Youtube Title Creator

YouTube Title & Thumbnail Creator

Purpose

Generate YouTube titles and thumbnails with a high degree of complementarity that achieve 6%+ click-through rates using the framework fitting method: extract your concept, match to 119 proven formulas ranked by performance, generate variations, select best.

Core Philosophy: Good titles follow proven frameworks. The skill is in matching your concept to the delivery mechanism that amplifies it best.

When to Use This Skill

Use this skill when you need to:

• Create YouTube titles for podcast episodes or educational videos
• Generate thumbnail text that complements titles
• A/B test title variations for same content

Do NOT use for:

• Blog post headlines (use hook-and-headline-writing instead)
• Social media captions (use social-caption-writer instead)
• Newsletter subject lines (use hook-and-headline-writing instead)

---

Channel Size Strategy (Critical Context)

Before generating titles, determine channel size tier. The right strategy depends on audience scale.

Tier 1: Growing Channels ( recognition** (unless George Clooney-famous)

• Stock photos > illustrations
• Show outcome/situation, not just person's face
• Dial to 11: "Most extreme version of that title?"
• Simplify text: Fewer words, higher contrast, more charged

Thumbnail Anti-Patterns (Especially for Podcasts):

• NEVER use a podcast recording screenshot as thumbnail. Two unknown faces in earbuds/mics = instant skip. This is the #1 podcast thumbnail mistake.
• NEVER use the guest's face alone unless they're famous. An unknown face with text overlay says "skip."
• NEVER make the thumbnail text a quote from the episode. Quotes require context. Thumbnails must work instantly.
• NEVER use more than 4 words of text. If the thumbnail needs explaining, it fails.

What Podcast Thumbnails SHOULD Show:

• The concept, not the conversation. If the episode is about removing TV from the home → show a living room with books where the TV should be.
• Visual contrast or transformation. Split images (before/after, vs., choice between two things) outperform single-scene thumbnails.
• Props, environments, or situations that tell the story without faces.
• If you must use a face: Show genuine emotion (shock, concern, joy) — not a neutral "podcast smile."

Thumbnail Text Formulas:

• "Here's how"
• "Number 3 will shock you"
• "This changed everything"
• "Schools are wrong"
• "CRAZY"
• "USE IT OR LOSE IT"
• "BEFORE → AFTER"
• "STOP DOING THIS"
• Question: "Is that even allowed?"

Step 3: Quality Test All Title Variations

Before finalizing your 5 titles, test each against these criteria:

The McDonald's Test (Andrew Muto - July 8): "Someone who works at McDonald's should understand your headline instantly."

• [ ] No jargon
• [ ] Accessible language
• [ ] Instant comprehension

The Ultimate Hook Test:

• [ ] Would this make ME stop scrolling?
• [ ] Creates curiosity or emotion in first 2 seconds?
• [ ] Can my grandmother understand instantly?
• [ ] Hints at payoff without giving it away?

Discard any titles that fail these tests. Keep the strongest 5.

Step 4: Generate 5 Titles with 3 Thumbnail Options Each

Generate 5 distinct title variations across different frameworks and strategic approaches. For each title, create 3 complementary thumbnail options that vary the visual approach while staying simple (3 elements max).

For each title, structure as:

• Title (single, clear concept)
• Framework reference
• 3 thumbnail options (each with visual description + on-screen text)

Why 3 thumbnails per title? Testing allows you to optimize visual delivery without changing the messaging. Different thumbnail styles appeal to different viewer segments.

Thumbnail variation strategies:

• Option A: Direct visual (show the problem/outcome)
• Option B: Emotional/reaction (show what it feels like)
• Option C: Curiosity/question (hint at the answer without giving it)

This approach gives the video production team multiple ready-to-test options without requiring additional iteration.

---

Output Format

Simplified Output for Practical Use

Create file: [Video_Topic]_YouTube_Titles_Thumbnails.md

# YouTube Titles & Thumbnails - [Topic]

---

## TITLE 1: "[Title]"

**Framework:** #3 - How To (Achieve goal)! Without (Unwanted action)

**Thumbnail Option A:**
- Visual: [Description of image/scene]
- Text: [On-screen text]

**Thumbnail Option B:**
- Visual: [Description of image/scene]
- Text: [On-screen text]

**Thumbnail Option C:**
- Visual: [Description of image/scene]
- Text: [On-screen text]

---

## TITLE 2: "[Title]"

**Framework:** #6 - The REALITY of (Polarizing entity) in (Current year)

**Thumbnail Option A:**
- Visual: [Description of image/scene]
- Text: [On-screen text]

**Thumbnail Option B:**
- Visual: [Description of image/scene]
- Text: [On-screen text]

**Thumbnail Option C:**
- Visual: [Description of image/scene]
- Text: [On-screen text]

---

[Repeat for Titles 3, 4, and 5]

