How many slides for a 20-minute presentation?
The honest answer is that slide count is the wrong variable. The one that has been measured is how much a person can hold at once, and it is smaller than your deck assumes.
There is no correct number, and slide count is the wrong thing to count. A twenty-minute talk can work with three slides or with forty. What cannot vary is how many separate ideas a listener can hold at once, and the research puts that at about four.
So count ideas, not slides. Three or four for twenty minutes, each supported rather than supplemented. And do not put your script on screen — reading aloud what the audience is already reading measurably reduces how much they take in.
The question gets asked because it feels answerable, unlike "is this any good". It has also attracted a famous rule: ten slides, twenty minutes, thirty-point font, from Guy Kawasaki's experience of venture capital pitches.
It is decent advice with a weak foundation. It was never a study — it is one experienced person's heuristic for one specific situation. The third of the three parts is the one that genuinely holds, and for a reason worth understanding.
The number that has actually been measured
Most people have heard that working memory holds "seven plus or minus two", from George Miller's 1956 paper. That figure has not held up well. Nelson Cowan's 2001 review of the evidence put the real capacity, once you prevent people from rehearsing and grouping, closer to four.
This is the constraint that matters, and it does not care how you distribute the content across slides. Four ideas spread over twenty slides is fine. Eight ideas compressed onto three slides is not, and it is the more common mistake.
Why reading your slides is worse than it feels
Richard Mayer's group has spent decades testing multimedia instruction, and one of the more counterintuitive results is the redundancy effect: adding on-screen text that duplicates what the speaker is saying reduces how much people learn, compared with narration and a graphic alone.
It seems as though it should help — two channels, same message. It does not, because written and spoken words compete for the same processing. The listener ends up reading ahead, hearing you behind, and reconciling the two instead of following the argument.
Which recasts the familiar complaint. "Do not read your slides" is usually delivered as a point of style, as though it were merely dull. It is a design problem: if the words are on the screen, saying them costs you attention rather than reinforcing them.
The practical version: a slide should carry either the words or your voice, not both. A picture, a chart, or a single sentence you do not read aloud — and then say something that adds to it.
So how many should you have?
Whatever it takes, with these constraints:
- One idea per slide. The moment a slide contains two, the audience chooses, and it will not be the one you wanted.
- Three or four ideas for twenty minutes. Not three or four sections — three or four things you would be content for someone to repeat afterwards.
- Big type. The thirty-point rule is the durable part of Kawasaki's heuristic, and not because of eyesight: a font size that large physically prevents you from putting your script on the slide. It is a constraint disguised as a formatting tip.
- Some blank ones. When you want the room looking at you rather than at the screen, give them nothing else to look at. This is the most underused slide in any deck.
A better question than "how many"
Go through the deck and ask of each slide: if this were removed, what would the audience fail to understand?
Most slides have no answer. They exist because the work was done and it felt wrong to leave it out, or because a section needed a heading. Those are reasons for an appendix, not for the room.
That test typically removes a third of a deck in ten minutes, and it removes it from the middle — which, given that the middle is where recall is weakest anyway, costs you almost nothing.
One honest limitation. Mayer's studies are mostly on instructional material with students learning a process, and Cowan's number comes from laboratory memory tasks. Neither is a board meeting. The mechanisms are general enough to act on, but anyone quoting a precise slide count as scientifically established has gone further than the evidence does.
The genuine answer, then, is unsatisfying and true: nobody has ever left a presentation thinking about how many slides there were. They leave with one thing or with nothing. Build for that, and the count settles itself.
Sources
- 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.
- Miller, G. A. (1956). The magical number seven, plus or minus two. Psychological Review, 63(2), 81–97. The original, and the source of the figure Cowan revised.
- Mayer, R. E., Heiser, J., & Lonn, S. (2001). Cognitive constraints on multimedia learning: when presenting more material results in less understanding. Journal of Educational Psychology, 93(1), 187–198.
- Sweller, J. (1988). Cognitive load during problem solving: effects on learning. Cognitive Science, 12(2), 257–285.