How to Give an Exceptionally Good Presentation

exceptionally-good-presentations.txt

"HOW TO GIVE AN EXCEPTIONALLY GOOD PRESENTATION"

CHARACTERS

Joe Cheng.............Moderator
Jennifer Bryan........As Herself
Hadley Wickham........As Himself
Aron Atkins...........Audience
Derrick Kearney.......Audience
Kirill Muller.........Audience

INT. INDUSTRY ROOM I, Hotel Republic, SAN DIEGO

ENTER JOE, HADLEY, JENNIFER

[ SCATTERED CONVERSATION DIES DOWN ]

JOE: We are exceptionally lucky to have Hadley and Jennifer
    with us today. I'll ask some questions to start with and
    then we'll open it up. To begin with -- What is the purpose
    of a 20 minute talk?

JENNIFER: You're trying to convince someone that they are capable
    of something. Not how to do something -- that it is
    possible, and possible for them.

HADLEY: It's a sales pitch. In 20 min you can't go deep
    technically; "I should learn more about this." Make sure
    there's pointers to learn more. 

JENNIFER: That talk is a confidence-building ad.

JOE: In some circles the point is to show that you're smart. 

JENNIFER: That's why we're not in those circles any more.

[ LAUGHTER ]

JOE: What are the most common mistakes?

JENNIFER: Front-loading the boring. "To be complete I have to show
    you these things..." and starting with those things. For
    example, when people start off with R by teaching a list
    of every data type. Save that for later, if at all.

HADLEY: You want to persuade people in the first minute that they
    should care about the rest of the talk. You need compelling
    motivation for later on.

SOMEONE: How about a table of contents slide?

HADLEY: I hate that.

JENNIFER: I think it's terrible.

HADLEY: But for longer talks, it's good to give a sense of
    structure, and sum up what you've done at the end. 

JENNIFER: The final slide should be one people can take a picture
    of. Tell them what you're going to teach them, then tell
    them what they've learned. Don't try to have a surprise
    in your talk. I've never had a satisfying reveal.

JOE: I have a funny story that's relevant here. The first
    time RStudio has ever demo'ed in public, I showed RStudio
    Server in Chrome the whole time, but full screen so you
    didn't see the browser. Then I pressed F11 at the end to
    show that I'd been running the whole thing in a browser.
    That was supposed to be a powerful moment, but the audience
    didn't get it right away.

JOE: I think for a lot of us who have not given a lot of
    talks, it's nerve-wracking, but it seems like it's so easy
    for you. What's the worst you've ever bombed?

HADLEY: I showed an XKCD comic and no one laughed. I realized the
    presentation was wrong for the audience, and there was
    nothing I could do. There's nothing to do but show the rest
    of the presentation. For small things you can rescue
    mid-flight, but when it's so far off you really can't. 

JENNIFER: I've had lots of versions of this, where the audience
    isn't into the topic. The most confidence-eroding was giving
    a talk that was very applied, and someone wanted to talk
    about the asymptotic approach of something. I ceded too much
    control and it made me feel less confident.

HADLEY: Dealing with questions is hard.

JENNIFER: I was not prepared to say nicely that your question is
    not relevant.

HADLEY: It's good to have canned responses like "I wish we could
    spend time on this."
    
[ LAUGHTER ]

JOE: You each have a different speaking voice. Is it intentional?
    How do you try to speak?

JENNIFER: I try to be more conversational. I notice on podcasts
    when people use fillers and back off on things. It's better
    to pause when I use filler words, and to skip softeners and
    backpedaling. Just own it.

HADLEY: A couple of years ago I did five sessions with a voice
    coach, just to learn how to project better. If you do a lot
    of public speaking, that is really useful. Part of the
    reason was because of Joe: there was a room with no mic and
    it was loud, and Joe said it sounded like I was shouting. I
    needed to learn skills to project without exhausting my
    voice.

JOE: The filler words issue. Can you explain that a little
    more? It seems really hard to get past.

JENNIFER: Being more patient; being willing to tolerate silence.
   Silence is better than weasel words and ums. They undermine
   your authority and soften everything. 

ARON: Do you do dry runs?

JENNIFER: I do little bits in my head; I go over it when I drive.

HADLEY: When you are silent as a speaker for a few seconds, it
   feels like an eternity. But your experience of your
   performance is very different than the audience's experience
   of that performance.

