The worst conference talk ever...

And how to avoid it being yours!

Huh? Isn't this blog usually about travel and merciless self-promotion?

Yes. But this entry is different. Nothing I say here is particularly original, but I'll attempt to synthesize into a shortish article containing some pointers for speakers. The subject matter is conference talks, but a lot of what I have to say applies to any public speaking event.

Communication!

A conference talk is about communication. Communication takes two parties - the speaker and the listeners. Fundamentally, you must ask yourself mercilessly if what you are doing is communication, or if you're improving the extent to which your talk is good communication.

Is my talk communicating well?

Context

A conference talk is not like a lab talk, a journal club, or a lecture course. At a conference, people working on similar or different things are getting a broad overview. In a 10 minute talk, there is not really time to give a comprehensive view. A conference talk's role is to update the audience on the state of research and knowledge, and to pique their interest to discover more. They can always find you over bagels or read your papers if they need more detail. The audience are generally professional physicists, but they know less about your field than the average peer reviewer, and that should be kept in mind.

What does my audience want to hear? What don't they already understand?

The speaking bit

Speaking is the most important part of a talk. You'd think, given the name "talk", that this would be obvious, but most people tend to treat what they say as optional ambiance for their slide show. This is a fundamental mistake. A good talk is well written and can stand on its own without slides. It could be read over the phone or heard on the radio. 

If you are not accustomed to public speaking, you should practise. A lot. Basic concepts include clarity, grammar, projection, speaking slowly, and so on. Many of the audience members will speak English as a second language, and are sitting some distance away. Repeat audience questions. Encourage the audience to wave if they can't hear. 

The content of the speech is also incredibly important, but often neglected. Many speakers read their slides. Words cannot convey how wrong this is. A speech should read like a pithy, preferably witty essay or short story. Given time constraints, it pays to think carefully about what you want to say, construct a compelling narrative, use evocative language, and synthesize ideas and concepts with the central topic regularly. Unless your results are perfect, acknowledging shortcomings with self deprecating humour is usually a plus.

Is it easy for people to listen to what I'm saying? 

Presentation

Comb your hair and wear nice clothes. People should feel comfortable looking at you and your face - it's very important for communication. Similarly, room lighting should be bright enough for people (even ancient professors) to read fine text. A lot of the conference is about networking and recruitment.

Dress up for a job interview or better.

Slides

This is by far the most abused part of conference talks. There are many ways to screw it up. Fundamentally it comes down to communication. A sequence of slides is great for telling a story with a strong sense of chronology, so write your talk accordingly. A sequence of slides is terrible for illustrating connections in complicated systems, so be careful.

Does my slideshow actually contribute to my talk?

Quantity of slides

If you try to split a paper or thesis up into a slideshow, you will have too many slides. As a general rule, one slide per minute of talk is about right. Less is fine. More is definitely not fine. If you have difficulty fitting your 10 minute talk onto about 10 slides, the answer is not to make your figures smaller. Break your talk up into 6-8 key concepts or results, and have a slide for each.

Does each slide actually add substantially to the whole?

Use of slides

What are slides good for? People can speak and listen more quickly than they can read, but a good diagram speaks better than thousands of words. As such, slides should be used almost exclusively for diagrams, figures, graphs, plots, and so on.

Is each element of my talk contributing the best it can?

Slide dos

Do use big diagrams. Do label diagrams appropriately. Use labels rather than keys, label axes, title graphs. Title each slide with a short, pithy, and specific label. Keep it simple. You can always answer a question with overwhelming detail and erudition.

Is my slideshow an exercise in communication?

Slide don'ts

Don't use bullet points. Especially don't use hierarchical bullet points. Don't use 'introduction', 'conclusion' or any other generic term as a slide subject. Don't use block text. Never read text on the slide.

Have I done anything which will make the audience want to shoot themselves or me?

Good figures

There are many resources on the art of making good figures. But in general, you know a good figure when you see one. Colours are used to transmit information. The colour palette should be chosen so that colour blind people won't be too confused. Diagrams and graphs should be devised to be as simple as possible. Axes should be clear but unobtrusive. Information should be dense. The most important basic information should be immediately obvious; subsidiary detail should not dominate the figure. Tables are better for small data sets. The take home message should be obvious. A figure might require explanation - that's why you're there, and that's okay.

Is my figure a work of art?

Lolcats

Visual punctuation is important, especially for didactic talks. If you want people to remember things, a change of tempo is a good idea. Eg "lolcat says always check your dimensions" is a lot more memorable than anything else. That said, overuse cheapens the effect and shouldn't be necessary at all in many contexts or with a sufficiently well written speech. Be aware of political correctness when executing any naughty jokes.

Am I comedian? How do I want to be perceived? Will people tune out?

Title slide

The title slide is important, since it'll sit there while you're being introduced. It should have a relevant figure on it, preferably enticing and visually stimulating. It should also contain the title of your talk, your name, your collaborators (be inclusive), your institution, and the conference you're at. This is to remind you what you're talking about, and to help you check you have the right slide show. A double title slide is a trick where you have two nearly identical cover slides. While you're being introduced you can switch between them to check the show is working, without distracting the audience too much.

Is my title slide a throwaway? Is it worthy of the subject matter?

Audience participation

Audience participation is generally a good thing, as it promotes engagement. The best talkers regularly pepper their talks (especially longer ones) with anecdotes and refer to people present in the audience, and even ask questions (especially in a lecturing context). When I'm giving a lab talk, I usually start by polling everyone in the group on their knowledge of the topic. It generates a framework. In the context of a conference talk, asking people to raise hands if they can hear you, especially if you prefer not to use a microphone, is a great way to turn their brains back on after hours of semi-comatose sitting. If your talk concerns something of a mystery, getting people to take a guess at a unknown result is also a useful way to engage. Eg "Show of hands who thinks the Higgs mass is between 120 and 130 Gev?" In a smaller seminar, get and keep everyone talking.

Do I have any evidence anyone is listening to me?

Lasers

Never under estimate the importance of lasers. Technically, you should be able to give a talk without a laser, but I find it's much more fun to secretly have a laser pointer capable of de-orbiting the space station. It helps to foment a feeling of awe in the audience. You can also claim its astronomical cost against tax. On the other hand, if you ever point it anywhere except the slide, you're an asshole who is gambling with people's vision.

Can I afford a bigger laser?

Flexibility of presentation

If the talk before yours introduces the same material and it's the same audience, skip your introduction. There is nothing worse than hearing "Professor Y spoke about this, but it's on my slides, so I'll go through it again anyway". There is nothing better than saying "Thankyou Professor Y for introducing my talk so well, this diagram is exactly the same as blah, next slide". If someone asks a good question during the talk, change the talk as necessary. Have spare diagrams at the end of your talk that you can jump forward to if you need to.

Is my talk structure more important than its purpose?

Audience etiquette

So now we know what makes a good show. How to spread the word? What is the best way to stage an intervention in a terrible talk? Some people walk out. Some people openly criticise. Some people play guerrilla laser tag with their laser pointer. Unfortunately I do not have any answers for this. But at least every time you see someone do something silly, you'll know not to do that yourself.

Is my life long enough to put up with this affront to informational aesthetics?