How I prepare a talk for a tech conference (2022)
chelseatroy.com
chelseatroy.com
It's always amazing to me when I see a talk by someone who has clearly not practiced it very much. I think it's usually better when you've run through it so many times so that you can deliver it without any notes, but it's still concise and complete. I practice double-digit amounts of times for every talk.
My tip is to record yourself practicing. Don't watch the recordings, no need for that. The thing is that blinking red circle seems to psychologically qualify as an audience (to me anyway) and it focuses me in on the performance part of giving a talk. One effect is that if I mess something up, and I'm recording it, I start improvising my way back on course rather than just starting over. There seems to be more consequences if you record it. I dunno, helps me.
One or two well-placed jokes can be great if they’re a very minor part of the presentation.
When someone has their slides stacked with memes or spends large amounts of time on vacuous entertainment content, it always feels disappointing.
There was a period of time where our biggest local JavaScript conference felt like one big entertainment competition. Presenters were singing songs, playing guitar, showing several minutes of clips from TV shows, and telling jokes more than they were presenting anything useful. The conference was a hit for young people and juniors, but it became known as a big waste of time for everyone else.
That said, I'm not really a "keynote" speaker. I've been doing it for most of my adult life, in front of small and large (but never huge) crowds.
I've definitely had some bombs, but most folks seem to think I do an adequate job (the fact I'm not a rubber chicken speaker means that it isn't much more than "adequate." I'm fine with that).
Most times, I wing it, but when I'm giving an important talk, like a tech class, or topic discussion, I practice. The tip someone gave about recording themselves is good. It's also how I know how long it will take. I'll often use a prompter app, while practicing, but don't use it, while giving the talk.
I've found a bit of humor (especially self-deprecating humor) can "humanize" things, but it should be short vignettes, and know your audience. If I don't really know who will be out there, I generally don't use humor. Most times, I know my audience quite well, and humor is appreciated.
I have found that speaking in the vernacular is generally appreciated by folks. There are definitely some folks that don't like making difficult stuff easy to understand, but I'm just an ol' high school dropout with a G. E. D., so I don't really pull off the "talk purdy" kind of thing, so good.
I've found that Keynote/PowerPoint shows can be useful, but I seldom use "wall o' text" slides. I use a lot of images and animations, and pack my notes (not the slides) with what I'm saying. I will usually insert
-- CLICK --
in the notes to denote when I advance the animation/slide.I'll put details and technology rabbitholes into the supergraphics.
People have widely varying views about what a “conference” is, or should be:
It's very nice to have a few people come up after a talk with "I saw what you did there...". It's a great way to connect with people who are on your wavelength.
The best talk I ever gave was when the scheduled speaker didn't show, and the organizer asked if anyone would volunteer. I volunteered, and gave an impromptu talk with no prep and no slides, just a whiteboard. I simply threw out questions to the audience, and let their responses guide things.
It was the best presentation I ever gave.
Too bad the video didn't work, either, and there's no record of it.
P.S. When traveling to a conference, I carry along the slides from my previous presentations. This enables me to fill in for missing presenters. I also had a talk that didn't go over well, so I just picked out a previous one and did that.
what works for me and is more time efficient is to get the introduction polished, then have a clear take away at the end. The middle stuff is better to go impromptu.
In general, the best speakers are those who a) know the material inside and out, and b) have practiced shit loads.
I remember the first time I went in front of a live audience to give a 2 minute intro. It was a meetup with ~60 people and all I had to do was go up and talk about a course I was selling.
On the train ride in, I scripted it out on paper and then re-read it like 30 times.
By the time I got off the train and closed the notebook I forgot everything and dreaded the walk to the venue. I got in front of everyone and tried to recall what I wrote, had a quick internal dialog with myself for about 3 seconds and mentally noted "you're a moron, just wing it". Then I winged it.
In the end it converted ~15% of the room on a tech topic that was ancillary to the main meetup's subject.
On that day I learned I can do well writing something out on paper but I can't remember shit when trying to deliver it live. I have to fall into category (A) and trust myself to deliver but really committing to that with no backup plan does make every experience interesting to say the least.
