You make one good design. Then you need it as a square, a portrait story, a wide banner, a small avatar, an email header and something with a logo in the corner for a partner site. Six shapes, one idea, and about two hours of dragging things around if you do it the obvious way.
The obvious way is also the wrong way, and understanding why makes the whole job faster.
Scaling is not the same as reformatting
When a design moves from 16:9 to 9:16, almost nothing about it survives a proportional scale. The headline was three words across a wide line; now it wraps to three lines and collides with the photograph. The logo sat in the corner; the corner is now somewhere else entirely. The subject of the image was on the left third of a wide frame and is now cropped out.
Stretching preserves the pixels and destroys the design. What has to change is the arrangement: how the elements relate to each other, where the emphasis sits, how much text there is room for.
That is a layout problem, not a scaling problem, and treating it as one is why the job takes two hours.
Design the middle version first
The single most useful habit: do not start with the extreme format. Start with something near the middle, a 4:5 or a square, where you have neither generous width nor generous height and every element has to justify its space.
A design that works in the middle usually adapts outward in both directions. A design built for a wide banner tends to be a horizontal composition that has no vertical idea at all, and turning it into a story means inventing a new design rather than adapting one.
Think in anchors, not coordinates
Here is the mental shift that does most of the work. Stop thinking of an element as being at a position, and start thinking of it as being attached to something.
A logo is not at 40 by 40 pixels. It is 4 per cent of the width in from the top left corner. A headline is not 900 pixels wide, it is 80 per cent of the content width, aligned to the same left edge as everything else. A call to action is not at the bottom of the canvas, it is a fixed distance up from whatever the bottom edge turns out to be.
Once every element is described relative to an edge, a center or another element, changing the canvas becomes a recalculation rather than a redesign. This is exactly how responsive web layout works, and the thinking transfers directly to static design.
What has to change per format
Type size does not scale linearly
If you halve the canvas and halve the type, the text becomes unreadable, because reading distance did not halve. Small formats need proportionally larger type: an avatar or a mobile story wants type that would look absurdly big on a poster.
The practical rule is to set a minimum readable size for each output and let the layout accommodate it, rather than setting a size from the layout and hoping it stays readable.
Word count is part of the design
A tagline that fits on one line in a wide banner takes four lines in a narrow one, and four lines is a paragraph rather than a tagline. The honest fix is a shorter version of the text for narrow formats.
Write the short version at the same time as the long one. Coming back later to shorten your own copy after the layout has broken is a worse process than deciding up front that there are two lengths.
The crop is a decision, not a calculation
Automatic cropping to the center works for perhaps half of all photographs and fails badly for the rest, because the subject is rarely dead center. Automatic subject detection does better, and still needs checking: it will faithfully center a person while cutting off the product they are holding.
For anything that matters, decide the crop for each shape yourself, and prefer a photograph with room around the subject, which is the single easiest way to make an image adaptable.
Safe areas differ per destination
Social platforms overlay interface elements on your design: profile pictures, captions, buttons, progress bars. A story with the key message in the bottom fifth will have it covered on some platforms and not others. Keep essential content inside a generous central region and treat the edges as decoration.
Build the set, then reuse it
Once you have made the same design in six shapes a few times, the shapes stop changing. That is the moment to stop treating each one as a fresh job.
A reusable set looks like this: a defined list of output formats with their exact dimensions, one master layout with elements anchored rather than positioned, and per-format overrides for the three or four things that genuinely need to differ, usually type size, text length and crop.
From there, producing a new campaign is filling in content, not rebuilding geometry. Tools differ in how much of this they automate, and the editor described elsewhere on this site has a reformat feature that re-lays out a design across a list of target sizes rather than scaling it, which is the distinction that matters. But the thinking is the valuable part and it works with any tool, including none.
Common formats worth having ready
- 1:1 square, the universal fallback, works nearly everywhere.
- 4:5 portrait, the largest footprint most social feeds allow.
- 9:16 full screen vertical, for stories and short video covers.
- 16:9 wide, for video thumbnails, presentations and headers.
- 1.91:1 the link preview shape used by most platforms when a page is shared.
- Something small and square, tested at 64 pixels, because avatars and favicons are where designs go to die.
The check that catches most mistakes
Put every version side by side at the same physical scale and look at them together. Two things become obvious immediately: whether the set reads as one campaign or six unrelated pictures, and which version was clearly made first and grudgingly adapted.
Then shrink the whole sheet until you cannot read the text. Anything that still communicates is doing its job. Anything that turns to mush at that size will turn to mush in a feed, which is where it will actually be seen.
