I love giving talks and I love attending a good talk. I actually have given 21 invited talks this year alone, which is a new record for me. I’d like to think I’ve gotten relatively decent at these in the past couple of years. And while I am certainly not an expert, I do want to throw the gauntlet. Academics give far more talks relative to many other professions, and yet their talks are often really mediocre (or worse, boring).

While I’ll be focusing this write-up primarily on the context of “paper” talks given by academics, I do think that most of this is pretty broadly relevant too.

Most advice for the format of giving a good talk centers on stuff that is standard like clarity/honesty, relevance, strong argumentation, and clutter-free, well-desiged slides. These are fine and worth paying attention to. You can read other blogs if you want advice on those fundamentals. But I won’t focus on those things in this write-up. There are 4 things I think are missing from most advice on giving a good talk that I really wish more academics paid attention to.

And while some advice (like the fundamentals I just mentioned) can be good. Some advice is really not good.

In terms of content within a talk, the prevailing advice given to budding academics tends to be: “take the headings from your recent paper, slap those into some slides, and then fill in the blanks” (typically using figures from the paper too). That’s a recipe for a really boring talk. Snores-ville express. I could just read the paper for all that. And if I can just read the paper, why would I ever attend that talk? (Even worse: some academic communities just have the speaker literally just read their paper as their talk. Awful!)

And just like I’ve heard the common wisdom to use paper headings as the outline of a talk, nearly as often I’ve heard the advice given to inexperienced academics that the goal of a talk should be to “convince people to read the paper.” What a weak goal! That goal fails to ask why you would want someone to read your paper in the first place. Do you just want citations? Or do you want to inspire future work? Those are two entirely different goals that are often confused with one another.

I often don’t want people to read my papers. Sure, it’s great if they do. But I wrote papers because I want an archival document of some work that I did. If someone does manage to read my papers, it really just means that (hopefully) they understand what I did, why I did it, and how I did it. That is the timeless and important role of a paper.

But a talk can actually accomplish convincing people to do something. And talks can do this much more effectively than a paper can! So if I can convince people to do good in a talk, I often don’t even need them to read my paper at all. They can read it if they want, but what I really care about at the end of the day is that they did something that makes the world a better place.

And to revisit my point about what papers are for: A paper is an artifact, hopefully of archival quality, of a research contribution. The paper is a record of how/why that research was done. Most of the time, a paper isn’t imperative, it is informative. The “rigor” that makes up a paper is really just about the careful, honest documentation of an attempt to engage a problem, gap, or idea as well as the results or outcome of that engagement.

Why not let a paper remain informative and focus your attention on making your talk the imperative part of your work? Use the right tool for the task at hand! Papers are archival and informative. Talks are imperative opportunities.

A talk might be informative. An academic one probably is, to some degree! But academic or not, a talk is about taking what really matters and instilling that motivation in your audience. So a talk shouldn’t be successful if your audience just reads your paper. A talk is successful when it produces newness in the audience: people pursue new ideas, gain new questions, establish new collaborations, or pivot their existing work in a different direction. That’s a good talk. A talk uses the limited time and space you have, shared with other people, to implore them to do something new.

So the very first thing that you need to do before assembling your own talk is to ask yourself, “who am I speaking to and what do I want them to do?” You must absolutely frame everything else in your talk based on what you know about your audience.

A talk should center on people

This is where academics start to miss the point. A talk is not a summary of your recent work because that’s what a paper abstract is for. (If you want to give that kind of talk, just read your paper abstract and call it good. That would honestly probably be a great format for a lightning talk.)

A talk, instead, is an opportunity. A paper is eternal, but a talk is temporary. And that temporary aspect is what makes it a powerful tool. A talk is an opportunity where you can connect with real people in real space, share something, and then have chances to strengthen or embolden your audience. Talks are interpersonal things that are limited edition. The specialness of sharing space with others is rarely used to its full potential in academic spaces!

