Last month I plugged a laptop into the TV in our Dubai office to watch a client cut, and I ended up staring at the YouTube home screen instead. From the sofa, about nine feet back, half the thumbnails on that grid were unreadable. Not ugly. Unreadable. Text I would have called perfectly legible on a monitor turned into grey mush, and the faces I would have sworn were the strongest part of the frame were competing with a title bar, a channel name, a progress line, and a hover state that blew the whole card up to twice its size.
That was the moment the TV app stopped being a footnote for me and became a design constraint I now check on every pack we ship.
The screen nobody designs for is the one growing fastest
Here is the awkward part of the job. Almost every thumbnail in the world is designed on a 27 inch monitor, at 100 percent zoom, in a quiet room, by someone sitting two feet away. It is then judged in a Figma frame next to five other versions, which is the most flattering possible viewing condition. Then it goes out and gets seen on a phone in a lift, or across a living room on a 55 inch panel where the viewer is holding a remote and not a mouse.
The living room case is the one most designers, me included until recently, treated as an afterthought. It should not be. Living room viewing is now a serious slice of watch time for anything long form: reviews, documentaries, tech breakdowns, finance explainers, anything people put on while they eat. If your channel has a meaningful chunk of TV watch time and you have never once looked at your own thumbnails on a television, you are designing blind for a real part of your audience.
What actually changes on a TV
The physical size goes up and the effective size goes down. That sounds like a contradiction, so let me break it down properly.
Viewing distance kills detail. A thumbnail on a phone is small but held close. A thumbnail on a TV is large but viewed from three or four times the distance. The net effect is roughly similar apparent size, except the TV version loses the one thing phone viewing preserves, which is pixel density right up against your eye. Fine strokes, thin outlines and small secondary text survive on a phone. They dissolve on a TV.
The card is not the image. On the TV app, thumbnails sit inside cards with the title stacked underneath and metadata below that. The focused card scales up and often dims or overlays. Your image is not the whole unit of attention, it is the top two thirds of a unit that also contains the title you wrote. This is the same relationship I wrote about in how thumbnails behave differently in search and the home feed, just pushed further. On TV, the title is unusually prominent relative to the image.
Colour is not your colour. Most consumer televisions ship with contrast and saturation cranked well past neutral, plus motion and sharpening processing nobody turns off. That saturated orange you picked because it pops in the feed becomes a glowing blob with a halo. Blacks crush. Subtle gradients band. You are not looking at your file, you are looking at a manufacturer's opinion of your file.
Navigation is directional. Nobody scrolls a TV feed with a thumb. They press right, right, right, and each press lands on one card at a time. That is a slower, more deliberate scan than the phone. The good news is your thumbnail gets more attention per impression. The bad news is that clutter that survives a fast thumb scroll gets fully examined and rejected here.
A phone viewer decides in half a second whether your thumbnail is worth a tap. A TV viewer sits on it for two seconds with a remote in hand. That extra time is not a gift. It is an audit.
The three failures I see most
When I started pulling our own back catalogue up on a TV, three problems came up again and again.
The first is secondary text. Almost every busy thumbnail has a hero phrase and then something smaller: a subtitle, a price, a number, a little tag in the corner. On a monitor that reads fine. On a TV at nine feet it becomes texture. Worse, it becomes texture that visually crowds the hero phrase, so you lose the small element and weaken the big one at the same time. That is the exact double penalty I get into in why your thumbnail looks cluttered.
The second is edge-hugging composition. Cards get rounded corners and the focused state can crop or scale. Anything you put within a few percent of the frame edge is living dangerously. I now treat the outer five percent on all sides as a soft margin: background only, nothing load bearing.
The third is face size. A face at 20 percent of the frame is expressive on a monitor and generic on a TV, because the features that carry the expression, eyes and mouth shape, fall below the threshold where the viewer can read them at distance. Bigger faces are not a MrBeast affectation. They are a distance-legibility decision that happens to have been popularised by one channel, which is a point I argue at length in the MrBeast effect on thumbnails.
