The Thumbnail Pack

The pre-publish CTR checklist

Run this before you hit publish. If a box won't tick, fix it first.

• Readable at a glance on a phone (shrink it to ~120px wide, or step back two metres).
• One clear subject and one clear idea. Nothing fighting for attention.
• Text is 3-4 words max, huge, and high-contrast.
• The text adds something new. It does not repeat the title.
• If there's a face, the emotion is genuine and legible at small size.
• The subject separates from the background (light-on-dark, or a rim of contrast).
• It looks like your channel. A returning viewer would recognise it as yours.
• It matches the video. The thumbnail's promise is one the video actually pays off.
• You compared it against your other recent thumbnails in a mock feed.
• Exported 1280x720, 16:9, under 2MB.
• You have a second version ready to test if this one underperforms.

The 3-second test

The single fastest way to catch a weak thumbnail.

1. Shrink it. View the thumbnail at real feed size, on a phone, next to other videos.
2. Give a stranger 3 seconds. Can they tell what it's about and why they'd click?
3. If not, subtract. Enlarge the subject, cut a word, raise the contrast, remove a distraction. Clarity beats detail every time.
4. Repeat until it passes. A thumbnail that survives the 3-second test at phone size is a thumbnail that works.

The principles, on one line each

• The thumbnail decides the click, and the click decides the reach.
• Design at the size it's seen, not the size you're working at.
• One subject, one idea. Detail disappears at thumbnail scale.
• Faces work when the emotion is real and readable, not as decoration.
• Contrast makes the subject pop. Separate it from its background.
• Words add clicks only when they add an idea the title doesn't.
• Your first thumbnail is a guess. Test it, then trust the data.