Stop making one ad and resizing it into five

Turning a square ad into 9:16 either cuts 43.8% of its width or leaves 43.8% of the frame empty. On Stories and Reels the interface then covers the top 14% and bottom 35%, leaving about 51% usable. Compose each placement for its own frame instead; if you must crop, start from 9:16.

Here is how most teams still make a campaign. Someone designs the ad — usually square, usually because that's what the template was. Then it gets resized: a 4:5 for feed, a 9:16 for Stories, maybe a 16:9. The resize is treated as production work. Export, crop, nudge the logo, done.

That workflow is inherited from display banners, where a 300×250 and a 728×90 genuinely were the same ad at different scales. It doesn't survive contact with the placements people actually buy now.

What happens when you crop one ad into every placement?

A composition is a set of relationships: where the eye lands, how much air sits above the product, the distance between headline and edge. Crop the frame and every one of those relationships changes, but nothing inside the frame knows it did.

Going from 1:1 to 9:16, you have exactly two options, and people talk about them as though they were one thing.

Fill the frame: scale the square until it covers 9:16, then cut the sides. That discards 43.8% of the width. Fit the frame: keep the whole square and pad above and below. That leaves 43.8% of the frame empty. There is no third option, and — pleasingly — the arithmetic is identical either way.

Fill, and the product that was comfortably centred is cropped at both edges. Fit, and it floats in the middle of a column with dead space above and below. Neither is the composition anyone signed off on.

Cropping preserves the pixels and destroys the composition. Those are not the same thing, and only one of them is what you shipped.

Then the safe zones finish the job. On Stories and Reels the interface sits on top of your ad, and Meta's guidance is to keep the top 14% and bottom 35% clear of anything that matters. Do that arithmetic and only 51% of the frame is actually yours — a band across the middle, a little over half the canvas.

A crop has no idea that band exists. So the copy lands under the CTA button, and nobody notices until it's live — because the exported file looked fine in Finder.

The tells

You can spot a resized ad without being told. Every one of these is a crop artifact:

• Logo drifting to a different corner between placements.
• Type that's noticeably larger in one format than the others, because the crop scaled it up with everything else.
• A product sitting dead-centre in vertical with dead space above and below.
• Text that reads fine in feed and collides with the caption in Stories.
• A background that's clearly one photo, hard-cropped, in three different ways.

None of these are fatal on their own. Collectively they read as cheap, which is the one thing an ad cannot afford to look like when it's competing with organic content in the same scroll.

Why the habit persists

Because for a long time the alternative was worse.

Designing each placement from scratch meant three or four times the work for one campaign. When a designer's time is the bottleneck, cropping is completely rational — a slightly compromised 9:16 that ships this week beats a perfect one that ships never. Nobody cropping a square into a story thinks it's the ideal output. It's triage, and triage is a reasonable response to a real constraint.

Automated resizing tools made the triage faster, but they didn't change the underlying trade. They crop more consistently than a person does. They still crop.

What replaces it

Render each placement natively. Not one master and four crops — four renders, each composed for its own frame, sharing the same brand, product, and message.

The thing that made this impractical was cost, and that's the part that changed. When composing for a placement stops taking an afternoon, the reason to crop mostly evaporates. You're no longer choosing between "perfect" and "shipped".

This is the bet Terrads is built on, so treat the recommendation with the appropriate suspicion. But you can test the claim without us: take your best-performing square ad, design a 9:16 from scratch with the safe zones in mind, and run them against each other. I'd expect a wider gap than the effort difference suggests. If your account says otherwise, that's worth knowing too — which is rather the point of testing it instead of taking my word for it.

If you're going to crop anyway

Sometimes you are. Deadline, one asset, no time. In that case:

1. Start from 9:16, not 1:1. Cropping down and out of the tallest frame loses less than stretching up from a square.
2. Keep the subject off dead-centre in the master, so a vertical crop doesn't strand it.
3. Never bake copy into the master. Add text per placement, after the crop, with the safe zones visible.
4. Check every placement in the preview, not the exported file.

That gets you most of the way. It just costs more attention per campaign than composing for each frame in the first place — which is the actual argument.

Sources

• Meta Ads Guide: image ad specs on Instagram Stories — "roughly 14% of the top, 35% of the bottom, and 6% on each side" (accessed 2026-09-15)
• Meta Ads Guide: image ad specs on Instagram Reels — the same guidance for Reels (accessed 2026-09-15)

---

Related: the four Meta ad sizes that matter, including the resolution nearly every spec sheet gets wrong — and the sixteen layouts almost every performance ad uses, which is the decision to make before you pick a size at all.