A Non-Talk on Giving Talks

The following is a collection of almost entirely subjective advice about giving talks. Mostly this advice is based on what has worked well for me, but it’s also based on what I have learned in the course of sitting through terrible talks. The only good thing about terrible talks is that they afford one the chance to contemplate the sources of one’s suffering, whereas good talks are engaging and keep one focused on the subject matter. This advice is mostly geared towards scientific talks for a variety of audiences, but hopefully these guidelines are generic enough to be useful outside of science communication. 

First, a bit of motivation.

Why Does Public Speaking Matter?

At the most basic level, giving talks is one of the primary ways of disseminating the results of your work to others. Academic publishing is where your results become part of the permanent record of research of course, and for better or worse, your publications are one of the primary metrics by which your success will be judged. Public speaking is more versatile than publishing, however, as public speaking is often how your work reaches those who would otherwise never know about it– be they others in your field, or those completely outside of science (“the public”). Your academic papers are likely to reach your immediate colleagues, those within your subfield– but through talks, you can reach both scientists outside of your immediate subfield, and the public at large.

And why does reaching a larger audience matter? Because the health of both your career and the field depend on it! 

If you are invited to give a professional (academic) talk somewhere, chances are that the people who know something about your work had a hand in inviting you in the first place. If they know about your work, and they invited you, they are already “on your side”– they think what you will say is interesting enough that they’re willing to hear you give that talk. So, in some sense, your talk isn’t really for them– it’s for the people in the audience who are outside your subfield. Your goal for the talk should be that they come away with a sense of how your work fits within the context of the field, and why it is interesting. If you are early in your career and hoping to get a job, people in unrelated subfields are the faculty swing vote– you need them to be pulling for you when the department is deciding who gets the job offer. 

Public speaking also helps your field as a whole: most of our work is funded, to some extent, by taxpayer dollars. Those taxpayers have immediate needs that likely supersede their interest in space and/or science in general. Therefore, if you are giving a talk to the broader public (generally speaking, anyone outside of your field), your ability to convey what you are learning represents a return on their investment. Personally, I love being an astronomer, and I feel that it’s a privilege to work on something that takes me so outside of myself on a regular basis. When you speak to people outside your field, they’re giving you some of their time in the hopes of sharing in that experience– respect that by giving a good talk.

And now, on to my utterly subjective guidelines for doing so.

Consider Your Audience

In my view, the number one contributing factor in giving a good talk is empathy. Empathy plays out in a number of ways in a talk, the most fundamental of which is the realization that the talk is not for YOU, the speaker, but for THE AUDIENCE. Before you create a single slide, the first thing you should do is consider who makes up your audience. Are they experts in your subfield, scientists from a completely different discipline, non-scientists..? If they aren’t scientists, who are they? “The public” is not a uniform bloc of people, so consider why they are attending your talk to begin with. In some sense, every audience consists of both novices and experts, and the trick is to create a talk that will reach them both. 

Missing ingredient.

Missing ingredient.

Above all, remember that no one in the room is more interested in your work than you are– if they were as interested as you are, they’d probably be doing it themselves! As the speaker, your job is to convince them that what you do is interesting and worthwhile. 

Start At The End

Whether your talk is a few minutes or an hour, you need to decide what you want the audience to know when they leave. This should not be a laundry list– it should be a single sentence that summarizes the main point you want people to walk away with. If you have lots of time to speak, you can nest a couple of subpoints under this main message, but the main message should be clear in your mind. 

Choose Your Content Wisely

The message you want your audience to take away should guide all your subsequent choices about content. What are the main pieces of information you need to include to bring your message to the audience? If it’s a short talk, you may include only a bit of background and few examples of your work (say, a plot or two), whereas if you have a lot of time you can include more detail, or spend longer on the examples you do include. The main point here is again to exercise EMPATHY: the talk is not for you, it’s for the audience. When you see people speeding through information, or there’s a bunch of plots in the talk that they don’t really explain (or worse have to click past), it means that they didn’t consciously think about what information to include and, as a result, have included too much that’s only tangentially relevant and probably uninteresting to the audience. 