What we changed in how we design
We test styles on our own channels before we put them in front of a client, so this is not theory I read somewhere. Here is what came out of it.
One idea, one size. If a thumbnail needs two levels of text hierarchy, it usually needs a better idea. We cut the secondary line and either fold it into the title or drop it. The title field on the TV card is doing that work for free, and it is rendered by the platform in a font that is guaranteed to be legible.
Design at thumbnail scale, then check at distance. I export at full size, then view the file scaled to about 15 percent on a monitor and physically walk back from it. Cheap, ugly, effective. If it holds at that scale from six feet, it holds on a television.
Contrast over colour. Luminance contrast survives aggressive TV picture processing. Hue contrast does not. Two colours of similar brightness can look punchy on a calibrated monitor and merge into one shape on a showroom-mode television. We now check every composition in greyscale before it ships. If the subject does not separate from the background in greyscale, the colour was doing all the work and colour is the least reliable thing you have.
Write the title as part of the frame. Because the TV card gives the title real estate, the smart move is to let the image carry the emotion and the title carry the specifics. That is why title consultation is bundled into every pack we sell rather than being an upsell. You cannot design the image well without knowing what the words next to it will say, which is the same reason I keep arguing for deciding the thumbnail before you film.
Where this does not apply
I want to be honest about the limits, because a lot of thumbnail advice gets stated as universal law and then quietly fails.
If your channel is overwhelmingly Shorts, none of this matters much. The Shorts surface has its own rules and its own failure modes, and I went through those in what changed for Shorts thumbnails. If your audience is mostly mobile, the TV constraint is a tiebreaker rather than a driver, and you should optimise for the thumb.
On CTR: I am not going to hand you a number and call it a rule. What I can tell you is that on a mid-size gaming channel we worked with, cutting the secondary text line and enlarging the face moved click-through on TV-heavy traffic noticeably while barely touching mobile, and on a finance channel around 80k subs the same change did nothing measurable at all. Same edit, two different outcomes, because the audience mix was different. Any agency quoting you a single percentage lift across all channels is selling you an average that describes nobody.
Live streams are their own case. Stream thumbnails often carry more information by necessity, schedule, guest names, episode numbers, and that information is precisely what TV distance destroys. It is worth thinking hard about hierarchy there, which is most of what we do in live stream thumbnail work.
The practical checklist
If you do nothing else, do this before you upload.
- View the file at 15 percent on your monitor and step back six feet. Can you still name the subject and read the text?
- Convert to greyscale. Does the subject still separate from the background?
- Keep the outer five percent free of anything essential.
- Cut one text element. Then look again. If nothing was lost, cut another.
- Read your title and the image together as a single unit, because that is how the TV card presents them.
- If you have a TV, cast one video to it and look at your own grid honestly.
That last one takes four minutes and it will change your mind faster than any article, including this one.
Where I have landed
The TV app has not created a new design language. It has done something more useful. It has punished the habits that were already weak and rewarded the ones that were already strong. Big clear subject, one idea, strong luminance separation, generous margins, no decorative clutter. That was always good thumbnail design. Television just removed the ability to fake it.
That is the reason every thumbnail we make is built per video rather than pulled off a template. A template is a set of decisions made for an average context, and the TV grid is where average decisions look most obviously average. You can see the range of what that produces in our work, and if you want a second pair of eyes on your own grid before you change anything, that is exactly what a thumbnail consultation is for.
One more thing, since it comes up. If you are also putting long form podcast episodes on YouTube, the TV surface matters even more, because that is where people actually watch them end to end. Our sister studio handles video podcast editing, and the same principle applies to the cut as to the thumbnail: design for the room the thing gets watched in, not the room it gets made in.
Go plug a laptop into your television. It is the cheapest research you will do this month.