The TED format is one idea, not twenty minutes of them

The short answer

The TED format is a personal story, then the insight that story produced, then what you want done about it, all of it in service of one idea. The shape is easy. The constraint is the hard part, and it is the whole format: everything that does not serve the one idea comes out, however good it is.

That is why most TED-style talks fail. Not on delivery. On the count.

The test that decides whether you are ready to write

Before a single slide, write this sentence down:

If they remember one thing from this in a year, it is ______.

If you cannot fill that blank in a sentence a colleague would repeat verbatim, you are not ready to build the talk yet. You have a topic, which is a different thing and a much less useful one.

I ask clients to do this on the call, out loud, and it is uncomfortable. Most people arrive with a subject. “Our approach to hiring”. “What I learned scaling”. A subject is not an idea. An idea has a claim in it. Somebody could disagree with it. “Hiring for potential costs more in year one and less in year three” is an idea. “Our approach to hiring” is a folder.

Once that sentence exists, the rest of the talk becomes an editing job rather than a writing job, which is a far easier task.

The three beats

The story that earns the idea

Open with something that happened to you. Not a general observation, not a statistic, and not a joke about the venue’s wifi. A specific occasion, with a place and a moment in it.

The reason is not warmth. It is standing. An audience will accept an argument from someone whose experience they have heard, and will resist the identical argument from someone who has only asserted it. The story is how you earn the right to make the claim that follows, and it should be chosen for that job: the episode that produced the idea, not your best anecdote.

Two to four minutes. If it runs longer, it stops being the on-ramp and starts being the talk.

The insight

Now say the thing. Plainly, in one sentence, and early enough that the room does not have to work out where you are going.

This is where people hedge. They arrive at the insight and then soften it into something nobody could argue with, which is also something nobody could remember. If your idea has no edge, the talk has no reason to exist. Say the version you would defend in a hostile Q&A, then spend the middle of the talk defending it. Evidence, a second example, the obvious objection answered before someone raises it.

The ask

End by naming what you want. A change of mind is a legitimate ask; so is a change of behaviour. What is not legitimate is trailing off into “so, something to think about”, which is the most common ending I hear and it wastes the only moment in the talk everybody is definitely listening to.

Fifteen per cent of your time. Most people give it thirty seconds after spending three minutes arriving.

Where it goes wrong

Three ideas because all three are good. This is the big one, and it does not feel like a mistake while you are writing. It feels like generosity. It is the single reliable way to be remembered for nothing.

A story with no insight attached. Charming, and the room leaves with a nice anecdote and no argument. The story has to have produced something.

An insight with no story under it. Now you are asserting. The room has no reason to grant it, and the talk feels like a slide deck read aloud.

Saving the idea for the end. Mystery is not a structure. If the room does not know what you are arguing by roughly a third of the way in, they are not following you, they are waiting.

Delivery notes specific to this format

A one-idea talk lives or dies on emphasis, because the room has to be able to tell which sentence is the sentence. That is a job for pause and pitch: a beat of silence before your core sentence and after it, and enough range that it does not land in the same tone as the sentence before it.

If you deliver your one idea at the same pace and pitch as your setup, the room will hear it and not notice it. I have watched people write an excellent talk and then bury the line it was built around, because they were nervous and speeding up, and the line went past at the same volume as everything else.

Where this sits

TED-style is one of five structures worth knowing, and it is the one to reach for when the room has to remember something. If they instead have to decide, approve or fund, use problem-agitate-solution. If they will interrupt, use a numbered framework, which is the only one of the five that survives being cut off.

The stories inside it have shapes of their own. STAR, CAR, nested loops and callbacks are worth knowing separately.

There is a lesson on opening in the free public speaking course, and building a one-idea talk from a blank page is most of the first fortnight of the public speaking course.

Common questions

What is the structure of a TED talk?

A personal story that earns your right to raise the idea, the insight that story produced, and what you want the room to do with it. One idea throughout. The story comes first because an audience accepts an argument from someone whose experience they have already heard.

How long should a TED-style talk be?

Eighteen minutes is the TED limit and it is a good ceiling for any one-idea talk. Below eight minutes there is not room for a story that earns anything. Above twenty, a single idea starts repeating itself and the room notices.

Can I use the TED format for a business presentation?

For a keynote or an all-hands, yes. For anything the room has to decide or fund, no. Use problem-agitate-solution instead. TED-style is built to be remembered, not to get a budget approved, and the two need different shapes.

What if I have more than one idea worth sharing?

Then it is more than one talk. A talk carrying three ideas is remembered as none of them. Pick the one you would want quoted back to you in a year, keep it, and put the others in a post, a follow-up, or a conversation afterwards.