Tutorial 16: giving good talks, part 4: delivery

December 17, 2011

And so we come, rather belatedly, to the fourth and final part of this series on preparing and giving talks at scientific conferences.  If you’ve followed the previous installments, you should have figured out a clear, compelling story that you want to tell from your research; you should have clear slides with striking, relevant images and no visual distractions; and you should have rehearsed your talk so that it’s clear and coherent, and fits into the time available with a minute or two spare.  Now we come to the business of actually delivering the talk.

Matt’s talk in Bonn

If you were paying attention, you’ll have noticed that I used the word “clear” three times in the previous paragraph — once for each of the three previous stages.  That’s not by accident: the whole business of giving talks is about clarity, and that’s not going to change as we talk about delivery.  Remember, once more, that your talk is one of maybe fifteen or twenty that your audience will hear that day: by the time they get to you, their brains will be half-frazzled.  They will thank you (internally if not out loud) for leading them by the hand through your argument.  You don’t want listening to your talk to be hard work.

Remember, too, that you are the expert on your subject.  Even if you’re a student giving your first ever conference talk, and even if the front row is full of Mike Bentons, Eric Buffetauts and David Normans, the magic of specialisation means that you know more than they do about your subject.  They want to know what you have to say.

And that leads us to Rule One, Rule Two, and Rule Three, which are so important that together they make up 90% of what you need to do to deliver your talk well.  Here they are.

The big three rules

  • Rule 1.  Speak up.
  • Rule 2.  Slow down.
  • Rule 3.  Don’t Panic.

So let’s look at these in order.

First: speak up.  It’s so simple that it seems almost insulting to mention it, but if people can’t hear what you’re saying, you might just as well not be saying it.  Not everyone has a naturally loud, resonant voice, but that’s OK: conference venues will provide a microphone.  Stay reasonably close to it — a foot or so will be fine.  Many microphones are extremely directional, so make sure it’s pointing directly at your mouth, not past your ear or at your belly-button.  If your natural tendency is to get quieter and quieter as you speak, have a friend sit at the back and wave at you if your volume falls too low.

And the second is like unto it: slow down.  Again, if people can’t hear what you’re saying, you might just as well not be saying it.  I know that I tend to go rather fast myself, but then I rarely finish before my time is up and I don’t think anyone has any problem making out my words, so it’s not a disaster.  Where it can hurt much more is if you’re nervous, and keep speeding up incrementally.  (Vanessa suffered from this tendency when she gave her talk at the sauropod meeting.  Fortunately we knew this from rehearsal, so I was ready with a big SLOW DOWN sign to hold up as a reminder when necessary.  I used it twice.  To my surprise, Matt needed a dose of the SLOW DOWN sign in his talk, too.  Maybe I should get a really nice one made and take it to all the conferences.)

You probably won’t need Rule 3; but if you get into a situation where it’s applicable, you’d better remember it.  First-timers are most likely to feel panic; and the only advice to give is, don’t panic.  Take your time.  Everyone’s on your side.  (It’s always interesting to see how an audience’s tenor changes from talk to talk: for some speakers there’s a sort of hunger to hear what’s coming next, and for some there’s a subtle challenging feel, as though everyone is silently saying “Oh, yeah? Show me.”  But without exception, every time I’ve heard someone giving their first talk, the audience has been polite, respectful and encouraging.  They want you to do well; so don’t panic.)

The best example I’ve seen of not panicking was at this year’s SVPCA in Lyme Regis.  A student, giving her first talk, lost her thread completely and just stopped.  It was a horrible position to be in and my heart — everyone’s, probably — went out to her.  But she just stood still and silent for as long as it took (felt like twenty seconds but was probably more like five), thought hard about the line of her argument, then when she’d found the thread, carried on like nothing had happened.  It was actually really impressive.  If I remember right, she didn’t even apologise, and that is definitely the right way to go.  The idea was just “I was gone for a moment there, but now I’m back; so ANYWAY …”  Good stuff.

Mike’s talk in Bonn

Different styles

It’s worth pointing out there are very different styles for giving a talk, and cliched as it sounds it really is true that none of them is better than another.  To pick examples from among the SV-POW!sketeers …

When Matt prepares and gives a talk, he does it not by learning the talk itself at all, but by learning the source material, and by knowing his own work inside out.  The slides then function (for him) as a visual cue, reminding him what to talk about next — even as they are also functioning (for the audience) as illustrations that support what he is saying.  The specific words that he uses are made up as he goes along.  (If real-time improvisation sounds hard, just think that this is exactly what you do every time you have a conversation.)  I’ve heard Matt rehearse a talk the night before, then give it on the day, and hardly a word is the same between the two performances; but the substance is the same.

