Video Title, Thumbnail and Content Retention Check
Professional check capability for video titles, thumbnails and content retention based on MrBeast cases, to improve click-through rate and watch time.
When to Use This Skill
This skill will be loaded automatically when the following scenarios are detected:
- Users say "check title and thumbnail", "check content retention"
- After the first draft of the video script is completed
- Users mention "low click-through rate", "short watch time"
- Need to optimize video title or thumbnail
Core Objectives
- Improve click-through rate: Titles and thumbnails attract users to click
- Improve watch time: Good content retention keeps users watching till the end
- Build reusable formulas: Standardized title, thumbnail and retention strategies
Check Dimensions
1️⃣ Title Check (Strong Contrast Strategy)
2️⃣ Thumbnail Check (Result Display Priority)
3️⃣ Content Retention Check (Opening Confirmation → Mid-section Surprise → Ending Delivery)
1️⃣ Title Check: Strong Contrast Strategy
Core Principle
Focus on the strongest contrast, do not list all content
5 Strong Contrast Formulas
Formula 1: Quantity Contrast
Format:
Small quantity vs Large quantity
Examples:
- ✅ "1 Prompt wrote 1000 lines of code"
- ✅ "3 people vs 300 people team, who will win?"
- ✅ "Replace 10 tools with 1 tool"
Applicable to: Tutorials, efficiency content, tool content
Formula 2: Price Contrast
Examples:
- ✅ "What's the difference between 200-yuan AI and 20,000-yuan AI?"
- ✅ "What can free AI do? I tested 10 tools"
- ✅ "Is Claude 10 times more expensive than GPT worth it for the same function?"
Applicable to: Review content, comparison content
Formula 3: Result Contrast
Format:
Low achievement vs High achievement
or
Examples:
- ✅ "300k followers vs 0 followers, I only did one thing right"
- ✅ "I used 3 tricks to get from 0 to 1000 views"
- ✅ "The same video got 10 times more views after changing the title"
Applicable to: Experience sharing content, growth content
Formula 4: Strength Contrast
Format:
Novice/Underdog vs Expert/Top player
Examples:
- ✅ "Novice vs Cursor: Who can write this function faster?"
- ✅ "I asked AI to compete with programmers, the result shocked me"
- ✅ "Beginner challenges to develop an App with AI"
Applicable to: Challenge content, experiment content
Formula 5: Time Contrast
Examples:
- ✅ "3 minutes vs 3 hours: Is AI programming really faster?"
- ✅ "I finished a week's work in 1 day"
- ✅ "It only takes 10 minutes to learn and get started with Cursor"
Applicable to: Tutorials, efficiency content
Title Type Selection Table
| Content Type | Recommended Title Type | Example |
|---|
| Tutorial | Result contrast, time contrast | "One Cursor feature doubled my efficiency" |
| Review | Price/performance contrast | "What's the difference between free AI and paid AI?" |
| Challenge | Strength/quantity contrast | "Beginner challenges to develop an App with AI" |
| Experience Sharing | Result contrast | "The same video got 10 times more views after changing the title" |
Title Types to Avoid
❌ List type: "5 common features of Cursor"
- Problem: No conflict, no suspense
- Improvement: Focus on 1 strongest feature → "One Cursor feature doubled my efficiency"
❌ Plain type: "Cursor usage tutorial"
- Problem: No attraction, no specific value
- Improvement: Add result contrast → "Get started with Cursor in 3 minutes, 5 times efficiency improvement"
❌ Abstract type: "Future trends of AI programming"
- Problem: Too abstract, hard to imagine scenes
- Improvement: Specific scenario → "I have been writing code with AI for 3 months, and found these problems"
Title Check List
When performing title check, confirm the following items:
2️⃣ Thumbnail Check: Result Display Strategy
Core Principle
Choose thumbnail strategy according to follower count and content type
Applicable Conditions for Face Thumbnails (Use with caution)
✅ Applicable scenarios:
- Follower count > 100k (Face has brand value)
- Content type: Challenge, Vlog, high-emotion content
- Extreme and consistent expressions (Surprise/excitement/shock)
- Fixed thumbnail style has been established
❌ Not applicable scenarios:
- Follower count < 100k (Face has no brand value, results are more direct)
- Technical tutorial content (Results are more important than emotions)
- Different expressions in each issue (Cannot build memory points)
Applicable Conditions for Result Display Thumbnails (Recommended to test first)
✅ Applicable scenarios:
- Tutorials: Before/After contrast
- Reviews: Product close-up + core data
- Technical content: Code/interface screenshot + result display
- Follower count < 100k (Results are more direct than faces)
Thumbnail Design Elements
1. Strong contrast
- Color contrast: Cold and warm color contrast (Blue + orange)
- Size contrast: Enlarge core elements, reduce secondary elements
- Number contrast: Large number vs small number (300k vs 0)
2. Visual focus
- 1-2 core elements: Do not put too many elements
- Clear visual center: What users see at first glance
3. Concise text
- 0-5 characters: Vision is information, use less text
- Unified font: Fixed font to build brand
4. Consistent style
- Fixed color system: E.g. blue + orange
- Fixed layout: E.g. result on the left, number on the right
- Fixed font: Consistent thumbnail style for each issue
Thumbnail Type Suggestions
| Content Type | Recommended Thumbnail Type | Example |
|---|
| Tutorial | Before/After contrast | Original code on the left, optimized version on the right |
| Review | Product close-up + data | Product interface + performance numbers |
| Challenge | Extreme scenario + emotion | Face optional (if the effect is good) |
| Experience Sharing | Data contrast | Views change curve |
Thumbnail Designs to Avoid
❌ Pure text thumbnail: Too much information, no visual impact
❌ Too many elements: Visual confusion, no focus
❌ Different style for each issue: Cannot build memory points
A/B Test Suggestions
Prepare 2 thumbnails for the same video
- A: Face thumbnail (if you want to test)
- B: Result display thumbnail
- Record click-through rate data, keep the one with better effect
Thumbnail Check List
When performing thumbnail check, confirm the following items:
3️⃣ Content Retention Check: Suspense Postponement Strategy
Core Principle
Build suspense at the beginning, exceed expectations in the middle, deliver promises at the end
Standard Retention Structure
标题/封面:核心悬念(最强对比/冲突)
↓
开头(前30秒-1分钟):
- 确认标题内容存在(建立信任)
- 建立悬念:"到底能不能做到?"
