A/B tests start after you publish. This one runs before: drop your thumbnail, name your topic, and see it in the feed next to the real top videos — then fix it while it still costs nothing.
Need the size rules? See the thumbnail size checker, or browse thumbnail text from high-performing videos.
A thumbnail never appears alone. It sits in a grid of other videos on the same topic, at a fraction of the size you designed it at. Testing it against those neighbours before you publish catches the problems a full-screen preview hides.
If every video on your topic uses a dark background with yellow text, yours doing the same makes it one of many. Look at the grid for a second: if your eye doesn't land on yours, change the one thing that makes it look like the rest — background colour, brightness or layout.
Dark studio shot with big yellow text, like the six videos around it
Same idea on a bright, plain background, so it's the one light tile in a dark row
Most views come from phones, where a thumbnail is a few centimetres wide. Text that is readable on your monitor can turn into texture. Three or four large words survive; a sentence doesn't.
"I tried every budget meal prep hack for 30 days" written across the image
"$2 MEALS?" — the title says the rest
Viewers read both together. Repeating the title in the image wastes the space; the thumbnail should show the result, the stakes or the question the title sets up.
Title "I tested 5 cheap mics" and thumbnail text "5 CHEAP MICS"
Same title, thumbnail showing the $12 mic next to a $200 one with "SAME?"
Rewrites are illustrative examples, not measured results.
Yes. The feed test and the measurements are free with no sign-up. An optional AI review of the design uses credits; a free account includes 15.
No. The image is read and measured in your browser and never sent to our servers.
They are real YouTube results for your topic from the YouTube Data API, ordered by how far each beat its own channel's median views. View counts and dates are the public numbers at the time they were fetched.
No. It shows how your thumbnail looks and measures how it differs from the ones around it. Only real viewers can tell you the click-through rate — once the video is live, use YouTube's own thumbnail testing.
A/B testing tools swap thumbnails on a published video and wait for views. This tests before you publish, against the specific videos yours will compete with, so you can fix it before the first impression.