Don’t Go Over Time

Seriously people: DON’T GO OVER YOUR FUCKING TIME. It’s disrespectful and makes you look unprepared. We all make mistakes, and sometimes perhaps you’ve misjudged slightly– but if the session chair tells you you have 30 more seconds and you go on for another five minutes (or another half hour, as I have seen in colloquia), you are being an all-caps ASSHAT. Once again, not everything you have done is relevant and interesting, so pick the part that is. 

Consider Making Your Talk Modular

If you’re giving a longer talk, part of choosing your content is thinking about how to make it digestible. Human beings have attention spans of roughly 10 minutes or so, which means that it can be helpful to remind people of your main points repeatedly during the talk. If possible, break results and subtopics into subsections in the talk, so that you have mini-conclusions throughout the talk. When you finish, you can revisit these mini-conclusions all together to provide a succinct snapshot of your main points.

Making your talk modular is also really helpful for adjusting the amount of content you present in real time. Different departments have different audience styles– for example, the style in my department at Princeton is to ask questions throughout the talk, rather than waiting until the end. That can be great, as it shows you the audience is engaged and curious, but it can also derail a talk that would have been the perfect amount of content if there wasn’t time out for discussion midway through. To account for the range of styles out there, I often try to keep some of my content in reserve and include an extra optional mini-section at the end. Then, if I get through the primary part of my talk and still have time, I can click on and present some bonus material, a work in progress, or even go into further detail about something I presented in broad strokes earlier. 

If It Doesn’t Convey Meaning, It Doesn’t Belong On Your Slide

You can include pictures to provide a bit of visual interest, but detailed plots, equations, and even schematics are information-rich graphics and do not belong on your slides unless you are going to discuss them. If I see one more person put up a plot and say “I’m not going to talk about this here…” UGGGGGGHHHHH WHY IS IT ON YOUR SLIDE YOU’RE HURTING AMERICA.

Do You Want Me To Listen Or Read?

Humans love the idea that they are good at multitasking, but the truth is that there is no evidence that multitasking actually exists. What DOES exist is rapid serial tasking, where you switch attention between multiple tasks. If your slide is full of words, your audience is forced to choose between listening to you and reading the slide. Chances are, though, that they will try to do both, and your message will be lost in that conflict of attention. 

Once again, wordy slides are often an indication that the speaker lacks empathy: the words are there because THEY want to be able to read them, so they don’t forget. Listen up: slides aren’t your notes. There is a notes section in the speaker tools, so use that if you need to, but really you should just suck it up and…


The gold standard is to have done a dry run of every talk you give, so that you know it by heart. For long talks this can be pretty time-consuming, however, so usually you can get away just going over your talk slide by slide and recalling your main point. For each slide, you should have a clear single point in your mind– the point of practicing is not to memorize every word, but rather to create an association in your mind between that main point and the visual of the slide itself. When you look at the slide, you should be able to recall your point. If you can’t, you either haven’t practiced enough, or maybe it’s just not a very illuminating slide. 

Practicing in front of an audience can be hugely helpful– if you can enlist a few friends and/or colleagues and ask them to give you honest, constructive criticism, it will greatly improve your talk. It will also help you to absorb this criticism if you listen carefully to feedback without defending your choices! When responding to critiques, challenge yourself to phrase your responses as questions. 

One other very helpful tool is to make a video of yourself practicing. This is hands-down my least favorite thing to do– I hate watching myself on camera, and invariably I discover that I have some weird physical tic that I didn’t know about, e.g. “Why am I looking at the ceiling so much? Wow, I gesture a LOT. What is that dumb thing I’m doing with my mouth while I’m thinking?!” Of course, the reason this is helpful is that being aware of these behaviors is the best way of getting rid of them. If they distract you watching yourself, they’ll be distracting to the audience too!

95% Of Your Presentation Software’s Features Do Not Exist

You know all those templates that are options in your favorite presentation software? Delete them. OK, you can keep the simplest one. Now go into the slide masters section and get rid of everything that isn’t a blank slide or a slide with a single text box for a title. Bullet points? You don’t need those. Columns? No. Prefab spots to put an image with bullets next to it? No, no no no. These things are dead to you from now on. 

