Preparing a talk for a conference


I distinctly recall a little non-event at a conference: I was scooting to catch a friend’s talk on time. I found him sitting in the hallway outside the room, slide carousel* in his lap. Grabbing a bunch of slides and putting them into his carousel. He was picking out slides, on the fly at literally the last minute. Figuring out both his content and his sequence

Then I saw his talk, which was stellar in every way. (This was no surprise, though, because he rocks.) I thought to myself, “I couldn’t give a presentation that was so conceptually crisp, unless I had gone through it far in advance.”

Slide carousels have been replaced by powerpoint presentations, and as a consequence last-minute talk construction is the norm. You don’t have to shoot your slides in advance, which was a huge annoyance in itself. But it meant that you had your final slides ready to go well before the actual meeting. Having a talk prepared days or weeks in advance is, now, the outlier.

I’m going to a meeting today, and giving a talk tomorrow. (If you’re at Goldschmidt, come on by!) And there is no frickin’ way that I’d show up at this meeting without knowing exactly what I’m going to say before I get there. I might do some small tweaks, but as far as I’m concerned, my talk has to be in the can before I arrive.

I don’t know how common my approach to preparing for my talk is, but here’s how I do it:

  • Make the slides (2-20 days in advance).

  • Think about how much time I have, and then delete half of my slides.

  • Delete all of the text from the slides and just leave pictures or figures. (Any text on slides is not advised because it causes people to read instead of listening. Well, maybe a few words are okay.

  • Practice giving the talk about out loud with a timer (a day or two in advance).

  • Restart the timer about 30 times, usually after about 30 seconds, because it takes that long to find the best few sentences. This is the most critical part of the talk.

  • Run through the whole talk 2-3 times once I have it totally down.

  • Give the talk at the meeting. Aside from the first 2-3 sentences, it barely resembles what was practiced because of going into improv mode at the last moment. But it usually seems to come out quite well.

This process takes several hours, to get a 10-12 minute talk down pat. (I do it pretty much the same way for a typical 45-50 minute seminar, too, on the occasions when I’m blessed with an invite. Next academic year is mostly open, hint hint.) I don’t want to make an ass of myself, this can only be avoided if I make sure the talk is ready before I show up at the meeting. I can improvise mighty well, I think, but it comes off a lot better if it looks spontaneous but isn’t.

Do you prepare any differently? More prep, less prep, different prep?


*This was probably around 2000. The Entomological Society of America was a late adopter of digital projection at conferences.

15 thoughts on “Preparing a talk for a conference

  1. Some good tips here. I tend to re-use slides from previous presentations, analogous to sitting with a slide carousel and choosing what to put in in the slide tray. But, like you, days in advance!

    Also I suspect I use more text than you do, but that’s for my benefit to give me prompts due to a generally lousy memory.

    Finally my rule of thumb is one slide per minute plus a maximum of 10%, not counting intro and acknowledgements. So a 15 minute presentation by me would usually have an absolute maximum of about 18 slides, including intro/outro. I’ve seen some people try to cram 25 to 30 slides into 15 minutes and it’s usually a garbled car crash.

    Oh, finally finally, avoid clever animations unless the animation is critical to the content, e.g. a slide I use in teaching and public talks showing pollen arriving on a stigma and pollen tubes germinating, moving down the style and fertilising the ovules. I was very pleased with that, though it’s looking a little old now :-)

  2. Thanks for the good tips.
    I agree that talks, seminars, presentations, lectures are theater and require rehearsing or some form of prep. There is a temptation with Powerpoint to add all kinds of things because you can. I prefer talks where the “things” support the science not just the show. My research involves analysis of high-speed videos of insect behavior. One movie can say a thousand words.
    Also the only text on my power points might be a title to orient the audience where we are in the talk or what question the data presented answers (or in small print, supporting statistics).

  3. I think that the “carousel” point is a good one. I started doing presentations at the tail end of the carousel era. At that point we had to make our slides on PPT and then print them out as physical copies. That enforced early talk preparation both due to the time required and due to the fact that the cost-per-slide wasn’t negligible. So your talk had to be mainly ready when you hit print.

    Great point by jeffollerton on animations. One more reason for that that I can add as a departmental seminar organizer for a number of years:

    If a seminar is going to be webcast (slowly becoming the norm these days) a lot of webcasting software doesn’t play well with animations when converting PPT to broadcast files. That, and the fact that simple is generally better and more certain to work on any system, argues against animations of any sort.

    Good point on removing text. When I run courses with a presentation component – and with my own grad students – I am continually harping on the fact that the less text the better. Less text means that you’ve given more thought to the content; that you’ve practiced more (not relying on the screen as your notes); and that your visuals are punchier. Some of the best talks that I’ve seen have had virtually no text at all throughout.

