Ten Steps to Better Public Speaking

Speaking to an audience can be difficult for many of us. For the audience, it can be painfully boring. Yet when done well, a talk becomes a magical moment. I've had a few of these, among a lot of disasters. I'm going to explain my strategy for becoming a better public speaker. Some people are more gifted than others. No matter: the key is not talent, it is patient effort in the right direction.

Why It Matters

Once in 2007 I organized a conference on software patents. We had a brilliant set of international speakers. The two keynotes were shockingly good, and for me, set the bar for every keynote speech I've ever given.

Mark Shuttleworth, dot-com billionaire and the first South African in space, told the story of the Internet revolution, and said a phrase I'll never forget: "Old money hates new money, and tries to destroy it every chance it gets." His keynote inspired me to write on the digital revolution, and this fight between old and new money. That became Culture & Empire.

Our closing keynote, Bill Kovacic, was an ex-commissioner of the US Federal Trade Commissioner. He didn't use a microphone, and didn't need one. His voice boomed out over the audience. Two things he said, which are not in his slides, stood out. "When Bush took office, every single anti-trust action in the USA stopped, dead." he said. And a little later, he explained why patent pools are so popular in large companies. "Patents trump anti-trust law," he explained. "When a group of CEOs get together to fix prices, that's a criminal offense and they will go to jail. But if they're cross-licensing patents, it's entirely legal!"

Such power in so few words.

For me, this is the reason to attend a conference. It is to be kicked out of my inertia. It is to hear people with years or decades of experience lift the veil on mystery. It is to leave with words burnt into my mind, changing me, and pushing me to do things I'd never dare, otherwise.

It is a tragedy to see brilliance turned into a robotic slide reader. Kovacic had slides, for sure, yet what I remember is his hammering voice, his words, and his deep passionate anger at a corrupt system.

In technical conferences, the best track is often the corridor. I love the corridor, and the long fertile chats. Yet the corridor mostly doesn't discuss the talks. "Did you see so-and-so's talk? Did you like it?" is about as far as it goes. And then we're back to more interesting subjects.

It's better than nothing. A good bubble conference like BuildStuff, bringing people from far away and locking them up with booze and food, for days, can be mind blowing. It is like going on a cruise with a hundred close friends. A few weak talks, or many weak talks, is still worthwhile, if it brings people together.

Yet we can do better. I know we can do better and I think most of us who speak at technical conferences want to do better. Which brings me to this article and ten steps to better public speaking. I'm not going to pull my punches. It took me decades of public speaking to realize some of these techniques were even possible, let alone get good at them.

1. Stop Using Slides

Slides are the training wheels of public speaking. They stop bad accidents that might scar you for life. Yet they are often the biggest barrier to improving your public speaking. Speaking without slides forces you to start learning the real skill of talking to, and not at, your audience.

Once in OSLO (NDC) I gave a talk on open source community building, and four people came. One was the sound engineer, one was an organizer checking on me, and two were there by accident. I gave the talk. The slides were, I think, excellent. My best slides ever, perhaps. And my last.

I've been giving public talks for a long time. Mostly they were the usual affair: twenty slides, five points each, weak delivery from behind a lectern. Good content, mediocre delivery, and a somewhat bored audience. And then sometimes, I'd give a brilliant, life-changing talk.

In June 2005 I came to talk at a small press conference-slash-debate organized by the European Patent Office and the FFII, fighting on opposite sides of the software patent war. This was the Christmas Football Match, a small moment of peace between the hostile camps. The EPO brought their usual suspects: bogus Microsoft-funded startups, industry lobbyists, and regulators. The FFII brought small businessmen, and me.

I asked to speak last. The two panels presented in turns. Each time the FFII's speakers were trying to show their slides, one of the organizers, from the EPO, kept adjusting the projector, playing with the audio and so on. It was not subtle.

When it was my turn I switched off the projector and said, "I'm not going to use slides."