I've recorded around 300 videos since then and I still need to do the intro about 20 times before I get into the flow to do the rest of the video in basically 1 take. They are usually unscripted with no preparation, I just pick a topic and go.
Fascinating. What I learned from your anecdote is that massed practice didn't work, just as the literature on human memory predicts. Spaced repetition still sounds like a winner. Practice, sleep, forget almost everything, then repeat until it's go time.
I don't know why people don't seem to understand that.
If it's a topic that is low stakes and I know a shit load about, I'll practice for a day or two. If it's a topic that I know and it's high stakes, I may practice daily for two or three weeks.
Everyone who knows me professionally remarks about how relaxed I seem when public speaking.
That's because it's all intentional. Every pause. Every single thing. I seem relaxed, because I am relaxed.
For one off talks, or decks that one doesn't use often, yes you're 100% right. When I'm in that situation I rehearse the talk over and over and over again to the point where I no longer need notes. My goal is to get to a point where I neither sound like I'm reading from a script, nor would I get thrown off if I get interrupted.
However for material that I've presented a bunch of times, I no longer need to rehearse. There was a period of my life where I could on the fly give a talk on a particular topic at a moment's notice. Ideally I'd have a bit of time to customize some bits to the specific audience/venue, and sure I'd rehearse those. But the bulk of it I could stitch together content in just about any order on the fly.
All of those previous presentations basically count as rehearsals. Depends on you, but if it's been a year or several since you gave one, it's probably a good idea to do at least a mini rehearsal before you do it again, but if you're doing one a quarter, once you're good, you should be good.
But if you tell me give some variation on your talk about $X, I can probably be ready in about 30 minutes. (And, in fact, I've dropped in a conference presentation in response to a 3am email because another speaker forgot they were supposed to be there.)
Intros and ends though I agree require careful thought to appeal to the right audience and give the right message.
I don't think this follows. There are many things I am an expert in that I'm not an expert in communicating about impromptu- I've been to many talks where the speaker is clearly an expert but was not prepared to speak.
I certainly think there are people for who expertise == ability to speak about it, but that's certainly not always or usually the case.
I remember the first meme like presentations back in 2002 at work and I hated it. It's like stop learning, now you have to find this picture funny (maybe it was some star wars reference, and I haven't watched them), and we will get back to it. It was cringe!
Made me chuckle anyway...
I think it can work in limited contexts. I've given an internal talk about SRE using the "this is fine" meme, representing a current emotional state, then contrasted that with the reverse meme[1] as a visual metaphor. Perhaps this_is_fine.png works because it does not rely on outside cultural context.
I've also used HTTP status cat memes on a carousel slide, to emphasize that SREs commonly understand them, before contrasting that with a custom non-HTTP protocol I support. If people laugh or pay attention, so much the better.
> I think it's usually better when you've run through it so many times so that you can deliver it without any notes, but it's still concise and complete. I practice double-digit amounts of times for every talk.
Agreed that rehearsal really helps and more people should do that. I wish I could put that 10+ rounds of practice into presenting but I'm typically doing internal stuff not anything with an honorarium or ticket price attached. Yet it seems like some kind of badge of honor in the tech speaker circuit to start building your slide deck the night before. Trust me: we can tell, especially when your first five minutes are panicked searches for projector compatible adaptors you forgot to pack or test.
[1]: https://www.reddit.com/r/ReverseMemes/comments/dnlizu/the_th...
I certainly think there is a balance.
If you're not confident about your talk or your ability to present it, folks should be disabused of the idea that adding some gifs or a "One does not simply" meme will fix that.
That said, I think of my talks as sharing interesting stories. When I tell stories, I often crack a few jokes along the way, and my presentations are no different. I feel like adding humour is often a way of giving people insight into your thought process.
Of course, everyone is different, and perhaps you are dead serious about your way of thinking about compilers or neural networks or whatever your work on.
The author does an excellent job, however, of underscoring the fact that jokes can rely on an in-crowd or shared experience that some in the audience don't share. It's important in all aspects of your talk to keep that sort of thing in mind, and make sure your humour isn't just some twitter in-joke.
Personally I'm not much of a reherser. I like to make a point with each slide, and as long as I know the point, I can "freestyle" the actual text.
