# Outlier Hunter

> Find outlier videos (3x+ above channel average) across niches. Extract the FORMAT, not the topic. Adapt with your brand voice. MrBeast's 'copy with taste' method. Use when hunting for video ideas.

- Skill: `ravsau/outlier-hunter` (Agent Skill)
- Install (CLI): `npx skillmds@latest add ravsau/outlier-hunter`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ravsau/outlier-hunter/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: ravsau (https://skillmd.com/u/ravsau)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/ravsau/outlier-hunter

---


# Outlier Hunter — "Copy With Taste"

Hunt outliers for: **$ARGUMENTS**

## Research-First Protocol

Do not guess when a recommendation depends on current market reality. If the user asks about a specific niche, channel, competitor, or trend, collect evidence first.

Minimum useful research set:
- **Top 10 current videos/channels** in the niche or adjacent niche
- **Outlier multiple** vs each channel's recent average
- **Packaging patterns**: title structure, thumbnail elements, first 5 seconds when available
- **Trend signal**: search interest, platform momentum, or recurring audience questions
- **Budget/ROI reality**: what the format costs and whether the upside can justify it

If public data is weak, ask for the creator's last 3 videos with CTR, AVD, retention at 30 seconds, and net subscribers. Without those numbers, label assumptions clearly.

## The Method (MrBeast)

"Don't copy and paste. Copy with taste."
"There's only one time you should aim for originality: when you're adapting winning formats to your unique brand."

## Step 1: Find Outliers

An outlier video gets **3x+ above the channel's average views**. These are format signals, not topic signals.

### Where to Look
1. **Your direct niche** — channels covering similar topics
2. **Adjacent niches** — similar audience, different topic
3. **Completely different niches** — same format, different everything
4. **Trending/breakout** — new channels that blew up with one video

### What to Measure
For each channel you study:
- Calculate average views per video (last 20 videos)
- Flag any video at 3x+ that average
- Note: exclude outliers caused by external events (news mentions, algorithm bugs)

### Study Cadence (Jake Thomas)
- Study outlier videos across **5 industries weekly**
- Spend 4 hours/week studying YouTube across niches (Paddy Galloway)
- After 50 of your own uploads: analyze YOUR top 5 and bottom 5

## Step 2: Extract the Format

For each outlier, deconstruct:

### Packaging
- **Title structure**: What pattern does the title follow? (Number? Question? Claim? Comparison?)
- **Thumbnail elements**: What's in the thumbnail? (Face? Object? Text? Contrast?)
- **Title-thumbnail relationship**: How do they complement each other?

### Content Structure
- **Hook** (first 30 seconds): What did they open with?
- **Pacing**: How are segments structured? Fast/slow/narrative?
- **Stakes**: What keeps you watching? What's the tension?
- **Payoff**: How does it end? Satisfying? Cliffhanger?

### The Format (the extractable pattern)
- Strip away the specific topic
- What's the underlying STRUCTURE?
- Example: "Person tries [extreme thing] for [time period] and documents results" — works for any niche
- Example: "[Tool A] vs [Tool B]: I tested both for [X] days" — works for any category
- Translate expensive formats down to the creator's budget: keep the core tension, reduce the spend.

## Step 3: Adapt to Your Brand

Take the format and apply:
- Your niche topic
- Your voice and tone
- Your unique angle (expertise, experience, contrarian take)
- Your audience's specific pain points

### Quality Check
- Does this feel authentic to your brand? (MrBeast: if it doesn't excite you, kill it)
- Can you deliver more value than the original? (expertise advantage)
- Is the packaging naturally different because of your angle?

## Step 4: Validate Before Filming

Run through these checks:
- [ ] Format has proven demand (3x+ outlier in at least one channel)
- [ ] Your adaptation adds unique value (not a copy)
- [ ] Title/thumbnail conceived before scripting (packaging sets ceiling)
- [ ] You're personally excited about this video
- [ ] The format serves your audience's needs (not just chasing views)

## Roberto Blake Stats for Context

- 88% of YouTube videos never reach 1,000 views
- You're competing with the top 5 in YOUR niche, not all of YouTube
- Educational content gets 80% of views 6+ months post-upload
- Fewer than 0.05% of channels reach 1M subscribers

## Output Format

```markdown
## Outlier Analysis: [Niche/Topic]

### Channels Analyzed
| Channel | Avg Views | Subscribers | Niche |
|---------|-----------|-------------|-------|
| [name] | [avg] | [subs] | [niche] |

### Outlier Videos Found
| Video | Channel | Views | Channel Avg | Multiple | Why It Hit |
|-------|---------|-------|-------------|----------|-----------|
| [title] | [channel] | [views] | [avg] | [X]x | [reason] |

### Formats Extracted
1. **[Format Name]**: [description of the pattern]
   - Proven by: [which outliers]
   - Your adaptation: "[your title idea using this format]"
   - Why it fits your brand: [reason]

2. **[Format Name]**: [description]
   [same structure]

### Top 3 Adaptations (Ready to Film)

**#1: "[Your Title]"**
- Format source: [which outlier inspired this]
- Your unique angle: [what makes yours different]
- Thumbnail direction: [what the thumbnail shows]
- Estimated ceiling: [views potential based on format performance]
- Excitement check: [honest gut feeling]

[Repeat for #2 and #3]
```