Plain Speaking wordmark: two horizontal rules above an open bracketPlain Speaking

Plain notes on how people speak, listen, read aloud and write.

N-11 · Speaking to a room · written 2026-07-30

Presenting without slides

Written for anyone who has to stand up and explain something.

What the room is actually doing

An audience is not evaluating you. It is trying, with a fair amount of effort and a limited budget of attention, to work out what your talk is for and whether it is going anywhere. Almost every failure of a talk is a failure to answer those two questions early enough.

The room is also thinking about other things, and will drop out and return several times. This is normal and cannot be prevented. What it means practically is that the structure has to be recoverable: someone who returns after ninety seconds away should be able to locate themselves.

Deciding what the talk is for

Before anything else, finish this sentence: after this talk, the room should be able to do or decide one specific thing. If you cannot finish it, the talk has no shape yet and no amount of preparation will supply one. Most rambling talks are not badly delivered; they were never given a destination.

Once the destination exists, most of the material selects itself. Anything that does not move the room towards that one thing is interesting rather than necessary, and interesting material is what you offer in the questions afterwards.

A shape that holds

The most durable structure for a short talk is also the least fashionable: say what you are going to argue, argue it in three moves, then say what follows. It is unfashionable because it is not suspenseful, and suspense is exactly the wrong instinct. You are not withholding a twist; you are trying to be followed.

Three moves is not a rule about the number three. It is an observation that most rooms will hold two or three things and will not hold seven, and that a speaker with seven points has usually not yet decided which three matter.

Talking without the deck

Slides are frequently doing the speaker's structural work, which is why a talk collapses when the projector fails. The test worth running is whether the talk survives without them: if you cannot give it from a single index card of headings, the deck is holding it up.

Notes should be headings, not sentences. Written-out sentences pull a speaker into reading, and reading aloud from a script is audibly different from talking, in a way that every audience detects and no audience enjoys. Headings force you to construct the sentence in the room, which is what makes a talk sound like speech.

The twenty minutes before

Rehearse out loud, standing, at least once end to end. Rehearsing in your head runs at roughly twice the speed of speech and skips every place where you do not actually know what you think, which are precisely the places that will fail in front of people.

In the last few minutes, stop revising. Late changes are almost always worse than the version you rehearsed, because they have never been spoken. Slow your first two sentences deliberately: nerves show up as speed, and beginning slowly is the one adjustment that reliably corrects everything that follows.

Questions

Questions are not an appendix. They are the part of the event where you find out what the room actually heard, and a speaker who treats them as an obligation misses the only feedback available. Repeat the question before answering it, both so everyone hears it and so you have a moment to understand what is really being asked.

When you do not know, say so plainly and stop. An audience forgives not knowing something immediately and never quite forgives the improvised paragraph that was produced instead.