My company had been fighting a software patent troll in the mobile app sector. I'd been talking with other mobile app firms to fight the troll. They all capitulated and paid. We refused, and the troll started attacking our clients. We had to shut down our platform. So I was motivated, and had good stories to tell.

I told off the fake startups for lying to the audience. I told off the EPO guy for trying to sabotage the other speakers. I told the audience, "patents don't create jobs. Hard work by quiet people creates jobs. Patents are a tax, a tool for lawyers to extort and blackmail. Anyone who claims they need patents to write innovative software is a crook or a liar. We will fight this directive, and we will destroy it."

I spoke for fifteen minutes, unrelenting and angry. "You think you can come into our world of software, and change the laws so you can claim and tax our hard work?" We did crush the software patent directive (thanks to the hard work of others), and the FFII later asked me to become its president.

It was a lucky accident that day. Nothing to lose and enough anger to overcome my fear of talking without slides. For a decade that experience has been my benchmark and my goal. A lot has to come together to give a successful talk. No slides seems to be an essential starting point.

Let me break it down further:

When you use slides, the audience watches the screen, not you. This reduces the strength of your communication. For instance your non-verbal body language is almost totally silent.
Slides turn the audience from "participants" to "consumers." Half their brain switches off. They can always catch up later. They open their laptops. They think about their work.
Slides interrupt your focus, as speaker. You switch from screen to audience and back. This switching is visible and makes the audience feel less important to you.
A presentation, no matter how good, is fixed. There is no way to improvise, no way to adapt your talk to the audience in real-time. And yet that is how the best talks happen.
Slides are an artificial device. We don't use them in real life, ever. No boy-met-girl story ever contained a 12 page presentation with bullet points. A successful sales pitch happens around food and drinks, not a meeting table.

I can go on. Enough to say that polishing your training wheels does not make you a better cyclist.

Let me quickly answer the common pro-slide arguments:

"I need them to remember my talk" - it's a question of practice, like cooking without recipes.
"They help the audience follow my talk" - if your talk doesn't make sense without slides, maybe you are fixing the wrong problem.
"They make it easier to remember my talk" - visual experiences are entertainment. Learning needs words and voice, work and argument.
"Organizers and audience demand them" - there are lots of conferences. Being a better speaker gives you more, not less choice.
"It gives people a chance to catch my talk, later" - this is what the videos are for. If your talk can truly be captured on slides, why even be present in person?
"I need a way to show data, source code, diagrams" - put them on line, and talk about issues that don't need such illustrations, rather than technical ones. Your talk will be far more interesting and useful.

We're so used to the crutch of slides that we've built our rituals around it. We treat our slot as a chance to dump data on the audience. And the audience expects it, groomed by years of mind-numbing meetings. Stockholm syndrome, perhaps. The audience pretends to understand and remember, and the speakers pretend to believe them. "Great slides! Where can I download them?" We all feel it's bogus yet can't quite verbalize it.

2. Remove the Barriers

When you talk to an audience, you must build trust rapidly. If you fail at this, everything you say will be rejected, no matter how plausible or well explained. Some audiences are more hostile than others. The majority will always be at least skeptical. The fastest way to build trust is to remove all props and barriers and show vulnerability.

Now we come to the deeper psychological reason for not using slides. It is about projection of power, the fixing of trust, and the opening of your audience's mind to your message.

In my work I've often had to resolve conflicts. There are some simple ways to build trust in a difficult meeting. You start by asking for token gestures, "could you pass me the water, please?" Then you remove barriers. Instead of sitting across the table from a person, you sit right beside them. Not so close to make them uncomfortable, yet within stabbing distance, so to speak.

Someone once described my speaking style as "vulnerable." This is indeed what I show, standing in front of the audience without any props to support me. And yet a show of vulnerability is not weakness. It is a claim of power. The other party either rebels, or accepts.