By contrast, I have the impression that Darren prepares every word, and that if he gives that same talk twice it will be the same both times.  That’s not to say that he won’t take the opportunity to throw in new comments that occur to him at the time; but basically I think he writes and memorises a script, which he then plays out when the time comes.

Both approaches can work really well.  (They both did in the sauropod conference, where both Matt’s and Darren’s talks were outstanding.) Doing it Matt’s way can have the advantage of feeling more informal, and it gives you more opportunity to engage the audience non-verbally — eye-contact, responding to laughs, and so on.  On the other hand, Darren’s approach lets you use the time more efficiently — you know you’re not going to waste time repeating yourself or going off on a tangent.  It also means your running-time is more predictable, so you can keep cramming in more material till you get right up to the end of the slot.

Personally, I use the first approach — I like the freedom of approaching each slide’s discussion in a way the reflects how people have been responding.  Many students start out doing it the other way — probably because, in an intimidating environment, it feels better to be fully in control of content.  Let me say again that neither approach is intrinsically superior — you should feel free to do it whichever way feels most natural to you, and that may or may not change over time.

The one thing you don’t want to do is write a script, then not memorise it, so that you have to read it out from a printed copy. That’s a recipe for not engaging.  All the time your eyes are on the printout, you’re not looking either at the audience or at the same images they’re seeing.  How can anyone in the audience join you on your journey if you’re off on your own somewhere in a stack of index cards?

Hints and tips

Face the audience.  It sounds obvious, but I’ve seen it not done.  At the Bonn conference, the talks were given from a laptop sitting on a surface to the side of the screen.  At one point in the conference, it got turned around so it faced more towards the screen than the audience, so one speaker gave her entire talk looking at the screen, with her back towards the audience.  Straight away, bam, you’ve lost people.  If they can’t see your face, you don’t come across as a human being.  They might just as well be paging through a copy of your .ppt.

More generally: engage.  Sense the mood of the audience.  Feel the atmosphere.  Make eye-contact with a friendly face.  If you crack a joke and people laugh, take a moment to enjoy it, don’t plough determinedly onwards through your list of points.  As a rule of thumb, treat it the way you would if you were giving a talk to a group of friends in a pub.  You don’t want to sound like a technical paper, you want to sound like a person.  (Example: when Matt ran his talk the night before his presentation, he said that in birds, the pneumatic diverticula run “subcutaneously and intermuscularly”.  But there was no reason for him to say that instead of “under the skin and between the muscles” — it’s harder to say and to understand, and carries no greater precision.  So he changed it for the actual presentation.)

Vanessa’s talk in Bonn

You don’t need to say “This work is from my Masters project”, or from your Ph.D, or whatever it may be.  Doesn’t matter.  Your work will be evaluated on its content, not on what stage in your academic career you’re at.  Some — no, most — of the best talks I’ve ever heard have been given by grad-students, or even on occasion undergrads.  Some of the least engaging were by seasoned, even respected, professionals. Basically, once you’re standing at the podium, the playing-field is level.  No-one cares about your status, they want to hear your ideas. In academia, nobody knows you’re a dog.

Leave the laser pointer alone.  If you really need to point at something, pick up the laser pointer, point to it, and then put it down again.  DO NOT WAVE IT AROUND.  Do not circle it repeatedly around the feature of interest.  These behaviours are distracting us from your talk.  Don’t you want us to concentrate on your talk?

Don’t read the slides out loud.  We can read.  If you say the words on the slide, we will be reading along with you; but because reading internally is faster than speech, we’ll be ahead of you, and frustrated that you’re not keeping up.  You don’t want that.  When you’re giving your talk on the subject that you know best out of everyone in the whole world, you don’t want to seem like the dumb one.

The worst possible thing you can do combines both of the last two DON’Ts, yet I’ve seen it done more than once.  The way you do it is, you slooowly read the words off the screen while tracking each word as you say it with the laser-pointer.  About two slides of that is enough to make the audience want to gnaw off their own heads so they won’t have to see any more (or, if they are clearer thinkers, gnaw off your own head).

Finally, a time-saver.  Many talks finish with a summary slide and an acknowledgements slide.  Both are very wordy and take a long time to read.  So don’t.  Skip the summary completely — we should be able to remember what you told us only fifteen minutes ago.  And don’t read the acknowledgements — just put them on the screen as you finish, and people can read them as you take the questions.

Handling questions

Not too much to say about this.  The questions at the end of talks fall into a few categories.

