A quarterly business review in a Mumbai bank, the kind with a long table and a short attention span. A vendor's account manager has twenty minutes. He uses them on nineteen slides, in the order his product team built them: company, platform, architecture, features, roadmap, case studies, pricing. At slide eleven the chief digital officer stands up, apologises, and leaves for another meeting. The account manager keeps going. Afterwards, asked what the vendor does, a junior analyst says "a lot of things".
I have given that presentation. I have also sat where the chief digital officer sat. The failure is not the content or the speaker's nerves. It is the structure, and structure is the one part of a presentation that memory research can tell you how to build.
What a listener can hold
In 2001 Nelson Cowan reviewed decades of memory experiments and concluded that the real capacity of working memory, once you remove rehearsal and chunking tricks, is about four items, not the seven that the older rule of thumb suggested. Four. That is how many things a listener can hold in mind at the same time while you are speaking.
Nineteen slides is not nineteen items; it is a stream the listener cannot hold at all, so they stop trying somewhere around slide five and wait for it to end. One idea, with three supports, is four items. That is the ceiling, and good structure lives under it.
The second finding is older and just as useful. In 1962 Bennet Murdock had people learn lists of words and recall them freely. They remembered the first few and the last few far better than the middle, the serial position effect. Primacy and recency. Applied to a talk: the opening and the close are remembered; the middle is where content goes to be forgotten. So the one thing you want remembered goes at the start and again at the end, and the middle carries the proof, which does not need to be recalled word for word, only believed.
Add Willis and Todorov's Princeton finding that judgements of the speaker form in about a tenth of a second, and Kahneman's point that the fast system decides and the slow system justifies, and you have the whole brief: say the one thing first, prove it in no more than three moves, say it again, and get out.
The shape
Here is the structure, for a twenty-minute talk, a five-minute update or a sixty-second pitch. Use it as a presentation outline, or as the answer to how to organize a talk of any length. The proportions change; the shape does not.
The one thing, first. Not an agenda. Not "thank you for the opportunity". The sentence you want the chief digital officer to repeat in the lift. "Your onboarding takes nine weeks; we can make it five, and here is how we know." If you cannot write this sentence, you are not ready to build slides. Most presentations fail at this line because the speaker has not decided what it is.
Why it matters to them, briefly. One minute. The cost of the current state, in their numbers if you have them. This is where the room decides whether to lean in. Neil Rackham's research on 35,000 sales calls found that the implication of a problem, what it costs, moved buyers more than any description of a solution. The same applies to a room.
Three supports, and only three. Each one is a claim, a piece of evidence, and what it means. A case, a number, a demonstration. Three is under Cowan's ceiling and leaves room for the one thing. If you have five supports, the two weakest are costing you the three strongest. Cut them.
The one thing, again, with the ask. Recency. Say the sentence again, almost word for word, then say what you want to happen next and who does it. "Five weeks, not nine. The next step is a two-hour workshop with your onboarding team; I can run it on the 14th."
Questions last, and planned for. Do not say "any questions?" and wait. Say "the question I usually get at this point is X" and answer it. Then open the floor. You have shown you expected the hard question, and the room is more likely to ask the ones you can answer.
Surviving interruption
The structure above has a property that matters in Indian boardrooms more than anywhere I have presented: it survives being cut off. If the chief digital officer leaves at minute eight, she leaves with the one thing, the cost, and at least one support. If she leaves at minute eight of the nineteen-slide version, she leaves with the company history.
Build for interruption deliberately. Put the sentence first. Put the strongest support first. Assume you will get half the time you were promised, because you often will.
The slide test
Take your current deck. For each slide, ask which of the five parts it belongs to. Slides that belong to none of them, usually the company introduction, the architecture diagram and the roadmap, are not supporting the one thing. Move them to an appendix. The deck will get shorter and the talk will get longer, because you will finally have time to pause.
Then read the first sentence of the talk aloud to someone who was not in the room. If they can say back what you want, the structure is working. If they say "a lot of things", start again at the first line.
The full method, including how to write the one thing when there are three stakeholders who want three different things, is Structure and Scripting. For why decks built in product-team order fail, read your pitch deck's story is backwards. The free Pitch Scorecard scores a sixty-second version against this shape in two minutes, and corporate teams get the method through Communication and Public Speaking training. On LinkedIn, comment PITCH on any of my posts and the pitch and demo kit comes to your inbox.
Frequently asked questions
What is the best structure for a presentation?
One central sentence first, a brief statement of why it matters to the audience, no more than three supporting points each with evidence, the central sentence repeated with a clear ask, and a planned first question. The shape works for a sixty-second pitch and a twenty-minute talk alike.
How many points should a presentation have?
Three supporting points under one main idea. Memory research by Nelson Cowan puts working memory capacity at about four items, so one idea plus three supports is the most a listener can hold while you speak.
Should you start a presentation with an agenda?
No. Start with the one sentence you want remembered. The serial position effect, first described by Murdock in 1962, shows that the opening and the close are remembered best; spending the opening on an agenda wastes the most memorable slot.
How do you structure a 20-minute presentation?
Roughly: one minute for the central sentence, two for why it matters, twelve for three supports, two for the repeated sentence and the ask, and three for a planned question and discussion. Build it so that if you are stopped at minute eight the audience already has the main point.
How do you make a presentation memorable?
Say the main point first and last, keep supports to three, use the audience's own numbers for the cost of the problem, and pause after key sentences. Memorability is a structure decision more than a delivery one.
References
- Cowan, N. (2001). The magical number 4 in short-term memory: a reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24(1), 87-114. doi:10.1017/S0140525X01003922
- Murdock, B. B. (1962). The serial position effect of free recall. Journal of Experimental Psychology, 64(5), 482-488. doi:10.1037/h0045106
- Willis, J. and Todorov, A. (2006). First impressions: making up your mind after a 100-ms exposure to a face. Psychological Science, 17(7), 592-598. doi:10.1111/j.1467-9280.2006.01750.x
- Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
- Rackham, N. (1988). SPIN Selling. McGraw-Hill.
Tell us what training you need
We run this analysis against your live deals, in your office, with your team. You keep the written gap report either way.
Tell us what you need