As Mark Rendle said once, this is like patting a large dog on the head. The dog can bite your naked hand. Yet if it does not, then it accepts your dominance. It looks like a friendly gesture, to pat a dog, yet it is all about power.

Slides are a prop. The moment you have something on the screen, you're not patting the dog's head, you're throwing it a stick. It is not the same thing at all. Watch an expert give a weak talk and you see this effect. The more they put on their slides, the less people believe them.

What other props do we tend to use? Obviously, the lectern. Don't stand behind anything. Walk down to the front of the stage and come close to your audience. If you can, get off the stage and get even closer.

How do you dress? Do your clothes make a statement? If so, they're a prop and working against you. I've learned to dress in plain, neutral colors. The less my clothes talk, the more the audience will accept my words.

How do you speak? Are you using jargon, making clever in-jokes, and referring to important companies and people? All of these are props, and they make you look weak. Learn to speak plainly, and if you tell jokes, make them about yourself.

Your own ego must disappear. The more you act or claim to be important, the less people trust you. The same goes for your technology or product. If you give a sales pitch, you are asking for rejection. Always talk from the audience's view. What problems do they encounter?

And here again, why slides are so toxic. How do you know the audience's experience unless you ask them? I like to discover this in the corridor, before my talk. Even a few conversations can give me the key to a talk. It is essentially unscriptable.

Don't use notes. I once had a last-minute panic, at a large event, and came on stage holding some small pieces of paper. The room was huge and filled, and instead of connecting to the audience, I connected with my hand. It was not a good talk.

3. Challenge Your Audience

Everyone is an expert in their own experience. Your best value as speaker is to ask the right questions, not provide the answers. As a layperson you can challenge an expert crowd. The best material you have is the data of your own experience. The best moments are entirely improvised, not rehearsed.

We come to the deeper question of how we think individually, and in a group. There is zero benefit in an exposition of existing finished material. "Here, I will now read the Wikipedia page on apple juice." Or worse, a corrupted version of that. "A fifty minute reading of my version of the Wikipedia page on apple juice." I call this Robot Reader Syndrome.

You must know the problems you're talking about in depth. This means collecting data from your daily experience, and using it to develop new crazy theories and models. It is just the scientific process: take real problems, develop wild answers, and get others to catch your errors. To present a finished product, with no space for negotiation or falsification, sterilizes the collective thinking process.

Let me give you an example. I've often seen conflict in open source communities. That's my data, and no-one can argue with it. Now I'll develop wild theories about where this conflict comes from, and how to solve it. That's my work. Then I present these wild theories to audiences who I hope have experienced the same problems. I wait, breathlessly, for them to tell me where I'm wrong. Please tell me where my models are broken, so I can improve them.

There are many ways to work on your material. Observe your own organizations and projects. Write, write, write. Talk with others, at meetups and in the corridors at conferences. Always, step away from selling an answer, and focus on identifying the real problems and challenges. While a Robot Reader is bad, a Robot Salesperson is worse.

If you are a good communicator, you can do this without even knowing the subject. It's how good consultants work. They enter a troubled organization. They ask staff, "what are the biggest problems?" They then ask others, "what are the best answers?" They write this down in a report, and charge a fat fee.

When you know your problems, you can learn to improvise around them, with time and practice. Don't give the talk people expect. Give the talk they need. Surprise your audience. I've done this a few times: propose one topic, then switch to a different one at the last minute. People mostly like that.

You may be tempted, as a speaker, to prepare a talk and then give the same talk several times. I've seen the same brilliant, interesting speakers give the same talk over months, even years. Part of this is the reliance on slides, again, which creates that sunk costs fallacy. "I've spent so long perfecting these slides that I might as well use them again." Part is the inability to improvise. Part is that culture of ex-cathedra monologues, the speeches and lectures.

Don't be that person who thinks or claims they're smarter than the room. No-one ever is. The only way to truly shine, in a roomful of people, is to be part of that room.

