One of the biggest hurdles to giving a good talk is convincing people that it's worth their mental energy to listen to you. This approach to speaking is designed to get that buy in from the audience, without them even realizing they are doing so. The key to this is exploitation of a simple fact: people are curious creatures by nature and will pay attention to a cool story as long as that story remains absolutely clear.
In the D.A. style of speaking, you are the narrator of an interesting story. The goal is to have a visually streamlined talk where the audience is so engaged with your presentation that they forget you're standing in front of them speaking. Instead, they're listening to your narrative and seeing the visuals that accompany your story, at no point do they have to stop and try to make sense of what you just said. Here are the key points:
Get into this mindset: your main job is to be an entertainer, not a scientist. Most scientists don't do this, which is why most scientific talks are bad. The fact of the matter is that if the audience doesn't understand and enjoy your talk, they won't care if your science is good.
Tell a story; don't simply talk about your methods and results.
A solid scientific narrative is critical for easy comprehension, and easy comprehension is critical for the D.A. style! The audience needs to understand every word you are saying, because once someone is no longer paying attention, the spell is broken. Your goal is to never lose someone's attention. Thus, a clear narrative arc is essential for success. See talk outline suggestions below.
Practice your transitions between slides and subjects. The most common place to lose people's attention is with a bad transition. Once someone is no longer following along with you, it's hard for them to get back into the flow of the talk.
Kill clutter. Remove text. Complete sentences are to be banished from your talk. Rather than write full sentences, speak them while pointing people's attention to visuals that reinforce your point. Exception: A single sentence (or sentence fragment) at the top of each slide that captures that slide's main point is a good idea.
Only show as much of the slide as people need to understand the point you are making at that precise moment. I often hide most of the slide, revealing additional detail only once it is needed to understand my narrative. For this to be most effective, learn to use custom animations in PowerPoint, Keynote, etc.
Be excited. If you're not excited about your work, why should the audience be excited?
Be engaged. Do not speak to the ceiling, floor, or (worse yet) your presentation. They are not the audience. Look at your audience, or if you're nervous, the back of the lecture hall. People will feel like you're engaging with them, and also makes them feel watched. They will pay better attention.
Jokes are tricky, you probably have more to lose than to gain. While your job is to be an entertainer, this is still a scientific talk, so keep silliness to a minimum. A good joke can really rally people to your side, but the audience has to already have a positive disposition towards you for this to work. A poorly timed or executed joke can do a lot of damage, so keep jokes to a minimum unless you really know your audience.
Try not to use a laser pointer. They're a crutch for you and are distracting to your audience. More to the point, there's no need for one if your slides are properly designed (uncrowded).
Setting up the talk. This is probably the most important and difficult part of the talk to write, but this section is critical for framing the narrative arc. The rest of the talk follows naturally from the intro. Keep it short and keep it direct.
Big picture context. Imagine David Attenborough talking about how cool Madagascar is while the screen shows aerial footage of a majestic mountain range. You want something big like this. I find it helpful to memorize the broad overview I give during my first slide.
Key questions you are investigating (zoom in from the big picture). Scientists love interesting questions even more than they love answers. You will set up a nice talk if you can ask questions in the beginning that you later answer.
Don't bury the lede. Get quickly to your main result in broad brush strokes, and then move on. People really don't like suspense if they're not already invested in your talk, and they have bad memories. So tell them the main answer before going into the details. Rather than try to figure out where you're going, they'll be able to concentrate on the finer details of your talk.
Methods and Results.
Keep methods brief. Most don't care about methods, it's a distraction from the story. Provide enough detail that they know what you did and have some confidence that you know what you are doing.
Answer the questions you raised earlier. Tell the story as clearly as possible, keeping in mind that you are answering the questions laid out in the intro.
Concluding the talk.
Briefly recap the answers to the questions laid out in the intro.
Provide context for why your results are important in the big picture. See the symmetry? We start with the big picture, and end with the big picture and how your results affect the way we think about the big picture.
The final steps are to:
Get feedback from your colleagues fairly early on, so that you are not entrenched in your way of doing things yet.
Practice your talk until you can do it in your sleep.
Some final tips for crafting your talk:
Know your audience. Present something that they will enjoy, for scientists this usually means tailoring the level of technical detail (and amount of introductory material) to your audience.
Anticipate questions and put the answers after your acknowledgements. In addition to allowing you to really nail the Q&A, this is a good way to handle very cool but tangential details that you omit from your main narrative of your talk.
Don't bury your conclusions in acknowledgements. You don't want to cleanse your conclusions from people's minds just before the Q&A with a 5 minute long thank you list. I like to integrate key acknowledgements into the talk. A good way to do this is by showing a thumbnail picture of key people in one corner of the screen when discussing their contribution.
If you have a genuine personal connection to the work, consider using this as motivation in the intro. People are empathic by nature and this can really engage them in your work.
Keep to your allotted time - it's disrespectful to your audience if you force them to stay late or leave your talk while you're still speaking. I guarantee people will think more highly of your talk if you finish on time, even if it means leaving out some great science.