How to: give a great talk
You were just invited to give a talk. Congratulations! Why did they pick you? No, really; why did they pick you over someone else? It was probably because you have something interesting to say. So be interesting. Don't make the person who invited you regret it. Don't be the reason your audience fell asleep -- they were nice enough to attend in the first place. And be memorable. Give a great talk and people will remember it [Nobel laureate Liz Blackburn said this too], and they just might invite you back. And maybe next time, the invitation will be to talk in a far-away, exotic location. But even if it's not, please learn how to give a good talk.
1. Who's your audience?
- The amount of introduction required; a small, specialized meeting on your topic won't require much introduction; a larger meeting with a scientific audience will require a brief introduction to the topic you will present and the scientific tools you use; a presentation to a lay audience will require a thorough introduction as well as putting your scientific question into the big picture, and justification for studying it.
- The depth of your data slides; a lab meeting presentation will show primary data; a longer scientific talk will show some data, but also summaries through model slides and schematics; a shorter scientific talk or a presentation to a lay audience will have few (if any) primary data slides.
- The take-home message; Begin with the end in mind. What are you trying to say? Your message will be consistent throughout your presentation, and make sure it's appropriate for your audience. What is your most exciting/significant finding? What does this mean? Again, lab meetings might end with future experimental directions, whereas your fund-raising presentation to the private foundation might end with the profound scientific implications of your findings. Make sure the take-home message is clear.
We'll assume for the purposes of this post that the audience is a scientific audience, and your talk is polished. [I will write elsewhere on how to give an unpolished talk i.e. lab meeting, and perhaps will write on giving a talk to a lay audience too.]
2. What's your message?
- Pick one message: you have a choice here: tell one story the audience will remember or tell multiple stories the audience won't remember.
- Stay focused on this message, and don't stray off tangents. Scientific writing and scientific thought is concise and focused; scientific presentations should be the same. Why are you showing 'this' slide? What does 'this' slide contribute to the message? Does it support it? Can it be removed?
- Why is your message important? You'll want to start with the importance, remind your audience why it's important in the middle, and conclude with the importance.
- Finally, practice; practice telling your message. Tell it to your spouse, your dog, goldfish, colleague. Practice your elevator pitch. Tell people with varying levels of expertise in your topic, which will help you hone your message, and force you to think about it in new, and perhaps more clear ways.
3. Start with a blank slate.
4. Show, don't tell.
- Take the listener on a journey, showing them what you're working on (background), and putting it into context into the rest of the world or science they might know. This story should be a narrative throughout your presentation, and hold the listener by the hand, taking them along with you as you layout your science.
- Use pictures, and remove as many words as you can from your slides. [Read Slide:ology by Nancy Duarte for more on this idea]. While humans have many skills, reading and listening at the same time is not one of them. Having too much text on your slide will cause your listener to stop listening to what you say, and instead will cause him/her to start reading your slides instead. And now you've lost your audience. You might have to learn how to draw in Illustrator, but your images will be pretty, and your slides will have less text and be more clear.
- This will also help you learn your talk. You shouldn't memorize word-for-word, and you shouldn't read off the slides. Instead, each slide contains an important point, and knowing that point will allow you to describe it to the audience.
- Use the 'slow reveal'. You already have one slide for one idea. However, if you have a slide with several parts to it, and you don't want your audience reading ahead or to become overly visually stimulated, use a slow reveal technique. Presentation software will allow you to sequentially add parts to your slide, add arrows, remove blocked panels, etc. Use these.
- Be clear. Don't be confusing. Don't be funny.
5. Practical notes
- Plan for 1 slide/minute (any more and you're speaking too fast, and risk going over your allotted time, any fewer and...well, yes...use fewer.)
- White background for informal presentations (use most of the time)
- Dark background for formal presentations (a special occasion for when you get the Nobel)
- Don't use too much color. Does your slide have a colorful, but distracting background color? Does your slide look like a Sony Brevia bouncy ball commercial? Make your colors meaningful, not just pretty.
- Start by introducing yourself, (at least stay your name even if you were just introduced, which can be very effective), and tell the audience what you'll be talking about. Don't read the title of your talk; the audience can do that themselves. Introducing yourself also marks the beginning of your talk.
- Explain exactly what you did or what you show (a blind person should be able to hear your talk and understand it completely)
- Use big font and descriptive text (a deaf person should be able read your presentation and understand it completely); titles should be descriptive (instead of: "Relationship between Gene X and Cancer", try: "Gene X Increases Risk of Cancer by 20%"); use models and diagrams whenever possible (instead of a text description of your protocol, try: a schematic showing the important steps).
- Look the part; scientific presentations are often given in informal environments. Informal is great; messy and unkept is not.
- Use a laser pointer to point, and that's it. Your listener's eyes will be drawn to where ever you point, so make sure you're not circling items on the screen with the pointer, make sure you're not keeping the pointer lit as it moves off the screen, and make sure you're not turning your back to the audience as you use it -- stay engaged.
- Use color: scientific presentations are often presented in black and white, because journals charge for color, and data is often pulled from previously prepared manuscript panels. Don't be afraid of color, and use it wisely. But be consistent: each time you perform a treatment, use the same color; each time you use the same sample, use the same color. Make your colors meaningful, not just pretty.
- Give credit for images that are not yours; however, don't feel the need to put the entire URL on the screen from where you got the image, as this can detract from the message on your slide.
- If your "Conclusions" spill onto a 2nd slide, you're trying too hard. Remind audience of your message.
- Let your audience know that you've finished; for example, "thank you for your attention and I'd be happy to answer any questions you have at this time". Omitting a closing statement can lead to an awkward silence and uncertainty among the listeners. Because you didn't go over your planned time, you have plenty of time for questions...right?
Additional Resources (in no particular order)
This is a post in a "How to:" series specifically aimed at scientists-in-training, including undergraduates, graduate students, and post-doctoral fellows. I'd like to acknowledge the resources listed above for contributing ideas on giving great talks, and ridding the world of terrible ones, and Frederic Terral of RightBrainTerrain for designing the blog entry artwork and creating awesome art in the public domain.