mrvsubstack/yt-thumbnail-brief

YouTube Thumbnail Brief

You give it the title, what happens in the video and what assets you have (your face, a screen, a product). It hands back 2 to 3 thumbnail concepts, each a brief a designer or an image tool can execute without asking you a single question.

This skill writes words, not pixels. The brief is the deliverable.

---

The core discovery: the thumbnail finishes the title's sentence

The title and the thumbnail are read together in under a second. When they say the same thing, you've wasted half your packaging.

Wasted: title "I Quit My Job to Freelance", thumbnail text "I QUIT MY JOB". Split: title "I Quit My Job to Freelance", thumbnail shows a bank app reading £0.00 and your face mid-wince. Text: "MONTH 1".

The title makes the claim. The thumbnail adds the emotion, the stakes or the proof. Together they tell a story neither tells alone. Every concept in this skill starts by writing that split down.

---

Intake

1. The title (final or working). If there isn't one, run yt-title-optimizer first or draft both together.
2. The payoff and the tone. What does the viewer get, and how should it feel (relief, shock, curiosity, calm)?
3. Assets. Can you appear on camera? Is there a screen, product, result or location worth showing?
4. The single most visual moment in the video.
5. Your channel look. Colours, fonts or a recurring style, if any.
6. Long-form or Short. Shorts thumbnails show far less often. Brief them lightly.

Default if no face is available: build around the object or the result. Never fake a moment that isn't in the video.

---

The five non-negotiables

1. One focal point. One face or one object owns the frame.
   • Bad → you, a laptop, three logos and a graph.
   • Good → your face on the right third, one glowing graph line behind you.
2. A real, readable emotion. If there is a face, name the emotion (dread, delight, disbelief, focus) and make sure the video earns it.
   • Bad → shocked face on a calm tutorial.
   • Good → quiet satisfied grin on a "finally fixed it" video.
3. High contrast and separation. The subject pops from the background. Pick a colour your niche feed isn't already drowning in.
4. Four words or fewer, never the title's words. No punctuation, no full sentences.
   • Bad → "HOW I GREW MY NEWSLETTER FAST!"
   • Good → "0 → 5,000"
5. Passes the phone test. Shrink it to the size of a fingertip. The emotion and the text must still read. If a detail turns to mush, cut it.

---

Steps

1. Write the split. "Title says: ___. Thumbnail carries: ___."
2. Pick the lever to test. Each concept varies one big thing so the A/B result means something:
   • face vs object
   • text vs no text
   • warm vs cool dominant colour
   • before vs after
3. Fill the brief for each concept using the template below.
4. Run the five non-negotiables on each. Flag any concept at risk on the phone test.
5. Recommend the lead concept. Test & Compare judges by watch time, not raw clicks, so favour the concept that attracts the right viewer over the one that attracts the most curious one.
6. Add an image-tool prompt if the user wants to generate it. One paragraph describing subject, emotion, composition, lighting and background. Text overlays are added afterwards in a design tool, because image models still mangle lettering.

---

Brief template (one per concept)

CONCEPT [A/B/C]: [short name]
Lever being tested: [face vs object / text vs none / colour / before-after]

SPLIT
Title says: [...]
Thumbnail carries: [...]

FOCAL SUBJECT
[face + named emotion, or the single object/result]

TEXT OVERLAY
Words: [4 or fewer, no punctuation, none from the title]
Placement: [opposite the face / upper left / etc.]
Style: [weight, colour, one emphasised word]

BACKGROUND & CONTRAST
[colour, lighting, how the subject separates]

COMPOSITION
[rule-of-thirds position, gaze direction, crop]

PHONE TEST
[what must survive at fingertip size; what to drop if it doesn't]

---

Worked example

Input: Title "I Let Claude Run My Newsletter for 30 Days". Payoff: open rate held, one issue flopped badly. Face available. Channel colours: navy and orange.

Concept A: The nervous handover

• Split: title says "Claude ran it for 30 days". Thumbnail carries the risk.
• Focal subject: your face, eyebrows up, teeth clenched (nervous anticipation).
• Text: "DAY 30" in orange.
• Background: dark navy, a laptop glow lighting your face from below.
• Composition: face on the right third looking left at the text.
• Phone test: face and "DAY 30" must read. Drop any visible screen text.

Concept B: The result, no face

• Lever: object vs face.
• Focal subject: a single phone screen showing an inbox with one subject line highlighted in red.
• Text: none. The red highlight does the work.
• Phone test: the red line must read. If not, enlarge the phone to fill 70% of the frame.

Lead with A. The video is a personal experiment and the emotion sells the risk. B tests whether your audience clicks on proof over personality.

---

Output format

## Thumbnail brief: [title]

Split: title says [x] / thumbnail carries [y]

[Concept A brief]
[Concept B brief]
[Concept C brief, optional]

Lead concept: [A/B/C], because [one line]
Image-tool prompt (optional): [one paragraph, no text in the image]

---

Sibling skills

• yt-title-optimizer: write or sharpen the title first. They're a pair.
• yt-hook-scripter: the opening seconds should match the thumbnail's emotion.
• yt-content-planner: plans title and thumbnail angles across a month.
• youtube-thumbnail-design (if installed): turns this brief into an actual image.

---

Pre-flight checklist

• The split is written down for every concept.
• One focal point each. A named, earned emotion if there's a face.
• Overlay text is 4 words or fewer, no punctuation, no title words.
• Every concept passes the phone test, or the risk is flagged.
• Concepts vary one big lever, not five small ones.
• Nothing shown that doesn't happen in the video.
• A lead concept is named with a reason.