— Someone doesn’t understand something you said in your talk and wants clarification.  Just explain.  It may be worth going back to the slide in question.

— Someone wants to demonstrate that they know a lot about your subject by commenting on an esoteric point.  Sad to say, yes, this does happen.  Quite a lot.  Best just to let him (it’s nearly always a man) get it off his chest, acknowledge the point, and move on.

— Someone has useful information for you.  This is the best kind of “question”.  For example, at the end of my Bonn talk, Phil Manning happened to know the relative thickness of the cartilage between the cervical vertebrae of a hadrosaur mummy, and threw that in.  Very useful.

— Finally, someone might have spotted a real flaw in your work.  This is rare — after all, you’ve been working on this stuff for months or years, and the audience are only just hearing about it for the first time — but it does happen.  In this case, it’s probably best to say as little as possible, beyond acknowledging the issue, and save the deep thinking for later.

Darren’s talk in Bonn

To finish, here is (what I think is) an example of the latter.  I was at SVP in Austin in 2007, when Jack Horner gave his talk claiming that Dracorex and Stygimoloch are successive growth stages of Pachycephalosaurus. At the end I stuck up my hand and said something like the following (Darren remembers the actual words much better than I do):

We’re used to seeing animals develop more elaborate crests and other display structures through ontogeny, but in the sequence you’re proposing here, the cranial ornamentation becomes progressively less flamboyant.  Do you know of any extant animals that follow this ontogenetic trajectory?

Horner paused briefly, then replied as follows (and this is word for word):


There is much to be said for such economy.

Update: one more thing (from Matt)

I hate to destroy the elegance of Mike’s ending by tacking something on, especially since he let me do an editing pass and I had a chance to get this in before he published. But it’s important, and it’s not worth a post of its own, so here goes.

Please don’t say at the beginning of the talk, “I hope in this talk to convince you that…”, and likewise don’t end the talk with, “I hope by now you’re convinced that…” Blecch. This is just gross and lame, for several reasons.

Reason 1: You shouldn’t hope that your audience is convinced by you. Either your argument is sound and your evidence solid, or not. If so, then your job is to make those things clear, and once you’ve done so, to some extent it is out of your hands. You can’t force people to come to the same conclusions. And if your argument and evidence are not good enough, then you don’t want people to be swayed by rhetoric–that might make you feel good, but it would be a loss for science. And clever people in the audience would notice anyway. (Also, if your stuff isn’t that convincing, give a different talk!)

Reason 2: It’s not your job. You’re there to talk about your science, not how the audience should feel or think about your science. So don’t pull people out of the talk by talking about the talk (another reason never to apologize for the talk during the talk). Get out of that meta-level and tell your story, simply and directly. Or quit and go into marketing.

Reason 3: It sounds pathetic. You absolutely can affect your audience, but you do so by connecting with them, being responsive, and above all knowing your stuff and presenting it clearly. Telling the audience that you hope that they’re convinced is pleading, plain and simple. It makes you sound weak, and corrodes your credibility. Just don’t do it.

19 Responses to “Tutorial 16: giving good talks, part 4: delivery”

  1. Thank you so much for all of this Mike – I’ll certainly be reading this more than my slides/notes the first time I have to give a talk! Academics from all disciplines could benefit from your advice :)

  2. Mike Taylor Says:

    Thanks for your kind words, protohedgehog. (Cool username by the way.) Those of you know who people in other academic disciplines, please do send them the links: it’s true that about 97% of what I’ve written in this series is equally relevant outside of palaeo.

  3. Lately I’ve been noticing speakers who start each sentence at a reasonable volume but drop to a whisper at the end of each sentence. Don’t.

  4. Andy Farke Says:

    Excellent post – especially on the point of not reading acknowledgments! I would also add that one should give thought to how to end the talk gracefully, and never finish with “and I’d be happy to take any questions.” I see this done in about 25-50% of talks, and it’s awkward for everyone! Awkward for the audience, because they’re not sure if they should applaud as is conventional at the end of a talk or risk interrupting someone who is going to ask a question, awkward for the moderators when the presenter has gone 10 minutes over time, and awkward for the presenter when he or she has to be told, “actually, we don’t have time for questions.” Instead, make your concluding statement, and finish with “Thank you.” This lets the audience comfortably applaud if they wish, and leaves the control of the session in the hands of the moderators.

    A second pitfall frequently seen in talks (especially with students) is too much self-deprecation. . .don’t continually apologize for things, because it’s probably not as bad as you think, and too much apology is annoying to the audience (who is thinking, “if the slide is so unreadable, don’t apologize – fix it before the talk!”)