So before you convince people of anything in that liminal time and place, you need to signal to your audience that you understand them.

You’ll need to demonstrate in your talk that you know who your audience is, what they really care about, and what their challenges are. You need to establish a connection to your audience. That connection is how you get your audience to help you build a better world.

Ironically, you might think that the human-computer interaction researchers of my own field (who spend most of their time considering humans to some degree) would be good at this part of talk framing. But often, they are not! Human-human interaction is a bit harder and more nuanced than HCI and shouldn’t necessarily be approached in the same way.

Part of a good talk (after connecting with the audience) actually involves demonstrating care. You’re probably going to tell them something they’ve never thought of before or show them something that could shift how they think about things. In order to do that, you have to overcome the dreaded curse of knowledge (or the curse of expertise). You will have to put yourself in your audience’s shoes and remember: what didn’t you know before you know what you do now?

In order to take care of your audience, they need to be caught up. You need to see where they are and bring them to where you started. That might involve a little bit of framing, motivation, or mentioning related work. But the point is that you take their hand and treat them kindly. If you’re going to hope to bring change into the world, you’ll need to take care of your audience and where their knowledge, understanding, and biases are. You need to anticipate what knowledge they need before you embark on the part of the talk that really matters, otherwise you risk leaving people behind. (Doing this well takes tons of practice! It’s really hard!)

And lastly, if you connect to your audience and you are taking care of them (to catch them up to where you’re at), the last thing you need to do is make your point obvious. A really good talk has the “point” revealed before the speaker even says it. If a problem or set up is presented and you do a good job informing people of the variables and pieces in play, how things fit together should be so easy to follow that your audience might even guess what your main point is before you even say it.

So in order to make sure that the solutions, innovations, methods, or imperatives that you give the audience are effective, they should seem almost obvious.

It is also an engaging performance

I’m a theater kid. I think other theater kids can clock me pretty fast (as I can sometimes do to them), but to average, everyday (non-theater) people, I may seem a bit… eccentric. At the very least, I am pretty obviously dramatic in most of my interactions with others. But my behavior is actually a heavily-rehearsed expertise. I am constantly adapting what I do and how I do it according to the people around me, our shared (or conflicting) goals, and the overall vibes of a situation. Plus, I like to keep things interesting (and funny) whenever I can. And as other high-masking individuals can probably attest, I’ve been performing for most of my life. I might not be a skilled or practiced actor, but I’m a pretty good performer.

And a talk is a performance. I really do think that most academics fail to recognize this. In fact, I’d go so far as to say that academics actively deny this. For those that do, they are wrong. A good talk is almost completely performative. A good talk requires us to carry ourselves differently, enunciate more clearly, hold fewer breaks in our speaking, and avoid certain phrases and verbal fallbacks (such as “umm”). We need to move our eyes across a room with commanding confidence, carry an awareness of the vibes in the room and how we come across, and we need to understand the importance of rising tension and timing.

This is all a performance. And while the fundamental performative parts of a talk aren’t too interesting at first (like just how to present yourself confidently and clearly), what makes a talk really exciting is when the speaker knows when to move beyond conventions. Remember: a talk is an opportunity, a rare chance to exist at the same time as other people and impart something onto them.

You want to really engage your audience. And I hate to admit it, but in the age of TikTok, we can see how bad most engagement strategies are in talk performances. You don’t need to design your talk like a 20 second video that is designed for people with 4 second attention spans, doomscrolling for dopamine. But it wouldn’t hurt to learn from some of the strategies that more interesting people employ, TikTok and social media included.

One of the best examples of a really fun and engaging strategy was employed by Don Knuth in his “Earthshaking Announcement” talk where he revealed iTex in 2010 to the TeX Users Group and rang a bell every time he said the name of his system “iTex.” The whole room was buzzing and laughing and absolutely riveted by his old school style of swapping transparencies on a projector instead of sharing his computer screen or having powerpoint slides.

