Where the eye lands on a 16:9 slide

The eye lands on the largest contrast difference in the frame, rides the strongest edge away from it, and then leaves for the speaker's face. The whole sequence takes about two seconds. A 16:9 slide therefore needs one object that is unambiguously the brightest or darkest thing on it, placed above the bottom tenth of the frame where a real room can still see it, and short enough to be finished before the audience looks back at the person talking.

A page is read at arm’s length by someone who controls the pace. A poster is read standing still, with nobody talking over it. A slide is neither: wide, far away, up for about as long as one paragraph of speech, and competing with a human being who is the actual medium. Almost every bad slide is a page or a poster that got projected.

Why is the F-pattern the wrong model for a projected slide?

The F-pattern comes from eye-tracking of people reading web pages: the reader enters top-left, sweeps right, drops, sweeps a shorter distance, then runs down the left margin hunting for a reason to stop. It arrives in slide design as the instruction to put the important thing top-left. That borrowing is a mistake, because every condition producing the F is missing in a room.

Start with the geometry. Sharp vision covers about two degrees of the visual field, roughly a thumbnail at arm’s length, and a laptop at half a metre lets you move that window across text several times a second. That is what a scan is made of. A 2.5-metre screen from six rows back spans twenty or thirty degrees, so most of the slide sits permanently in peripheral vision — which does not read. It registers brightness and edges, and tells the fovea where to jump.

Then the motive. A web reader scans because the page is the only source of information. An audience member already has a source — the person talking — so the slide is a glance taken to check the picture against the sentence they just heard. Two or three fixations, then back to the face. And a slide that does get scanned has cost you the room: for the eight seconds someone spends reading your third bullet, they are not listening to you.

What does the eye actually find first on a slide?

The largest contrast difference, and it is not close. The visual system computes luminance boundaries long before it resolves a letter, so the first fixation is placed by brightness, not by meaning and not by position. A small word in near-black on a pale wash gets found before a large word in mid-grey sitting obediently top-left, because the mid-grey word is not different enough to earn a jump.

The second thing is the strongest edge. Once the eye has landed it travels along whatever continuous boundary leaves the landing site: the join between a photograph and a flat panel, a rule under a kicker, the height at which a gradient stops holding its colour. An edge is something you aim, which is why a background is a composition decision.

Three tests, none of which need equipment:

• Blur the slide until the letterforms dissolve. Whatever is still the lightest or darkest shape gets read first. If that is your logo, your logo is your headline.
• Halve the screen brightness. Anything that vanishes was never going to survive a lamp two thousand hours old.
• Look at the thumbnail in the rail. At rail width a slide reads as the shape of a layout rather than as text, which is roughly what the back row gets.

What is the top-left corner for, and what is the lower third for?

They do different jobs, and confusing them is the most common structural mistake in a deck. The upper-left quadrant is the register: it answers where am I, and it is what the eye checks on the way back from the speaker’s face. It wants the smallest, most stable text on the slide — a section name, a running label, a date. It does not need contrast, because nobody is hunting for it; it needs to sit in the same place on every slide. A kicker slot with a 60-character budget is that job, sized.

The lower third is the payload band, for two physical reasons. It is the last stop of any edge running down the frame, so the eye is already heading there. And in most rooms it is the part of the screen nearest the speaker, making the trip from the sentence to the face saying it the shortest available. That is why strong cover layouts stack a quiet kicker and a loud headline against the base of the frame rather than centring them vertically. Swap the two jobs and both fail: the register becomes furniture competing with the point, and the point gets found third, after the photograph and the chart.

Should a headline be centred or aligned left?

Centring is a decision about pace rather than taste, and there is a mechanism behind it. Left-aligned text gives every line the same starting x, so the eye returns to a known point and the next line begins immediately. Centred text moves the start of every line, so the eye has to locate the beginning again: a search, then a jump, then the read. That costs a fraction of a second, every line.

Which makes centring a brake, and brakes have uses. Centre when slowing the room is the point: a single statement of one or two lines, the slide you want landed rather than skimmed. The statement layout carries a 60-character budget and the section divider carries 34, and those numbers exist because a centred line stops being a brake and becomes an obstacle around the third line. Never centre a paragraph or a list — a 260-character slot set centred has a ragged left edge, and the reader pays that cost on every line of it.

Centre a line to make the room stop. Align it left to make the room continue. Choosing between them at random is choosing not to control the pace.

Why is the bottom of a 16:9 frame dangerous in a real room?

Because you do not own it. Four things eat the base of a projected frame, and all four are common enough to design around rather than hope against. Front-row heads occlude a screen mounted at table height. A display on a credenza puts the last band behind a laptop lid. Keystone correction and overscan crop two to five per cent off every edge and tell nobody. And on a video call the participant filmstrip sits exactly there.

So treat the bottom ten per cent as sacrificial. On a 1920 by 1080 deck that is the last 108 pixels — fine for a page number, a footer rule or a repeated logo, things whose disappearance costs nothing. Anything that has to be read stops above it. This is not an argument against the lower third: the lower third of 1080 begins at 720, and the usable part runs to about 972. The payload band and the dead band are neighbours, which is why so many decks push a headline into the dead one.

Exporting to PDF at vector quality rescues none of this. The PDF is faithful, which is the point of it: it reproduces your bottom-edge headline perfectly, and a projector crops it anyway. Only the layout can fix a layout.

Why is one idea per slide a layout constraint rather than a piety?

Because a slide has one contrast peak and the peak does not divide. Put two ideas on a slide and either you made one louder, in which case the other is not read second — it is not read at all — or you made them equal, in which case the eye has no reason to jump anywhere and returns to the speaker having resolved neither. A second point does not halve the attention the first gets. It usually deletes both. The rule is not about respecting the audience. It is a claim about how many objects two or three fixations resolve at that distance: one headline, one supporting shape.

The character budget is how that gets enforced without anyone having to remember it. Every slot carries a maxChars — the length the box was drawn for: 70 characters for a cover headline, 44 for a two-line title, 34 for a section name, 180 for a pull quote, 260 to 320 for a body paragraph. Those are not preferences. They come off 94 layout archetypes taken from decks that were actually presented, so each budget is text that already survived a real room.

Writing past a budget is what breaks a layout, and it breaks it in the two ways described above. Either the type shrinks to fit, so the headline stops being the highest-contrast object and the eye lands on the chart instead. Or the block grows downward into the bottom tenth, where the front row cannot see it. The budget is not a limit on your thinking. It is the geometry of the room written as a number, in the one place you will actually meet it: the box you are typing into.