Growth

Designing thumbnails for phones first

You design on a monitor. Your audience is overwhelmingly on a phone. That gap is responsible for more failed thumbnails than any question of taste, and closing it does not require better design skills — just a different order of operations.

The size difference is larger than people assume

On a phone home feed, a thumbnail occupies most of the screen width, which sounds generous until you remember the screen is a few inches across. In the suggested list below a playing video, thumbnails shrink to a fraction of that.

Meanwhile you are designing at 1280 pixels or more on a screen where that fills a large window. You are making decisions at perhaps eight times the scale at which they will be judged.

Design small, deliver large

The fix is to invert the workflow. Set the canvas at full export resolution because that is what you deliver, but do your looking at a small preview and make every decision from it.

This feels wrong for about ten minutes and then becomes obvious, because you stop making changes that are invisible at real size — the subtle gradient, the second line of text, the fine outline — and start making the ones that matter.

What phone viewing changes about composition

Elements need more separation than a monitor suggests. Overlap that looks like layering at full size looks like a single confused shape when small. Give the face and the text their own territory with clear space between them.

Fine detail is wasted effort. Texture, small props, background interest — none of it survives, and all of it competes with the things that do. A simpler composition is not a compromise on a phone; it is the correct answer.

What happens to faces, logos and screenshots

Text gets all the attention in this discussion, but everything else in the frame is being shrunk by the same factor, and each kind of element fails in its own particular way.

Faces survive better than anything else, which is why they are used so heavily — but they survive as expression, not as identity. At around 210 pixels wide a face reads as delighted, alarmed or unimpressed, and that is the whole of what it carries. Crop close enough that the expression occupies real area, keep the eyes near the upper third, and do not count on anyone recognising whose face it is unless the head is large.

Logos usually fail. A mark designed to work on a business card is being displayed at about the size of a piece of punctuation, and what arrives is a coloured smudge holding space that could have carried something legible. If a logo has to be there, use one strong part of it — the symbol without the wordmark — at a size that would look excessive at full resolution.

Product shots need to be one object against a clean field. Three products side by side become a row of ambiguous blobs, while a single product, large and angled and lit against flat colour, still reads as a specific thing. If the product is small or dark, put it against a contrasting field rather than against another photograph.

Screenshots are the hardest case, because the thing that makes them interesting — the number, the interface, the message — is exactly the thing that disappears. Do not paste the whole screenshot. Take the one region that matters, crop to it, scale it until it occupies a substantial part of the frame, and let the rest of the interface go. A graph is usable if you keep the line and drop the axes.

Brightness and the real viewing environment

Phones are used outdoors, at low brightness to save battery, and with automatic brightness that is often wrong. A thumbnail with subtle shadow detail can be genuinely invisible in those conditions.

Higher contrast is therefore not just an attention tactic, it is a robustness tactic. Design so the composition survives being viewed dim and slightly washed out, because a good share of your impressions will be.

Test on the device, not just at the size

Shrinking a window on your monitor approximates the size but not the conditions. Send the image to your phone and look at it there, held at normal distance, ideally outdoors.

It takes a minute and it catches things no amount of on-screen preview will: colour shifts, contrast that collapses at low brightness, text that was fine on a calibrated monitor and is not fine in daylight.

A bench test that takes two minutes

Everything above collapses into one habit: look at the design at real size, on a real phone, before you upload it. Most people skip it because it feels like a formality, and it catches something roughly one time in three.

The quickest version of the test is this. Export the file, send it to yourself in any messaging app, and look at it on your phone without opening it full screen. The preview in the message list is close to feed size, which is exactly the comparison you want. Do not tap in — tapping in is not what the viewer does.

Then judge it against three questions, in this order, at arm's length. What is this an image of? What does the text say? Whose channel is it? If any of the three takes more than a second to answer, the design has a problem that no amount of admiring it at full size is going to fix.

Do it in the conditions your impressions actually happen in. Drop the brightness to about a quarter, and if you can, step outside. A composition that holds up dim, in daylight, at arm's length is a composition that holds up everywhere else.

Finally, look at it in company. Your video sits in a column with other people's, so scroll your own subscriptions feed and ask whether your design would stop you if it were not yours. Being good in isolation is not the standard — being distinguishable in a crowded column is.

Editing on the phone is a legitimate workflow

Many creators shoot, edit and upload from a phone, and building the thumbnail somewhere else breaks that. A browser-based editor that works with touch means the thumbnail is made on the same device that judges it, which is exactly the right place to make it.

This editor supports dragging, rotating and pinching with a finger for that reason, and the live preview shows the real feed sizes on the same screen the viewer will use.

Quick answers

What percentage of viewers are on mobile?

It varies enormously by niche and audience. The safe assumption is that mobile is the majority for most channels, which is why designing for the smaller case is the low-risk choice.

Should I make separate thumbnails for mobile and desktop?

You cannot — one thumbnail serves every surface. That is exactly why it has to be designed for the hardest case.

Does this apply to podcast covers too?

Even more so. Cover art in a phone app is displayed at roughly icon size, smaller than almost any thumbnail.

What is the fastest way to preview at real size?

Send the exported file to yourself in a messaging app and look at the small preview in the message list without opening it. That preview is close to feed size, and the whole check takes a few seconds.

Keep reading

← All guides