the knowledge that your mode is always higher than your mean. But then again, are you lecturing for educational impact or evaluation rankings?
One of the most important features of my lectures is their liveliness, the sense of presence that I project to the audience. I’m really there, talking to them like a real live human being, rather than a wooden automaton hiding behind a lectern, reading notes.
There’s a great deal more I could relate about good lecturing, but I just wanted to make the simple point that there are many different approaches to the problem. Of course, my methods are best suited to my experience and personality. My experience is much different from most of the GDC speakers: I have given literally thousands of presentations in my lifetime, to audiences of all kinds. I have delivered lectures to groups as large as a thousand and as small as a dozen. I have lectured to people who don’t understand English (that’s a trick!), to physicists, doctors, 7th graders, artists, grad students — just about every conceivable grouping of audiences. Thus, my techniques are probably not well suited to the average GDC speaker. But they are, as I said, interesting.