The exception are high-speed sessions, where I have a slide, or two, per sentence. They have to be highly scripted and so I practise them a few times. But, while they're a lot of fun, I don't do them often.
For me, prep is much more about what to leave out. It's easier to do a 2 hour session than a 1 hour, and harder still to do 30 mins. Slides mostly help to keep me on track and not get distracted.
You might realize that you are either speaking too fast or too slowly. You might notice that you are using your squeaky voice, there are ways to improve it. You might realize that you assume some piece of tech is known to everyone even though it's pretty obscure. You might realize you need one more slide for transitioning to a new topic better.
I agree though with your other point, the recording can help enforce that there are no redos before a live audience.
Also, if you mess up a sentence or slide all the time, maybe your phrasing is too backwards, so you should probably simplify it to something that feels natural to you.
It probably matters if it is related to the talk or not. A slightly related joke or funny picture at the beginning that introduces where you are going and then comes back with a deeper meaning at the end is very powerful.
All these theories, rules, and advice about giving presentations are just useless fads that are often impractical and unnatural for most people.
People often confuse an eloquent and natural speaker with having a "good" presentation. I have been to many talks (Tufte, for example) where the presenter talks smoothly for an hour, yet in the end does not say anything useful to me. Your are entertained, in a sense, but you don't actually learn anything.
To many presenters, the presentation is actually about themselves rather then giving something truly useful to the audience. A lot of people can talk well, but I rather have somebody that can teach me something, even if the delivery is not perfect.
Why would you throw Kevlin Henney under the bus?
But in general he had a lot of one off examples and interesting pictures, but without a solid and applicable theory behind them. He was an articulate speaker but I did not feel like I learned anything from the talk. I have also "read" his books but the only thing that I remember and make use of is the idea of information density on charts.
Another advantage is that the speaker doesn't always get the question. Either can't hear it well, or the question itself is weird, complex, off-topic, or some variation on not-even-wrong. So they end up answering something else. If they repeat/summarize the question first, at least they're answering a question, and when this happens, the question which gets answered is often better than the one which was asked.
It happened so many times that I had no clue what someone asked, and the presenter repeats the question with their own words, usually much more clearly communicating the question that was asked.
Just go find the speaker later and ask questions if you have them.
But there’s an art to Q&A that’s as hard as public speaking but practiced by 1/10th the people. A simple hand vote is the most basic tool. After a Q, I’ll summarize it and ask how many people face this problem? Etc. why get an audience of people in a room if you don’t use them?
I particularly like the bit about examples. I think that stories and metaphors are much more powerful than plain statements. It really drives the point home, especially when it's somewhat abstract. My all-time favourite is Steve Job's "bicycle of the mind".[0]
And practice. Please practice. That's what makes the timing impactful & the jokes work.
The one thing I'd add (though this is more in the hands of the conference organizers) is that shorter (25-30 minute) presentations are often better than 45-50 minute ones. Some of this relates to what the post says at the beginning. If it's a bad presentation, I've wasted less time. But the other thing is that a presentation can cue me in to something being interesting more than it can teach me everything about that something.
I've been seeing this as a general trend although I have colleagues who hate that trend and respond by trying to cram 60 minutes of content into 25 minutes grumbling all the way.
And, oh, stay on schedule. It's really rude to the next presenter (and your audience) if you don't. I might end up going a minute or 2 over if I get off-track but I feel badly when I do.
It’s shocking how many conferences don’t do that. I’m still amazed when I go to other events and see things get 15 or 20 or even 30 minutes behind the published schedule. It’s just so disrespectful to the audience to do that in my opinion.
I'm autistic, I'd happily cut off your microphone on time, not matter if you're the pope or a Nobel laureate. This is why I'm not allowed to chair sessions.
Generally speaking I'm not a huge fan of public raising of hands type Q&A's. They turn into more of a comment than a question and showing off sorts of things as you say. Maybe a few pre-submitted questions electronically or just save them for later 1-on-1s.
But rule number 1 for any talk is "know your audience." Since the audience is tech people who care about tech issues, talking about tech is a better way to establish your credentials than listing your CV.