4. Learn the Technical Skills

Public speaking demands a set of technical skills. Some of us are naturally talented speakers. The rest of us can learn these skills with guidance and practice. This is a solved problem. One good solution: find your local Toastmasters club, and work through the program.

We are mostly better at walking than running. Yet almost anyone can complete a marathon, with practice and determination. We are mostly much better at talking to individuals than a group, yet anyone can learn to talk to a large audience.

There are a number of aspects to learn:

How to speak clearly and slowly, without, uhm, filler noises. The most effective and brutal way is to record yourself speaking a text. Play it back, and note the problems. Too fast, too quiet, dropping sounds, hesitating, and so on. If you can't hear your own errors, get someone to help. As a general rule, speak more slowly, be more precise about the sounds, and lower your timbre.
How to lose or shift your accent. Some accents are barriers, some are attractively vulnerable. It depends on the culture. In general, the more neutral you can sound with respect to your audience, the better. This can mean adopting the accent of your audience, or a neutral international accent.
How to structure a talk. There are classic structures you can learn and use, e.g. introduction, three points, and conclusion. Or, take a classic story plot like overcoming a monster, a common theme in the software industry. My preference is for less structure and more dialog, which we'll come to later.
How to time yourself. This is an essential basic skill. You must at the least offer the audience time for questions. It is a rude and inconsiderate speaker who takes all their time reading their slides, and then starts to break the schedule by continuing to talk. The conference should cut speakers off if needed, yet for me, self-timing is as important as not mumbling. Start by learning to talk for exactly five minutes. Then work up over years until you can structure a 50-minute talk on the fly.
How to use your body language. This means stepping away from the lectern, and standing in front of your audience. It can be terrifying. Nonetheless, so is driving a car the first time. Good body language for talking is relaxed, open, and vulnerable. For example, arms open and using your hands, yet never raising your hands above your shoulders, nor making aggressive gestures.
How to connect visually with your audience. Again, removing barriers is a first step. You cannot connect when you are reading your laptop screen. Then, speaking to specific people in the crowd. Speak to people close to you but not at the front. At the end of every phrase, switch to someone else, so you scan the room. If the room permits it, come down to the audience and literally talk to them. All the time, stay relaxed and smiling.

5. Learn from Failure

You will always learn more from your failures than your successes. This means always pushing yourself out of your safe zone, and then using your stumbles to become a better speaker.

It can be intensely painful, humiliating, and discouraging to give a bad talk. When you aim for perfection and you set your standards high, you will fail (by your own measure) more often. The good part is that your worst talks will still be better than average.

For me, for example, a talk is a failure if I don't get more questions than I can answer in the time. My goal as speaker is to motivate people to do further research. So a cold room (or an empty one) is data that I did something wrong.

To avoid traumatic disaster, start with short talks in forgiving environments. Build up your confidence and skills over time. After every talk, do this:

Get one or two points for improvement from someone in the audience whom you trust. "It was great," is not a help. "You spoke too rapidly," is helpful.
Get someone to record you, and watch yourself on video. It can be unpleasant to watch yourself speak yet you need to do this to see yourself improve. It can often take months for conferences to publish videos, so don't depend on that.
Experiment with different formats so you build up your range. If you don't find events that offer you what you want, organize your own. For example, one of my favorite formats is a multi-day workshop, which I organize myself.

Watch other talks and identify what doesn't work for you. Be honest about your irritation and then ask whether you cause the same feeling when you talk.

When you do have a bad talk, use this as material. It makes a great opening, and I've done this. "Anyone here from Oslo? Yes? Well, I hate Oslo. Nothing personal, but last time I spoke there exactly four people came to my talk, and two of those had come to the wrong room."

Of course, it's never Oslo's fault. As speaker, you did something wrong, and you can fix that, and make it better next time. (My main lesson from Oslo was that the talk title makes all the difference.)

6. Maintain Protocol

Public speaking is an old art with a well-defined protocol. If a conference doesn't maintain protocol, organize it yourself.