JOE: I felt like those pauses made me look stupid. But
   they're actually what make you sound authoritative. Own
   those spaces, and speak into them.
   
HADLEY: It's odd when you first start practicing that. You
   recognize when you've said "um", and you can't unsay it, but
   you need to be aware of it.
   
DEREK: The pauses are natural; it makes a conference talk
    sound more conversational. 

KIRILL: How do you allocate time? How do you make sure
    you're not spending 20 hours on a slide deck?

HADLEY: We never create a new presentation. 

JENNIFER: Talk more often! Manufacture opportunities to talk more
    often, so that you're not starting from scratch. 

HADLEY: There may be pathologies where you spend hours tweaking
   slides. Maybe you need to start by scratching things out. My
   pathology is buying new fonts.

HADLEY: I will rough out the whole thing, and then let it sit
   overnight. If you can, prepare a week before the talk, and
   then don't look at it until the day before the talk.

ARON: I started with a story first and an outline. The
    slides were the absolute last thing that I did. 

JOE: Can you talk about the relationship between the concept
    and the slides? 

HADLEY: One of the goals of this is being brutally honest about
    what this is.

JENNIFER: Another is that if a few slides from this get circulated,
    what should they be? Put those slides in first and then
    connect the dots. It's easy to connect the micro-stories and
    then never get to the conclusion you want to reach.

HADLEY: You don't want your slide deck to be self-contained. Why
    would they waste time watching you deliver it when they
    could read it?
    
[ INAUDIBLE QUESTION, PRESUMABLY RELATED TO NOTES ]

JENNIFER: I use the presenter notes in keynote, so that when I look
    at the slides before my talk I have them. When you're
    switching between RStudio and demos, you can't see your
    notes. 

HADLEY: There's a certain level of preparation. If you end up
    preparing so much, and you forget one, the audience is not
    going to know that you missed one -- they don't know what
    you're going to cover! It distracts you more. I don't use
    speaker notes. Sometimes I don't know what the next slide is
    going to be. 
    
JOE: But you're looking at it just to remind yourself what
    the point is. Some speakers are basically just reading the
    slides when they come up. That means you're trying to teach
    too many concepts. Your first sentence after the slide comes
    up should come out with conviction. 

HADLEY: One of the most useful slide layouts, which is what I
    have a template, is assertion plus evidence. The goal of the
    slide is to supply evidence for the assertion. This makes it
    less likely to create bullet points. 

JOE: All this advice is awesome, but some people don't need
    it. Whatever the hell Yihui does is working. And as I was
    matching Mara, I was thinking "Only Mara could give this
    talk." Mara, you did the opposite of all these things and it
    worked great. You don't have to be Hadley or Jenny. You can
    be Mara or Yihui.

JOE: What other problems comes up?

HADLEY: Demos don't work. Hecklers. 

JENNIFER: I wouldn't plan a demo I wasn't completely sure would
    work. Do not call the Internet! If there is a video I plan a
    static version of the code in case it doesn't work.

HADLEY: Never assume that the wifi is going to be any good,
    despite any assurances to the contrary. Code demos where you
    type the code, even if you make mistakes, it's really good.
    It shows that developers make mistakes too.

HADLEY: One piece of advice for asking questions is that you
    should count in your head up to 7. That's how long it takes
    form a question in your head. 

JOE: So when you've given a talk, how do you judge success?

HADLEY: You can see it in the faces of people.

JENNIFER: If you get any reaction, any questions that seem good.
    You find out later if it was succcessful or not. 

HADLEY: If you're teaching, you can tell if people are following
    along or if their eyes are glazing over. There are things
    that happen that make you think your talk is terrible, like
    people falling asleep. That doesn't mean that things are
    going terribly, it means they are sleepy. Accept these
    things. 

JOE: Can we say thanks to these two?

[ EXTENDED APPLAUSE ]

JENNIFER: I have two more things to add. Take your font and
    make it bigger. And then make it bigger again.
    Anticipate that it needs to be huge. Put the link at the
    front, and put it in like 150pt font size.  People often
    want to take a picture.  Leave a slide up for questions. 

HADLEY: Don't have a slide that says "Questions". Have a slide
    with useful information.

JENNIFER: The value of putting the slides up is greatly enhanced if
    it's put up from the start and at slide 0. If you're going
    to bother to put them up, there's a much bigger payoff that way.

[ SCATTERED CONVERSATION ]

EXIT JOE, JENNIFER, HADLEY