But then again, how the people in the audience know you aren't just talking smartly?
My favorite tech talks are by people who jump straight into the tech issues without introduction. But I'm fully aware there's a chance they could be straight up bullshitting.
Hi, I'm X, I work on Y at Z and maybe I worked on Y' at Z', welcome to my presentation on Topic establishes some context. And lets you know if you're in the wrong room. I don't think you need much more than that at the beginning. Maybe in the depths of the presentation you talk about options for a solution and how you tried some of the options on a project/at a company, when it's relevant.
We don't usually need to know a lot of other stuff from a CV, such as how long you were anywhere or where you went to school or a lot of details about job responsibilities. Just a little info to have some idea of the scope of 'databases' and 'embedded' for you --- there's a lot of computer words that have similar but different meanings depending on your industry.
The part that got me was right at the start:
"Why: I get very little out of 95% of them. I much prefer online recordings, which can come vetted/recommended. Also I can bump them up to 1.5x or 2x speed, so that at least if the talk isn’t that good I waste less time on it."
I also just don't enjoy these conferences at all. I rarely attend and have zero interest in ever speaking. Watching recorded presentations is far superior.
I've found that this is domain specific. You see it more for tech conferences for instance, but less so in biotech conferences
He gives a fantastic meta-talk; you should listen to it.
One of my friends casually mentioned that the A/V team had our podium set up. Podium? For what? So that you’ll have some place to set your laptop when you’re presenting the project to the large room we booked!
That’s how and when I found out I was giving a tech talk. It went well. Honestly, that’s about as much advance notice as I like. You can’t worry or procrastinate too much when you’ve got about 20 minutes to get ready.
Yet, there is an element of YMMV. Like, I only ever do live demos, including lots of live coding. My "slides" are my org-mode file, with org-babel, and org-tree-slide. My premise is that, in-conference, peoples' eyes will glaze over sooner or later because of information overload from other talks and/or intense social engagement. So I need to land one maybe two engaging moments. Interested parties will engage in the "hallway track", post-talk. At-home viewers will scrub through the recording. All people have access to slides and/or a detailed blog post. Disinterested parties are free to use "the law of two feet" and leave if it's not working for them.
My talks have never been polished diamonds --- all elements of serendipity and demofail are embraced. So far, nobody has booed me off-stage, and I've always had fun after-talk conversations.
Also, in terms of actually doing the thing, this is what my, ah, "process" ends up looking like [1]. I discovered many people relate to it.
[1] https://www.evalapply.org/posts/how-to-give-a-conference-tal...
The other trick is to make material available to everyone, post-haste. This is especially important for remote talks [1].
Apropos embracing failure. I work hard to set up a smooth path. I don't want things to fail. Yet, I actively chose to be open to it because undoing the failure has a habit of creating a learning moment for someone among the dear listeners (they too try to debug in their head, and arrive at their own insight).
[1] In 2022, I gave a talk as a live demo at a remote conference. Three different networks at three different locations in my area flaked on me. My home network because of digging in my area, another mid-way through my talk because of power failures, and an otherwise-pretty-good wireless network because it lost packets exactly over my presentation and was just fine ten minutes after.
This was after I prayed to the demo gods at the start. See slide #5: https://github.com/adityaathalye/slideware/blob/master/n-way...
And the talk (the zone was deadpan, laconic): https://www.youtube.com/watch?v=BTouODWov-A
And the accompanying blog post: https://www.evalapply.org/posts/n-ways-to-fizzbuzz-in-clojur...
But those slides; the walls of text. If I'm reading what's on the slide, I'm not listening to what you're saying. Worse, I'm trying to do both, so neither is leaving an imprint in my brain.
Text-heavy slides are great if I'm emailing a deck to someone, but for presenting, 5-6 words max per slide.
Remember - Stephen Hawking presented his findings. It wasn't his awesome presentation skills.
"What new skills do I want my audience to have, and know how to use, when they leave this room?"
That's effectively the same thing as the way I look at it. "What actionable thing do I want people to take away from this talk?"
I will add: Practice! You should have given your talk multiple times before you ever give it publicly, and you should have watched it. That means recording yourself and watching yourself and listening to yourself. Yes, this takes time. But it's important.