I'm shocked by how few conferences know and maintain protocol. It's not complex, yet it does make a real difference. Here is how it tends to work:

Speaker fiddles with laptop, while audience comes into room.
After some extended silence, speaker asks, "everyone ready?" and starts talking.
When speaker is done, audience applauds and walks out of room.

Here is how it should happen:

The Master of Ceremonies (MC) stands on stage, watching the audience as they take their places.
MC encourages audience to move closer, asks, "everyone ready?" and waits for an answer.
MC welcomes the audience and thanks them for coming. He then introduces the speaker by name, and cites some of their accomplishments.
MC asks the audience for a warm hand of applause, and welcomes the speaker up on stage.
MC passes the mike, shakes hands with the speaker, and walks off.
When the talk is over, MC asks the room to applaud, while speaker takes a bow (or looks bashful), and then MC takes back the mike.

This protocol isn't just decoration. It creates trust between speaker and audience, and it provides a fail safe in case the microphone isn't working (it's the MC who suffers, not the speaker). It is less stressful for the speaker. A good MC can save a flailing speaker from disaster, by intervening if the room is too cold to ask questions. And an MC can be a natural timekeeper, intervening when a speaker talks too long or bores the audience with endless slides.

If you organize a conference, please have an MC on every stage. If you are a speaker without such support, get another speaker to be your MC.

7. Build a Dialogue

Personally, both in the audience and on stage, I want a dialog. I want to be able to interrupt and challenge the speaker, and as speaker, I want the audience to do this. It is not about argument: it is about thinking together.

As speaker, you face Newton's First Law of Thermodynamics: people will sit quiet and not talk unless they feel provoked. So creating a dialog requires some planning and effort.

Start with the talk title. My best ever title was click-bait, "One Weird Trick for Making Perfect Software". I asked the organizer, is this acceptable, and she laughed, and told me it was great. And it was. A good title attracts and annoys, fifty-fifty. It fills the room with people who already want to argue with you.

Next, take control of the room. There are many ways to do this. It depends on the room, the audience, the culture, the time of your talk, and how much beer you had the night before. Here are some of my techniques:

Watch the audience and make them wait, just a little too long, until there is total silence. This works in a larger room.
Ask everyone to stand up and wait until literally the entire room is standing. Then ask them to sit again. I like to do this when I see people with open laptops.
Ask everyone to stand up, and shake the hand of the person behind them. I've tried this twice in large rooms. One time the whole audience broke out in laughter. The second time, not a smile.
Tell a self-deprecating joke about yourself. When you can get a room to laugh, you have taken control, as every comic knows.
Start the talk by asking a simple yet difficult question, then give a reward to the person who answers correctly. I bring along copies of my books for this. It focuses attention and wakes everyone up.
Thank the organizers again, and ask the audience to applaud. It's cheesy, yet it works.
When you've taken control of the room, you can start to talk seriously. You must anchor your talk in the experience of your audience. Again, I like to ask questions that show how common a problem is, or how rare a good answer is.

With large audiences (over 1,000) it can be helpful to bring someone on stage as your foil. If you follow my advice about having an MC, you have your foil. What the foil does is challenge you, in name of the audience. It is hard to chat with thousands of people, so the foil substitutes for the crowd.

I've also used a Twitter feed for questions, and that can work nicely. There is the danger of breaking the connection to your audience, as you read the tweets. The conference where I used that technique had arranged large screens at floor level, aiming up at the speaker, which was incredibly useful.

Every room is different, so you must spend time in the room and watch other talks in that same space if you can. In small rooms you don't need to pass microphones around. In larger rooms, you must. Or, you must repeat the question for the camera and rest of the room.

You can also change the layout of the room, in some cases. I've done this successfully in smaller events, where the seats started out in classic lines facing the stage, like a school class room. Ten minutes of moving stuff around, and we had a nice circle of sofas and chairs, almost like a campfire.

