Thumbnail Psychology

Thumbnail Psychology — Design Brief Generator

Create a thumbnail design brief for: $ARGUMENTS

The Science

Viewers decide in 0.05 seconds. You have 1.8 seconds during scroll. Every choice must maximize that window.

Step 1: Analyze the Title

Read the title. The thumbnail must complement the title, NOT repeat it. Each carries unique information. Together they tell a story neither tells alone.

• If title has a number → thumbnail shows the visual result
• If title has a question → thumbnail shows the tension
• If title has a transformation → thumbnail shows before OR after (not both)

Step 2: Generate 3-5 Thumbnail Concepts

For each concept, specify:

Visual Elements (max 3)

• Subject: Who/what is the focal point?
• Visual Hook: What single element creates intrigue?
• Message: What text (4 words MAX) appears?

The 3-Part Thumbnail Test

Strong thumbnails usually contain:

• One face: clear emotion, readable at phone size.
• One object: the money, tool, finished result, threat, or absurd visual proof.
• One question: the image makes the viewer wonder what happened or what happens next.

Run the zoom-out test: shrink the thumbnail to mobile home-page size. If the focal point, emotion, and question are not obvious, simplify it.

Research-Backed Rules

Faces:

• Faces get +35-50% higher CTR
• Close-up with eye contact performs best
• Emotion: shock, joy, fear, confusion
• Closed-mouth smiles beat exaggerated expressions (MrBeast A/B tested, 100% win rate)
• Changed hundreds of old thumbnails based on this finding

Color:

• Design for dark mode (60-70% of viewers)
• High contrast: +20-40% CTR improvement
• Red/yellow highlights: up to +20% CTR
• Bright colors (red, orange, yellow) grab attention on dark backgrounds

Text:

• 4 words or fewer: +30% CTR vs text-heavy
• Teasers > explanations
• Questions > descriptions
• Large, bold, readable at phone size

Composition:

• Maximum 3 visual elements total
• One clear focal point
• Human figure + item together = highest impact (Derral Eves finding)
• Negative space directs attention
• If the title already says the claim, the thumbnail should show proof, consequence, scale, or emotion instead of repeating the claim.

Step 3: A/B Test Plan

• Design 3 variations of top concept
• Variation A: primary concept
• Variation B: different emotion/expression
• Variation C: different text or no text
• Have backups ready for rescue swaps (if video underperforms in first hours, swap immediately)

Output Format

## Thumbnail Brief: "[Video Title]"

### Title-Thumbnail Relationship
- Title communicates: [what the title says]
- Thumbnail must communicate: [what the thumbnail adds — DIFFERENT from title]

### Concept 1: [Name] (RECOMMENDED)
- **Subject**: [description]
- **Expression/Emotion**: [specific emotion, closed-mouth smile / shock / etc]
- **Visual Hook**: [the one element that creates intrigue]
- **Text Overlay**: "[4 words max]" — [color], [position], [font style]
- **Color Palette**: [primary], [accent], [background]
- **Dark Mode Check**: [does it pop on dark background? Y/N]
- **Composition**: [describe layout — rule of thirds, focal point placement]
- **Why It Works**: [which research-backed triggers does it hit?]

### Concept 2: [Name]
[same format]

### Concept 3: [Name]
[same format]

### A/B Test Plan
| Variant | Change | Hypothesis |
|---------|--------|-----------|
| A | Base concept 1 | Control |
| B | [specific change] | Tests [specific variable] |
| C | [specific change] | Tests [specific variable] |

### Anti-Patterns to Avoid
- [specific things NOT to do for this video's thumbnail]

Rules

• Never repeat the title in the thumbnail
• MrBeast generates 50+ concepts per video — aim for at least 5 solid ones
• Watch Time Share matters more than CTR alone (clickbait with low retention = penalty)
• For educational channels, the thumbnail can be simpler — the title does more heavy lifting
• Always have backup thumbnails ready before publish