And I think of “the mother of all demos” by Douglas Englebart as another excellent example of engagement. The part of that demo, from 1968, that really stood out to me was that he opens by mentioning a “team” of people supporting him from over 30 miles away and then eventually shows Don Andrews’ hand manipulating what would eventually become known as the modern computer mouse. He dropped hypertext, the computer mouse, collaborative text editing, video conferencing AND collaborating teleconferencing all in the same demo. I can’t imagine how monumental it must have been to experience this talk.

Most recently, my favorite talk I attend that had excellent engagement was by Andrea Batch at IEEE VIS 2023. Batch was speaking on “Wizualization: a Hard-Magic Visualization System for Immersive and Ubiquitous Analytics.” The whole research project and talk were themed around magic, and Batch as well as several supporters and co-authors present were dressed as wizards (with beards and wide-brimmed, pointy hats). It was utterly delightful and I was absolutely engrossed in the performance of the talk.

A good performance requires some basics (which many other guides focus on), but pacing, charisma, confidence, and ingenuity/humor are some of the things that can make a talk especially impactful and memorable.

You should inspire new possibilities

Englebart’s, Knuth’s, and Batch’s talks all focus on a common approach: the best part of the whole talk is simply what is now possible. Yes, the performance was riveting. But the performance is really just a way to get people’s attention long enough to help them realize new possibilities.

This is because research, when done well, is the production of knowledge. While knowledge production is super important, knowledge is useless until someone figures out what to do with it. And that is what the talk is for. A talk should include why and a little bit of how some work was done, but actually leave most of those details in the paper. The majority of a talk’s real estate should be devoted to really showing why that knowledge actually matters.

Going back to this earlier piece of not-so-great advice: if the goal of a talk was just to get someone to read your paper, you’ve failed to consider what you want them to do after they read the paper. I think many academics believe something like, “if only people read my paper, then they would go out and do good things.” Bzzt. Wrong. A paper, if well-executed, is simply not a great source of possibilities for inspiration. Most of the real estate in a paper needs to be dedicated to framing, methods, results, and then near the end there is some itty-bitty room for discussion and “future work.”

(While I do suggest that “future work” sections should be devoted to what you, the author, want your readers to do now, most authors use that space for what do I, the author, want to do now, which tends to come off as a bit infertile and individualistic.)

Papers really don’t get people to do too much. Sorry academics! But I spent years outside of academia and pretty much never read a paper and then (because of that paper) changed how I did my work. But talks, blogs, books, and tools? Those things were hugely influential on how I did things.

I’d even argue that a single post on social media could probably have more imperative force than an academic paper does. And this is because papers aren’t necessarily meant to be imperative forces of inspiration.

And a good talk inspires possibilities. A talk is a chance to show and demonstrate, to convince, and maybe even to put a soul-crushing fear into people’s hearts when they recognize they’ve been doing things wrong for years. (Okay, maybe not necessary to be soul-crushing, but just consider that one option for a good direction that a talk could take.)

And you need to share your passion

And wrapped up in people, performance, and possibilities is passion. You can be weird and engaging and have plenty of passion for your work. But it is important to note that a talk occupies this magical, otherworldly space where you have the opportunity to transfer your passions to someone else in synchronized time. Relatively speaking, it is pretty hard to transfer passion from yourself to others in pretty much any other format. A talk is a special chance to be a vector for something new for someone else.

So a talk, like many social acts, has this infectious sort of role to play. If you know your audience well enough, you put on an engaging performance, and you demonstrate plenty of possibilities, then the last thing you want to do is leave people with something that lasts: desire.

I’ve had great ideas over the years. But ideas that stick with me are typically ones I am either consistently frustrated with (the same awful problem keeps boiling up in my thoughts) or I am consistently interested in (I just can’t get my brain to want to do anything else). You want your passion to stick to someone else.