There are some terrible rooms: huge flat rectangles designed to hold a thousand people all sleeping at the same time. In such rooms, bully the organizers into switching off all video, then move left and right on the stage so you can speak to most of the crowd. You may also kick off by asking people to move into the center, if the room is not full.

I like to poll for questions about half way through the time, and then drive the talk by answering questions. When this works, it is great. When it fails, it's dramatic. Recently there were no questions and so I simply stopped the talk early: the audience (like me) was exhausted from the party the evening before. Having an MC on stage is a good backup: I learned my lesson there, again.

Answering questions well takes skill at improvising. You don't actually need to answer the question. Rather, it can be an excuse for exploring another interesting topic. The best questions are a direct challenge to the ideas you are talking about.

8. Keep It Simple

This is perhaps the hardest skill of all: to explore one idea in depth rather than touch the surface of a hundred different ideas. Yet it is the difference between garbage and gold. Every piece of your talk should be carrying your story forwards, not telling other random stories.

There are logical (yet false) reasons why speakers try to cover way too much. They feel they have to fill their speaking slot (rather than give the audience time to think and argue). They still use slides, so work backwards from the time: "in one hour, I can convey ten big ideas." They suffer from the "ten slides good, twenty slides better" type of false reasoning (it works better as "ten slides bad, twenty slides worse.")

I think public speaking is like writing, carpentry, cooking, or music. You add value by making your product simpler and easier to digest, not by adding more. Any fool can make complexity. Simplicity is the real challenge.

The core of your talk should be an idea or argument or model that is rare, controversial, and valuable enough to justify the moment. It should be worth the time the audience takes to travel, and the cost to you and the organizers to be there. Above all, you should be answering real questions the audience is facing now or will face soon.

You can then explain and describe your model from many angles. These are not different stories. They are the same story, told in more and more detail. Each explanation gives the audience another perspective, and if you are doing your job well, they start to see the whole picture.

Do not expect people to understand a complex new idea in a single sitting. That is not how it works. All you can do is break the ice of skepticism and plant a tiny seed. If conditions are right, the seed will grow and many months or years later, come to fruit.

If you are challenging established culture, or authority, or habits, you will face resistance. The more investment in out-of-date models, the more resistance to change. Imagine this as wind. The simpler your idea and the more focused your presentation of it, the smaller your wind profile. And so, the lower the resistance.

9. Keep the Discussion Going

Your work starts long before you give a talk, and ends long after. Subtle and deep truths can take years to hit home. And your own thinking will evolve and deepen over time. So you help yourself, and your audience, by keeping the discussion going over months and years.

I use various techniques for this. Obviously, there's this blog and my books, which act as shorthands. Instead of repeating myself, I can refer to an existing chain of argument.

This article, for instance, came from a corridor discussion (well, more of a beer and rock music discussion) with Dylan Beattie. Perhaps in the future it would make a subject for a talk, in its own right. When you can, connect the past with the future like this. It just takes the habit of writing short pieces regularly.

However, articles are opinion. Facts are more compelling. So a key part of the long term discussion is code and formal documents such as RFCs. These are not truth, yet they are easily falsifiable. If no-one uses a particular piece of code, or a given contract, you can assume they are inaccurate or irrelevant.

I also like to collect my videos, as it gives people a view that stretches over years. It is also about self-promotion, something you must do as a speaker. People need to want to come to my talks. Empty rooms are no fun for anyone.

Self-promotion sounds negative, yet it's essential. Partly you need to build and understand your own success, to be a better speaker. This is to deal with impostor syndrome, which I'll explain in the next section. Partly, you need access to interesting conferences, and that means gaining a reputation and a following.

My advice is to build your reputation by your work, no more or less. The world cares about results, not ideas or vision, or even history. Don't tell us who you worked for, who your best friends are. Tell us what you made and give us a link so we can check it out.

