Your poster is not ugly, it is unranked

The wrong diagnosis

Someone makes their first event poster. The band name, the date, the venue, the ticket price and a QR code all sit there at roughly the same size, in roughly the same weight, spread evenly across the page. It looks wrong. They are told they do not have an eye for it, or that they lack taste, and they believe it.

That diagnosis is almost always false. The fault is not taste. Every element on that poster is claiming to be the most important thing, so none of them is. The reader's eye arrives, finds five equal candidates, and gives up. Nothing about colour, typeface or software fixes that, which is why the second version usually looks the same as the first with a gradient on it.

The eye does not scan, it jumps

Human vision is not a camera panning across a page. The eye moves in jerks — saccades — landing three or four times a second, and during the jump itself you are effectively blind. Sharp detail only exists in a tiny central patch, about the width of your thumbnail at arm's length. Everything else is a blur with colour and shape in it.

So a viewer does not "take in" your layout. They land on three to five spots, in some order, and build an impression from those. Design is the business of choosing those spots and that order.

What pulls a landing, roughly strongest first:

• Size difference. The blunt instrument. Reliable and boring.
• Contrast against the background. A dark thing on a pale field, or the reverse.
• Isolation. The underrated one. A small item alone in a wide empty area beats a large item in a crowd.
• Position, relative to the direction the reader starts from.
• Saturation, then weight, then shape oddity — one round thing among rectangles.

Notice that four of those six are relative. They exist only by comparison with the rest of the page. That is the whole idea.

Reading direction is a weak cue, and that is good news

A reader of English or Hindi starts at the top left. A reader of Urdu, Arabic or Hebrew starts at the top right. Position matters, but it is a weaker cue than size, contrast and isolation — which means a layout built on those three works for both audiences without redesign. Build the hierarchy out of the strong cues and position becomes a refinement rather than a dependency.

One more thing worth unlearning: the F-pattern. It is a real observation from eye-tracking of text-heavy pages with no visual structure, not a law of vision. On a page with a clear hierarchy, people look at the big thing first. That is the point.

Name first, second, third out loud

Before you open any software, write three lines on paper:

1. 1What must a stranger see in one second?
2. 2What should they see next?
3. 3What is everything else?

On a concert poster: band name / date and venue / all the rest. On a shop sign: what you sell / where you are / opening hours. On a CV: your name and what you do / your last two jobs / everything else.

If you cannot name them, the design cannot show them. That is the actual reason layouts drift: when you have not decided, every element gets nudged up a little each time you look at it.

That is the second failure mode, and it deserves a name: emphasis inflation. Each round of feedback adds emphasis to one more item. The phone number goes bold. Then the price goes bold and red. Then the date goes bold, red and bigger, so the band name goes bigger again. After four rounds everything is loud, which is identical to everything being quiet.

The rule that stops it: to promote one thing, demote another. Emphasis is zero-sum. If your fix is only ever additive, you are not designing a hierarchy, you are inflating one.

Make the difference obvious or do not make it

An 18pt heading over 16pt body text does not read as a rank. It reads as a mistake. A difference that a viewer cannot detect at a glance costs you the same effort as a difference they can, and buys you nothing.

Use steps you can see. A workable set for a page of text: body 16, subhead 20, heading 32, display 48. Each step is roughly 1.25 to 1.5 times the last. Rule of thumb: if you have to compare two items side by side to tell which is bigger, they are the same size as far as the reader is concerned.

The same applies to weight, to colour, to spacing. Half a difference is worse than none, because it looks accidental.

Four free ways to test it, one of which is your eyelids

• Squint until detail disappears. What survives is your hierarchy. If nothing survives, there isn't one.
• Greyscale it. In Photopea, GIMP or Krita, pull saturation to −100 (Image, then Adjustments or Colors, then Hue-Saturation). If the structure collapses, you built hierarchy out of colour, and colour is the least reliable channel you have.
• Shrink it. Zoom to 25 per cent, or take a screenshot on your phone and pinch it down to thumbnail size. The squint test by other means, and it needs no software.
• Three seconds. Show it to a person for three seconds, take it away, ask what they remember. More honest than any opinion given with the page still in front of them.

Today

Take something you have already made. Write down what you want seen first, second and third. Then screenshot it, hand it to someone for three seconds, and ask them what they saw. Where their answer and your list disagree, you have found the work.