Look at that font menu! Now stop looking at the font menu! Pick a single font and use only one font in a given presentation. Don’t make it Comic Sans, Papyrus, Curlz or anything that resembles these fonts (I’M LOOKING AT YOU, CERN). You probably also don’t need to put anything in bold or italics either. Emphasis can often be conveyed effectively (and more subtly) by using differences in font size or slight differences in color saturation (as in a pale blue versus a slightly darker blue).

Slide transitions: these also do not exist*. They are distracting and unreliable, being one of the top things that often gets screwed up when transferring a talk between laptops. I once gave a talk that I’d written with a relatively mild fade between slides– but when deployed on the presentation laptop, that fade had been transformed into a horrifying pixel dissolve. Finally, the all-transitions-all-the-time software Prezi 100% doesn’t exist at all. If you want to be seasick, go on a boat ride.

* One caveat here: Dan Foreman-Mackey convinced me that there is, in fact, one acceptable slide transition… anvil. Anvil may ONLY be used to deliver the punchline of a joke, and that joke had better be funny. Also, you can only use anvil once a year. 

Anvil in action.

Anvil in action.

Give The Audience A Guided Tour Of Your Plots

Many times, plots that may be acceptable for publication don’t display well in a talk– but a little bit of extra attention can really help your audience out. Use the text boxes in your presentation software to relabel axes, provide arrows to features you want to point out, and use shapes to selectively block out potentially distracting features (e.g. complex plot legends) until you’re ready to walk the audience through them. 

It can also be helpful to use the slide title to state the point of the slide. No one needs you to title your results slide “Results 1”– that’s perfectly good real estate you could be using for “No Trend Seen With Stellar Color” (or something relevant to your work). 

Give Credit Where Credit Is Due

Always include attribution where appropriate, whether to your colleagues or yourself. If you use a figure or cite a result from someone else’s work, include a short reference (e.g. Batygin et al. 2012) nearby on the slide. To keep the visual noise low, make it in a lighter, smaller font– but don’t leave it out! Proper citations indicate that the speaker is not only collegial, but well-versed in the literature. Similarly, if you use images off the internet, try to find public domain work and include attribution wherever possible. 

Don’t forget to give yourself credit, too– especially in job talks, it’s important that your audience (which includes the hiring committee) have a clear picture of what part of your work YOU are actually responsible for. Of course, be gracious and acknowledge your co-investigators and their contributions, and don’t represent yourself as having done work you didn’t directly do… but you can still alternate between using “my colleagues/team/co-investigators” and “I”. 

On Answering Questions

Audience members ask questions for one of two reasons:

  1. They have something to say
  2. They have something to ask

Unfortunately I sometimes think #1 is more common than #2, even when #1 is couched in a sentence that ends in a question mark. If you are relatively new to presenting your research, questions can be harrowing– it can feel like you are being tested and that makes many people anxious. The top thing to remember is to stay calm! You did the work, and that makes you an expert on the piece of it that you did. You may not know everything about the topic, but that’s OK– give the best answer you can, and if you need to think about it a bit offline, just say so. You can also offer to discuss it further after the talk is over. Remember that questions (even if persistent or aggressive) are a sign that the audience is paying attention, and are vastly preferable to a room full of silence and ennui.

Relax & Have Fun

Inhale, exhale, repeat! It’s a privilege to have the attention of your audience, so try to enjoy the moment as much as you can. The people whose talks I like are typically relaxed and conversational, where I can see that they like their work and are looking forward to telling me about it– it’s really hard to watch someone who is really uncomfortable speaking. I personally get very nervous right before talks, but I try to remember that the unpleasantness of pre-talk jitters will go away shortly after I start talking.

A Few Additional Links

For additional reading, here are a couple of other good guides to giving talks:



Lastly, if you find color choice challenging, this site can help you choose an appealing color palette for your talk:


The one caveat to the above, though, is that you should limit the number of colors in your talk to about three at most– many of those palettes are really bright and busy, and will be tiring to the eyes for most viewers. 

This entry was posted in career, science and tagged . Bookmark the permalink.

One Response to A Non-Talk on Giving Talks

  1. *Kudos* for the idea of using the chart title for the punch line. Bigger print, featured billing, etc. I’ve done it often with a bottom line but a top line is even better. The other advice is great and though often mentioned it is just as often neglected.

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s