    Great post, Terry. Thanks for this. I’ll likely be pointing my students here often in the future.

  4. This is exactly how I prepare, right down to the improv during the talk. During my first year of teaching, I did this with every lecture also. As you can imagine, that was a huge time suck, but I didn’t feel confident going into lecture without having done it. Now, things are much more loosey-goosey with my lecture prep, but conference talks are different. Unless they are very small and regional conferences, in which case I do tend to be more lax with the prep.

    I recall going to my first conference and my roommate, a good friend and fellow PhD candidate was putting her talk together the night before in our hotel room. She looked so stressed out and couldn’t enjoy the dinner and a dip in the hot tub afterward. Thankfully, my advisor hammered the practice talks in my head, so while she was nervous for the 24 hours leading up to her talk, I didn’t even look at mine until the morning of, and that was just a glance through. I spend the night before networking and relaxing in the hot tub.

  5. Good advice for all of us and pretty much identical to what I do – although unlike Jeff I have a few more slides – but that is because I like to show a lot of pictures – if my talk is data rich then yes I would go for a slide a minute.

  6. How did one make an old-school slide? I am solidly from the Powerpoint-era.

    • Oh my gosh. In my era, what you did was design it in powerpoint. Then you bring your file to the computer that had the expensive machine hooked up to it. You loaded 35mm slide film into this machine, and it shot the slides from the computer onto the film. This was a clunky process prone to failure and weird problems. Then, you take the film to a camera shop to be developed. Then you loaded your carousel.

      Also, you could use actual pictures from a camera taken on slide film, of your field site, organism, people, and such.

      Before that era, labs would have set up a flat board with bright lights on it, and a mount for a camera. Then you’d print out your graphs, text, what have you, and take actual photos of it with a camera. Then expose it.

      There is a workaround that I also used, which I thought was really slick but people didn’t really use. What I did was take my 8.5×11 printout of my slide (like a figure) and then photocopy it to reduce it. Then I do that again. And again. And again until it’s tiny. Then I photocopy it onto a sheet of transparency paper, like we used for overhead projectors. Then you cut that out of the transparency and slide it into the frame for the slide. It worked really well and nobody suspected that I just didn’t shoot it as a regular slide.

      And I started grad school in 1994, 20 years ago. The slide projector era totally ended in the early 2000s, I think.

      • I am currently feeling very, very grateful that I started grad school in 2009. That all sounds terrible.

    • For my thesis defense back in the day, I took pictures of hand drawn (drafted with pen on vellum) figures, typed tables, and of course glorious Kodachrome slides of my field site, my birds, the fruit they ate and their habitat. I took the same slides to meetings by handing a carousel to the person “running” the session. It cost $300 an hour to use a computer station that went to a main frame. It did not have a graphics program.
      We are SO fortunate now but it comes at a price of greater perfection

    • “How did one make an old-school slide? I am solidly from the Powerpoint-era”

      As a grad student I was lucky enough to have Derek Whitely as the departmental graphics technician in the biology department, so it was a matter of printing out everything that needed turning into slides and handing it over to him to photograph onto slide film. Derek was a great artist who illustrated some volumes in the “Oxford Book of….” series, for example:

      Derek completed a PhD part time on snail colour polymorphisms. Sadly he died a number of years ago, I think from a major stroke; his wife found him at home in his studio, pen in hand, working on an illustration.

  7. When I started giving talks (late 70s), if we wanted a data slide we had to make a hand drawn version of a graph. This was given to a “medical artist” who made a neat copy in indian ink on white paper. The slide was then made by photographing this picture. The slides were called diazos as they came out white (lines and text) on blue background.
    This took time but it didn’t stop people putting talks together at the last minute because they accumulated a large number of slides and just picked out what they thought would work. My postdoc adviser would put together his talks on transatlantic flights.

  8. I still have my carousel from my defense (Jan 2001). It was an art form to create a great slide that was correct the first time without having to be re-shot. The process was similar to now, but you had to have things done way in advance. I can remember the pre-photoshop era when we added stickers to x-ray films to label the lanes and then took a physical photograph of the film (not that long ago; early 90’s). The absolute worst though was finding and photocopying papers from the library. I am so spoiled by pubmed and pdfs now.

  9. We made figures and printed them out – one per page. text slides printed one per page. images were taken in the field (or at the bench) with your (non-digital) camera or hand drawn. you would anchor your camera to the departmental camera stand, place a printed page on the stand, add the nonglare glass on top, take a picture. develop roll (some did this in the dept. darkroom, others took it to the local photo shop with same day service). load carousel. see how blurry, how dark, how overexposed, how small, how uncentered, etc. reprint pages. buy more film. reshoot. develop roll. repeat until you either ran out of time, or had adequate slides.

Leave a Reply