If you are, like me, in the software business, invest in open source. It is the number one way to build your knowledge, and reputation. Your Github profile is your CV. This is mostly true today and it is almost entirely true tomorrow. Show me a solid history of work and I trust what you say. Show me a blank page, and I'm going to start questioning your motives.

Once you drink the github koolaid, you can use it for everything. Put your talks there, and accept pull requests. Put all your writing there, and accept pull requests! Use it for your blog (and accept pull requests).

10. Manage Your Emotions

Lastly, and hardest for many of us, is dealing with the fear and anxiety of speaking to a crowd of strangers. Many of us don't feel competent, or confident enough. We are afraid of saying the wrong things, of looking foolish, and of being rejected. Yet there are ways to overcome all these emotions.

The fear of performing is called "impostor syndrome." It affects many speakers, as it affects many people who put themselves in front of an audience. The feeling is one of, "I don't deserve to be here," and it can be crippling. It is worse for anyone who suspects they were invited for reasons other than their skills and experience. It is often worse for women than for men.

All our emotions can be tamed, it is a matter of technique and practice. I've written about this on my blog and in my book The Psychopath Code. This also applies to impostor syndrome, which is essentially the fear of exposure as a fraud, and rejection.

The fear is based on a chain of subconscious assumptions. One, that we're not really good, just lucky. Two, that we will stumble, and reveal our incompetence. Three, that people will react by rejecting us in horror.

Let's tackle competence first. This is, I feel, a symptom of society placing too much burden on the individual to be special, and different. Accept that we're all ants, little pieces of a complex system that works astoundingly well. Sure, we can flicker with individual brilliance now and then. It means little. Our superpower comes from working together. Simply by existing, you add value.

Next, fear of failure and exposure. This is a symptom of culture asking us to be perfect. I've said, embrace your mistakes and learn from them. Fail with happiness and grace, recover, and continue. When there is nothing to hide, there is nothing to expose.

Last, fear of rejection. There's a small mantra you can repeat to yourself. It's our vulnerabilities that make us attractive to others, not our strengths. We reject people who are difficult to work with because they are anti-social, arrogant, and deceptive. Such people never feel impostor syndrome. The simple fact you're afraid to fail makes you precious to others.

Here are other techniques I recommend to fix impostor syndrome:

Remove the barriers between you and the audience. It feels more scary at first and then you realize that the less you try to hide, the less fear you have of exposure.
Treat failure as science. It is better to try ten things, and see eight fail, than to try just one and hope it works. For me, this means starting lots of projects, writing lots of articles, and speaking at lots of talks. Many are failures. The ones that survive are my successes.
Before the audience comes into the room, go onto the stage, and walk around. Tell yourself, "this is my stage, this is my place, and the audience are my guests and friends."
When you talk, never lie, bluff, or make claims you can't back up. When you don't know something, just say, "I don't know." There are people who are professional liars: you're not one of them.
Stay busy. If you have a talk coming up, write an article about the topic. Keep researching it, and try out different story lines on people.
Trust yourself. You are the child of a billion generations of survivors. Good luck has nothing to do with it. You have winning genes backed up by hard work.

Conclusions

If you're passionate about your work, then sooner or later you will want to share that passion with others. There are many reasons to want to become a good public speaker. Perhaps the main one is simply to get out into the world and meet interesting people. For me, it is a way to learn from others, and shake up my ideas in a way that's impossible working alone.

I've explained my best techniques for capturing an audience's attention, and getting them to consider your ideas seriously. It is not easy. The typical conference attendee remembers only 9% of what they hear, by the end of the week, and a month later that figure is down to 0.1%. OK, I'm making this up, yet you get my point.

Stop using slides, get closer to your audience, and challenge them. Engage them in dialog, and make them think. Keep your voice and body focused on the room. Use no props, gimmicks, or notes. Leave your fear off-stage, and if a talk bombs, accept that with grace.

It is a slow, long process, so be patient and always kind to the organizers and staff.