- 预告后续:"但在那之前,先..."
↓
中段(主要内容):
- 主要教学/演示内容
- 加入"意外惊喜"(超出预期)
- 持续提醒悬念:"快了,马上就到..."
↓
结尾(最后1-3分钟):
- 兑现标题承诺(展示核心内容)
- 总结要点
- Call to Action(关注/点赞/评论)
Duration Control Guide
| Total Video Duration | Opening Confirmation | Mid-section Content | Ending Delivery |
|---|
| 3-5 minutes | 30 seconds | 2-3 minutes | 1 minute |
| 5-10 minutes | 1 minute | 6-7 minutes | 2 minutes |
| 10-15 minutes | 1-2 minutes | 8-10 minutes | 3 minutes |
Retention Check Example
Example: Tutorial Video
Title: One Cursor feature doubled my efficiency
✅ Correct Retention:
Opening (30 seconds):
"Cursor has a feature that saves me 3 hours per project.
What feature? I'll demonstrate later.
But before that, let's quickly go through the basic usage of Cursor,
so you can understand why this feature is so powerful."
← Confirm the promise exists, build suspense
Mid-section (5-8 minutes):
- Basic function introduction
- Installation and configuration
- Common problems
(Audience: Oh, there are these things, I learned a lot)
← Unexpected surprise, exceed expectations
Ending (2 minutes):
"Okay, now let's talk about that core feature.
This is it (demonstrate)... Look, isn't it powerful?
I have used it for 3 months, and it saves 3 hours per project..."
← Deliver the promise, complete the closed loop
❌ Wrong Retention:
Opening does not mention the title content at all:
"Today I will teach you to use Cursor. Cursor is an AI programming tool..."
Problem: Audience doesn't know when the "efficiency doubling feature" will appear → Drop off halfway
Tell everything at the opening:
"Cursor has a feature called XX that doubles efficiency. Use it like this..."
Problem: No suspense later → Short watch time
Title says A, video is all about B:
Title: "One Cursor feature doubled my efficiency"
Video: Full of basic Cursor usage, no mention of "efficiency doubling feature"
Problem: Sense of deception → Lose followers
Suspense is delayed too long:
10-minute video, only mention the "efficiency doubling feature" at the 8th minute
Problem: Audience loses patience → Drop off halfway
Common Mistakes (Must Avoid)
| Mistake Type | Consequence | Correct Practice |
|---|
| Opening does not mention title content at all | Audience doesn't know when it will appear → Drop off halfway | Confirm the promise in the first 30 seconds to 1 minute |
| Tell everything at the opening | No suspense later → Short watch time | Postpone suspense to the end |
| Title says A, video is all about B | Sense of deception → Lose followers | Title promises must be delivered |
| Suspense is delayed too long | Audience loses patience → Drop off halfway | Allocate reasonably according to duration |
Content Retention Check List
When performing content retention check, confirm the following items:
Complete Check Process
Step 1: Title Check
- Identify title type (5 contrast formulas)
- Check if it focuses on 1 core point
- Check if suspense is built
- Check title length (15-25 Chinese characters)
Step 2: Thumbnail Check
- Choose thumbnail strategy according to follower count and content type
- Check if there is strong contrast (color, size, number)
- Check visual focus (1-2 core elements)
- Check concise text (0-5 characters)
- Check style consistency
Step 3: Content Retention Check
- Check if the opening confirms the title promise (30 seconds to 1 minute)
- Check if there is unexpected surprise in the mid-section
- Check if the promise is delivered at the end
- Check if the suspense postponement duration is reasonable
- Check if there is a Call to Action
Step 4: Comprehensive Evaluation
Reference Resources
/视频创作/_knowledge_base/MrBeast视频封标与承接策略-2025年10月.md
- Detailed analysis of MrBeast cases
/自媒体经验梳理/内容策略/封面标题与内容承接-MrBeast案例.md
- Practical experience summary
Success Cases
- A certain AI programming tutorial video: Used "3 minutes to get started vs 3 hours to learn" contrast title, click-through rate increased by 40%
- A certain product review video: Used result display thumbnail, watch time increased by 25%
Last Updated: 2025-11-07
Applicable Project: Video creation
Maintainer: Hua Sheng
Produced by Uncle Hua | AI Native Coder · Independent Developer
Official Account "Uncle Hua" | 300k+ Followers | AI Tools & Efficiency Improvement
Representative works: Kitten Fill Light (Top 1 on AppStore Paid List) · "Master DeepSeek in One Book"