I recently gave a talk at ASSETS about my collaboration with Noor Hammad at Carnegie Mellon University. We built neat system that makes watching video game streams more accessible. It’s a pretty cool technology, but I wanted the audience to realize the thing that has been stuck in my head ever since I started working on this project: the way we made streams more accessible opens up a universe of new things. We hardly scratched the surface. I am burning with so many ideas and I’ll never be able to chase them all! I’m utterly convinced that you could completely reshape how we envision making new collaborative, multiplayer games because of this tech. I wanted others in the audience at ASSETS to recognize this and to have my passion for this infect them, even just a little bit, to want to do something themselves.

And sure enough, after the talk I had 3 really interesting conversations with folks, two of which may or may not lead to tangible projects (or collaborations) down the road. One of them even emailed me after the conference ended and we have had a couple discussions since then. To me? That means that the talk was successful. Someone else wasn’t just imagining new possibilities, but had enough left over interest to want to actually do something about it.

The ultimate measurement of whether or not you shared yourself and your passion is in what other people end up actually doing. In the end, this is what everything else in a talk should lead to: actions that people take that they wouldn’t have made otherwise.

The success of a talk, and arguably any sort of meaningful and intentional social activity we engage in, is in waiting to see what others do with what we have given them.

My “4Ps of a Talk” Rubric

Since the scope of this write-up is actually about giving an undergrad advice on how to give a good talk (and a talk I will be grading them on), I figure that it might also be valuable to put my grading rubric here too.

I’ve broken the previous 4 Ps (People, Performance, Possibilities, and Passion) into the following questions that can be asked during a talk. Is this actually a really good, comprehensive rubric for judging a talk you see out in the wild? Probably not. I think if any 1 of these are done well enough, it can sometimes compensate for everything else. So the rubric isn’t really a good system to rely on. But that being said, I do think these questions are worth keeping in mind when you’re making your own talk.

People

1. Do you connect with your audience and demonstrate that you know what struggles and interests they have; ie who they are?
2. Did you explain what your audience will need to know before they can be convinced of what you’re about to say?
3. Is the intended outcome of your talk (what you want your audience to do) pretty easy to figure out early on?

Performance

1. Is your talk well-rehearsed and your messaging/wording clear?
2. Do you have confidence, charisma, and a general belief in your work?
3. Do you keep your audience engaged through demonstration and mood-setting, such as humor or varying levels of seriousness?

Possibilities

1. Is it clear to your audience what is possible to do now that might not have been possible (or easy) before?
2. Are there imperatives, calls for involvement, or clear actions that your audience can take based on your talk?
3. Is it clear what still needs to be done, in order to achieve whatever ideal future is that you’ve imagined and communicated to your audience?

Passion

1. Are you passionate (filled with joy/frustration/strength of will/etc) about the contents of your talk?
2. Is your passion presented in a way that invites others to join you and how they might do that?
3. (After a while) Have others followed up with you, furthered work based on your talk, or otherwise taken action because they were your audience?

Generally, all of these areas should be seen more like spices that could, in the right combination, make for a good talk. Academic talks in particular also rely on many other things in order to be successful (that I mention I won’t be covering at the start of this writeup), like:

1. honesty
2. relevance
3. logical/rational argumentation/defense
4. and effective slide design.

Obviously you should be honest. And you need to work to situate your work in a way that remains relevant to your audience. And of course, if your content is academic (and so is your audience), you often may need to defend your work a little bit during the talk (which depends a bit on different cultures in different academic communities). Lastly, you need good slide design if you use slides. That one can be a talk killer when done poorly, to be honest.

But these last 4 points are so fundamental that I didn’t care to spend time with them here in this write-up. Plenty of other folks have written about how to give a good academic talk (and how to make good slides) based on these. I wanted to touch on the 4 things that I think most people really miss the mark on (and why).

People matter more than anything. And the best parts of a talk are because you get to perform a little bit, share new possibilities, and infect people with your passions. Talks are a special moment we share. Take advantage of that, because this kind of thing doesn’t happen too often elsewhere in life.