Also: do NOT plan to do live demos or live coding. You can easily record these and present them and know they work. They are edited, there are no mistakes, and things happen. You can even make the joke that it was "live coding when you recorded it."
Yes, if you have to answer a question with some live coding, fine. But it doesn't need to be live. No one wants to watch you typo stuff, complain about the network, etc. Just record it. If "recording it takes too much time" then doing it live is going to suck.
I would rather not give the presentation than do this. It is a cringe and self-embarassing experience. Once, I almost quit my job because certain types of high-level presentations at my company requires dry-runs and I refuse to do it. And I regularly give presentations at a few conferences a year with sufficiently good delivery.
I'm the same way, and honestly I don't understand why.
I'm not sure it's embarrassment per se, because nobody else is in the room.
But something about doing a dry run, in private, is so unpleasant that I can't make myself do it. Despite the apparent benefits.
Fad of the day at my company is evidence-assertion style. And I am sick of it...
As someone who has presented a LOT, practice is useful. I always do at least a bunch of mental run-throughs. That said, a real run-through, which I don't always do, results in a better Take 2. So, yeah, do a real rehearsal when you can.
(I won't practice dozens of time though. At peak I presented a lot and practicing for a given presentation that much would have been impracticable.)
On another note, there's now a huge pressure on people to do presentations - many requirements for senior engineer include things like presenting at meet ups and public speaking.
I find this very unfair on people. If you don't enjoy doing something then you really shouldn't feel obligated to do it. Talking at a meet up or a conference is very different from running a meeting or doing a presentation at work.
I asked him how he managed to do it so well and he said "I don't do talks anymore.". Despite being very good at it he decided that putting himself through the wringer was just not worth it.
I disagree, legitimacy by proxy is a very useful initial signal. Someone who works on the gRPC team at a large API distributor is more likely to have better advice on best practices for protobufs, than someone who works on optimizing ray-tracers for TempleOS. Relying on proxies, and likelihoods, is crucial for the timely processing of information. We heavily rely on at least some level of trust - otherwise we would have to completely verify every single thing that anyone says, before we can begin to do anything with the information.
Association alone isn't very convincing. If people really feel they need to establish credentials they should talk about what they do. It is almost like at a job interview. I don't need to know that you worked for X - I want to know what you did at X.
Introductions should be no more than what you can fit into one breath. After that it is just some person going on about themselves, and that isn't very attractive. Get to the meat first, and then we can talk.
Slide with title "about me"
Next slide: "who cares"
Then I say "you came here to learn about <topic>, not me. Feel free to Google me if you want".
* someone has read or at least scanned my speaker bio
* folks want to get straight to the content
* my expertise is shown by the fact I'm speaking and the content I'm conveying
But maybe I'm wrong.
Probably not.
As I wrote in another comment, a quick context-set and contact information in probably useful but quick is the operative phrase especially for any background that isn't directly relevant to the material at hand.
I'll do it after, if your talk catches my attention.
This is why I now love attending talks remotely. If you annoy me, I can mute you and focus on something else and unmute the stream when the next speaker comes up.
"you came here to learn about <topic>, not me. Feel free to Google me if you want. Just please keep in mind that all teenagers make mistakes. It's not my fault you only made boring mistakes."
To make it work, you need to deliver the last line while seemingly preoccupied with something presentation-related. Otherwise, it seems hostile. An example could be as simple as pretending to change settings on microphone or moving one slide back, then forward twice. It conveys it was intended to be inner monologue
If you do everything that the author recommends, but you don't have anything valuable to share, your talk will flop. So make sure you have something important to tell people.
What I'd really like to read is an account of how to make interesting content, how to arrive at great conclusions while working on great problems.
It's basically coming out of the gate saying that you don't respect the time/attention of the audience.
Are there other plausible explanations?
E.g., perhaps the presenter struggles with ADHD-based procrastination, and is only able to make slides under imminent deadline pressure.
Or maybe they tend to be neurotic and over-think things, and are experimenting with being more spontaneous in their preparation.
(Asking as a neurotic ADHD person with moderate social anxiety.)
Technical presentations have a different set of "rules".