    The final pitfall is the opposite of the second, arrogance. Also common with students. There is a fine line between being confident and talking down to your audience. I’ve seen this more than a few times, and it’s always irritating (even if the Q&A session usually brings BS to light).

  5. Oh yeah, arrogance is awful during talks – I’ve noticed it a lot over here, typically from Oxbridge students. If you’re at a conference, people are going to know at least the basics of what is being talking about (like, you wouldn’t have someone at a Sauropod conference who didn’t know what one was..), so there’s no point in patronising the audience about fundamentals.

    As for finishing talks, I usually finish (as with most things) by saying “Shazzam”. Goes down pretty well..

  6. You don’t need to say “This work is from my Masters project”
    Indeed – except when you know there are a lot of holes in your work, and you present at a conference where that level is usually not presented (e.g., one-week school project at SVP). Then, it is often a good idea to work the “I had only a few month to look into this” thing into a sentence. It keeps the know-it-alls from pouncing on you hard in the Q&A.

  7. Jeroen Says:

    Thanks a lot for these tutorials.
    It’s a great help, especially for talk newbies like me. :p
    I will be giving a small talk on tuesday so this post came just in time. XD

  8. Darren Naish Says:

    The Horner question was at Bristol.

  9. kattato Garu Says:

    It was a good question. And what should be the answer – can sexually selected display traits regress (for lack of a better word) to a less elaborate, cumbersome or costly state? I thought there might be some examples amongst well understood island phylogenies e.g. anole lizards but frankly cannot think of or find a clear case study.

  10. 220mya Says:

    Heinrich – rather than saying “this is an MS project”, “this is an undergraduate project”, “I’ve only had a month to look into this”, or similar things, I think its just better to say “these are preliminary data”, or this is “a preliminary study that I plan on expanding”…

  11. 220mya – depends. I can see situations for both ways of doing this. “preliminary study” says you didn’t do a thorough study (yet). “BS project” additionally tells people that you’re very inexperienced, basically granting you puppy protection.

  12. Mike Taylor Says:

    But you don’t want to go in saying, “Oh, please be gentle with me, I am only an undergrad”. You don’t want the audience to patronise you. You want to be telling them “Here is my project, there are my data, this is what follows”, Bam, bam, bam. You want them to see you as an equal participant in the process of making science, not as some poor waif or stray in need of nurture.

  13. Mike Taylor Says:

    kattato Garu asked:

    It was a good question. And what should be the answer – can sexually selected display traits regress (for lack of a better word) to a less elaborate, cumbersome or costly state? I thought there might be some examples amongst well understood island phylogenies e.g. anole lizards but frankly cannot think of or find a clear case study.

    It was a good question and not one that I have seen answered in the affirmative yet. (That’s not to say that there aren’t examples; but if they’re out there, they are at least a tiny minority of all sexually-selected allometric ontogenetic trajectories.)

    I’ve not read all the Horner ontogeny papers. Does anyone know if he addresses this issue in any of them?

  14. Mike, that really depends. If you are knowingly stepping into a minefield it can really be a good idea to tell people you’re not at their level of expertise.

  15. Andy Farke Says:

    @Mike – Ontogenetic allometry of cranial ornaments in modern taxa hasn’t really been addressed by the MOR research group. That said, the general (exclusive?) trend tends to be positive allometry for horns, crests, or whatever, judging from the comparative literature.

    There is some precedent for regression of sorts in ceratopsids – pachyrhinosaurs start off with a “traditional” triangular nasal horn, which is remodeled into a boss. Some other centrosaurines round off their triangular brow horns late in ontogeny, too. One thing I wonder is whether this was really reflected in external appearances – i.e., was the overall shape maintained by keratin?

  16. Darren Naish Says:

    With reference to Andy’s comment… actually, isometry and even negative allometry is common for sexual display ornaments in insects and – shock horror – birds. However, it seems to be a fair generalisation that horns and cranial crests exhibit positive allometry during ontogeny. I seem to be saying the same thing a lot at the moment: the ‘ontogenetic morphing’ hypothesis (Torosaurus is Triceratops, Stygimoloch is Pachycephalosaurus etc.) is a HYPOTHESIS – is anyone outside the MOR research group actually confident that it might be right?

  17. Mike Taylor Says:

    “The Horner question was at Bristol.”

    No, at Austin.

  18. […] and you will — finally! — be ready to actually deliver the talk at the conference.  We’ll discuss that next time. Share this:FacebookRedditTwitterLike this:LikeOne blogger likes this. Posted by Mike Taylor […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: