Delivering a keynote that actually matters
Five principles. One job.
After twenty years of speaking at conferences from Silicon Valley to Mumbai, I’ve watched more keynotes than I can count.
A lot are forgettable.
Most are unsurpricingly similar.
Some are actively painful.
And a tiny handful genuinely change how people think about their work.
The difference isn’t the speaker’s résumé. It’s not the production budget or the clicker choreography. It’s whether the speaker understands what a keynote actually is.
A keynote isn’t a presentation. It’s a performance with a purpose. And in my book, if you get these five things right, you’ll deliver something people remember long after the conference swag hits the recycling bin.
Priciple one: It’s a story, not a slide deck
If your “keynote” still works when you email the deck around afterwards, you didn’t have a keynote. You had a document.
When I grabbed the mic at Superweek to talk about “It’s the process, stupid,” I didn’t start with an agenda slide. I started with a moment, the one where someone says, “Let’s feed the analytics data into our AI system,” and you smile politely while your stomach drops, because you know your data quality isn’t ready for that conversation.
That’s a story. It has tension, stakes, and a character facing a choice.
Too many keynotes are slide decks wearing a human. Bullet points dressed up with stock photos. The speaker becomes a narrator for PowerPoint, reading what everyone can already see. Slides should be visual punctuation, not paragraphs. When I presented “The Age of Activation,” the slides weren’t the talk, they were the beats of the talk. A chart here. A single phrase there. Something that made the point land harder, not something that made me redundant.
Think about it. When did you last remember a bullet point from a conference?
You don’t.
But you remember the speaker who told you about the time their entire analytics infrastructure collapsed because someone changed one tag and didn’t tell anyone. Because that’s a story.
-
In practice: before you open Keynote or PowerPoint, write the story. What changes? What’s at risk? What did you learn that your audience needs to know?
If you can’t tell it over coffee, you don’t have a keynote yet.
Principle two: It’s about the audience, not you
Here’s the trap every experienced speaker falls into. You’ve done interesting work. You’ve solved hard problems. You’ve collected war stories. And you want to share all of them.
Don’t.
Nobody cares about your clever solution to a problem they don’t have. They care about their problems. Your job is to help them see those problems differently or give them something they can actually use.
When I spoke about digital maturity in Amsterdam, I could have spent forty-five minutes explaining IIH Nordic’s framework.
Instead, I focused on the questions everyone in the room was already asking themselves: Are we keeping pace or leading? Are our investments creating impact or just noise?
It’s a subtle shift, but it changes everything. One approach centers on your expertise. The other centers their reality. And that’s the whole point.
People already see you as a expert, that why both you and they are there.
-
In practice: for every story you plan to tell, ask yourself, “What decision does this help my audience make better?” If the answer is “none,” cut it. If the answer is “it shows how smart I am,” definitely cut it.
Principle three: It has to be memorable
Three months after your keynote, what will people remember?
If your answer is “the comprehensive framework I explained in detail,” you’re in trouble. They won’t remember it. They’ll remember the moment they laughed, the metaphor that clicked, or the single sentence that made them uncomfortable, in a useful way.
When I talked about decision analytics at MeasureCamp Munich, people didn’t like it because I had the most data. It was because the message stuck: We’re optimizing for a world that won’t exist. Not “things are changing.” Not “we need to adapt.”
That kind of sentence has a hook in it. It doesn’t let you shrug and move on.
At AllWeb in Tirana, I spoke about campaign tracking, one of the driest topics in our industry, but framed it as., There have never been more ways to go wrong and wrapped it in a story about solving murders.
Everyone in the room felt that in their bones, because they’d gone wrong recently.
-
In practice: try to build your keynote around one sentence. A single idea that people can repeat in the hallway and remember when they get up the next morning or the morning after that. Everything else exists to support that sentence. If people can’t summarize your talk in one line, you gave them too much, and they’ll remember too little.
Principle four: It has to be uncomplicated
Complexity is easy. Simplicity is hard.
I could fill an hour with measurement protocols, governance frameworks, and attribution nuance and I’d lose most of the room before the 7th slide. The goal isn’t to prove you know complex things. The goal is to make complex things clear.
Adding a new connection between two know points.
When I spoke about data governance at the Google Conference in Milan, I didn’t try to teach governance. I tried to land one insight. That for twenty years, digital analytics tried to be everything to everybody, and what most companies have to show for it is data in silos. The solution isn’t more tools. It’s just right for somebody.
That’s simple. Not simplistic. There’s depth underneath it, but it’s clear enough that people can hold it in their heads while you build on it. That’s the craft.
-
In practice: if you can’t explain your core message to someone outside your field, it’s too complicated. Cut the jargon. Remove the nested frameworks. Find the simple truth underneath the complexity. Your audience will think you’re smarter, not dumber.
Principle five: It has to change something
This is the non-negotiable. If your keynote doesn’t change something, it’s not a keynote. It’s a theatre.
There are three ways a keynote can matter. It can change minds, and have people leave thinking differently. When I argue that we’re building a “faster horse” by using AI to optimize yesterday’s methods instead of rethinking what customers actually need, that creates a pause. The next time someone pitches an AI optimization project, they ask a better question.
It can change hearts; people leave caring about something they didn’t care about before. My Superweek talk “Infants and Guns” wasn’t about cleverness. It was about responsibility. Data ethics isn’t a compliance exercise. It’s a human one.
And it can change actions. Get people do actually do something different on Monday. When I talk about the analytics circus, the action is simple: stop waiting for management to be impressed by reports they don’t understand, and start building insights that drive decisions.
The best feedback I receive is never about slides or stage presence. It’s words like “enlightening,” “engaging,” and “deeply valuable.” That’s the language of change.
-
In practice: before you agree to speak, ask yourself what you want to change. Then structure everything, story, examples, call to action, around that change. If you can’t articulate what changes, don’t take the stage.
The compound effect
When you get all five right, something interesting happens.
Your story makes people lean in. Your focus on them earns trust. Your framing makes the idea stick. Your simplicity makes it shareable. And your intent to change makes it matter.
The bar for conference keynotes is criminally low. Most speakers show up with the same deck they’ve delivered twenty times, swap in the event logo, and call it new. They deliver information that the audience could have read in a blog post. They finish to polite applause and are forgotten before lunch.
You can do better than that.
Your audience gave you their time and attention. Give them a story they’ll remember, an insight they’ll use, and a sentence that changes how they see their work.
Give them a keynote that actually matters.
But be your own judge if this is just me blabbering or if I actually walk my own talk.