Key Rules:

• 5 titles total
• 3 thumbnail options per title
• Each thumbnail: 3 elements max (visual + text, keep it simple)
• No detailed explanation of principles or Hook Scores
• Focus on actionable output ready for video production

Bundled Resources

Framework Libraries

• references/creator-hooks-ranked.md - All 119 frameworks sorted by Hook Score
• references/psychological-principles-guide.md - Deep dive on each trigger
• references/ctr-benchmarks.md - Performance expectations and troubleshooting

Strategy Guides

• references/universalization-examples.md - Before/after comparisons
• references/thumbnail-complementarity.md - Visual pairing strategies
• references/andrew-muto-formulas.md - Additional formulas from sessions

Examples

• examples/katie-kimball-walkthrough.md - Complete framework fitting process
• examples/dyslexia-universalization.md - Narrowing vs. broadening analysis
• examples/george-washington-from-x-to-y.md - Transformation formula

---

Common Pitfalls to Avoid

Phase 1 Issues

❌ Vague content summary - Can't match frameworks without clarity ❌ Missing elements - Incomplete extraction leads to weak options ❌ Wrong video type - Misclassifying content limits framework options

Phase 2 Issues

❌ First-match bias - Using first framework that fits ❌ Insufficient volume - Testing only 2-3 frameworks ❌ Ignoring Hook Scores - Not prioritizing proven performers ❌ Over-filtering - Adding too many audience limiters ❌ Descriptive over provocative - "What X Looks Like" describes content; "Your Brain Can Do This at 8 But NOT at 25" provokes emotion. Always ask: does this title describe or provoke? ❌ Leading with unknown guest name - Guest names are only hooks if the guest is famous. Use credentials as adjectives instead.

Phase 3 Issues

❌ Thumbnail repetition - Title and thumbnail say same thing ❌ Failing McDonald's Test - Using jargon or complexity ❌ Single variation - Not creating A/B test options ❌ Low principle count - Triggering only 1-2 psychological levers ❌ Podcast screenshot thumbnails - Two faces in mics/earbuds is the lowest-performing thumbnail archetype. Show the concept, not the conversation. ❌ Quote-as-thumbnail-text - Episode quotes need context. Thumbnail text must land instantly without context ("USE IT OR LOSE IT" beats "TV'S NOT WHO WE ARE").

---

Advanced Techniques

The False Binary

Present two extremes as only options: "Burnout or Boredom: Where Most Careers End"

The Credibility Bomb

Drop credentials mid-controversy: "After Managing 300 Teams: This Generation Can't..."

The Mirror Moment

Make viewer see themselves: "If Your Team Can't Do This One Thing..."

Time-Based Urgency

Add current year or timeframe: "The Reality of Remote Work in 2025"

Reverse Expectations

Flip the expected outcome: "My First Launch Failed. Perfect."

---

---

Related Skills

• hook-and-headline-writing - Provides foundational hook strategies
• cold-open-creator - Creates opening that delivers on title promise
• social-caption-writer - Adapts titles for social platforms

---

Version History

• v1.1 (2026-02-12): Channel-size strategy & podcast-specific guidance
• Added Channel Size Strategy section (Tier 1/2/3) — topic carries title for small channels, not guest name
• Added Podcast-Specific Title Guidance — extract provocative claims, avoid interview-format titles
• Added Thumbnail Anti-Patterns for podcasts — never use podcast screenshots, never use episode quotes as thumbnail text
• Added "descriptive vs provocative" pitfall — titles must provoke emotion, not describe content
• Added new thumbnail text formulas (USE IT OR LOSE IT, BEFORE → AFTER, STOP DOING THIS)
• Synced updates to podcast-production skill (Checkpoint 3) and checkpoint-3-template

• v1.0 (2025-10-28): Initial skill creation
• Framework fitting method from social-content-creation
• 119 Creator Hooks frameworks integrated
• Andrew Muto strategies (universalization, dial to 11, CTR benchmarks)
• Complementarity principle
• 3-phase workflow
• Quality control systems

---

For complete framework library, psychological principles, and examples, see references folder

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: cdeistopened
• Source: cdeistopened/content-os
• License: MIT

Install and usage instructions live in the source repository linked above.