30x Growth Marketing Panel

Skill by ara.so — Marketing Skills collection.

An AI-powered expert panel of 11 world-class marketing experts distilled from 4,000+ YouTube videos. Get answers from the right expert(s) in their voice, using their actual frameworks.

What It Does

The 30x Growth Marketing Panel uses a dual-layer architecture to provide authentic expert advice:

• Layer 1 (Brain): NotebookLM retrieval from 4,000+ indexed YouTube videos
• Layer 2 (Soul): Persona Protocol with expert personality, frameworks, and anti-patterns
• Semantic routing: Automatically matches your question to the right expert(s)
• Anti-hallucination: Retrieve-first protocol ensures responses are grounded in actual expert content

Installation

npx skills add norahe0304-art/30x-growth-marketing-panel

Works with Claude Code, Cursor, Codex, and 45+ AI coding agents.

The Expert Panel

Usage Patterns

Single Expert Consultation

Ask focused questions to get advice from the most relevant expert:

# Pricing question → routes to Alex Hormozi
"How should I price my B2B SaaS product?"

# SEO question → routes to Neil Patel or Nathan Gotch
"What's the best AI SEO strategy for 2026?"

# Content question → routes to Julia McCoy
"How do I build a consistent content voice?"

Named Expert Request

Explicitly request a specific expert:

"Ask Alex Hormozi about my offer: [describe your offer]"

"What would Neil Patel say about this landing page conversion issue?"

"Get Greg Isenberg's take on community-led growth for an AI tool"

Multi-Expert Roundtable

Broad strategic questions trigger multiple experts:

"How should I go to market with a new AI marketing tool?"
# Returns perspectives from Greg Isenberg, Neil Patel, Ryan Doser

"What's the best growth strategy for a bootstrapped SaaS?"
# Returns perspectives from Alex Hormozi, Dan Koe, Growth Tribe

Expert Knowledge Base Structure

Each expert has two components:

1. NotebookLM Brain (Raw Retrieval)

# 200-300 YouTube videos per expert
# Indexed in NotebookLM Pro (300 sources/notebook)
# Zero information loss from original content

2. Persona Protocol (Personality)

Located in expert_kb.md for each expert:

## Role
Who the expert is, their background, core expertise

## Thinking Models
Frameworks they use (e.g., Hormozi's Value Equation, AARRR funnel)

## Tone & Communication
How they speak, teaching style, personality markers

## Anti-Patterns
What they avoid, common mistakes they call out

## Retrieval Logic
How to search their NotebookLM notebook effectively

Anti-Hallucination Protocol

The panel follows strict retrieval rules:

1. Retrieve first: Must search NotebookLM before generating responses
2. Dual verification: Cross-reference retrieval with KB persona
3. Explicit marking: Extrapolations from core principles marked with ⚠️
4. Never fabricate: If an expert hasn't covered a topic, say so

Example output structure:

**Alex Hormozi's Perspective:**

[Retrieved content from NotebookLM]

Framework: Value Equation
- Dream Outcome: [specific to your question]
- Perceived Likelihood: [specific analysis]
- Time Delay: [specific analysis]  
- Effort & Sacrifice: [specific analysis]

⚠️ *Extrapolating from core principles:* [only if needed]

Distilling Your Own Expert

Use the distill_anyone.md prompt template:

# 1. Copy the prompt from distill_anyone.md
# 2. Change 3 variables:
#    - Expert name
#    - YouTube channel/playlist URL
#    - Domain expertise
# 3. Run in Claude Code

# The pipeline automatically:
# - Collects YouTube URLs with yt-dlp
# - Creates NotebookLM notebook
# - Bulk adds videos with notebooklm-py
# - Generates Persona Protocol KB
# - Creates semantic routing rules

Variables to Configure

EXPERT_NAME="Your Expert"
YOUTUBE_SOURCE="https://youtube.com/@channel or playlist URL"
DOMAIN="their core expertise area"

Key Commands

Query the Panel

# General query (auto-routes)
"[Your marketing question]"

# Named expert
"Ask [Expert Name] about [topic]"

# Multi-expert
"Get a roundtable on [strategic question]"

Inspect Expert Knowledge

# List all experts and domains
"Show me the expert panel"

# View expert frameworks
"What frameworks does Alex Hormozi use?"

# Check NotebookLM coverage
"How many videos are indexed for Neil Patel?"

Configuration

Language Support

Responses automatically match your query language. Framework names stay in English:

# Query in Spanish → Response in Spanish
"¿Cómo debería fijar el precio de mi SaaS?"

# Query in French → Response in French  
"Quelle est la meilleure stratégie SEO?"

Retrieval Depth

Adjust how many NotebookLM sources to search:

# Default: 5-10 most relevant sources
# Deep research: 20-30 sources (slower, more comprehensive)
# Quick answer: 3-5 sources (faster, still accurate)

Real Code Examples

Example 1: Pricing Strategy (Alex Hormozi)

Query:

"How should I price my B2B SaaS with a $10k LTV and 6-month sales cycle?"

Expected Response:

**Alex Hormozi's Pricing Framework:**

[Retrieved from NotebookLM: Hormozi's pricing videos]

**Value Equation Analysis:**
1. Dream Outcome: Enterprise efficiency gain (high)
2. Perceived Likelihood: 6-month cycle = trust-building needed
3. Time Delay: 6 months is long → charge more upfront
4. Effort & Sacrifice: Implementation complexity matters

**Pricing Recommendation:**
- Annual contract: $5,000-$7,500/year
- Upfront discount: 20% off for annual vs monthly
- Reduce time delay perception: 30-day quick wins
- Stack value: Implementation support included

**Hormozi's Principle:** "Charge for the dream outcome, not your cost"

Example 2: SEO Strategy (Neil Patel + Nathan Gotch)

Query:

"What's the best AI SEO strategy for a new marketing tool in 2026?"

Expected Response:

**Neil Patel's Take:**
[Retrieved content about AI content + E-E-A-T]

- Focus on helpful content, not just keywords
- Build topical authority clusters
- Optimize for Google's AI Overviews

**Nathan Gotch's SEO Approach:**
[Retrieved content about Search Everywhere Optimization]

- SEO = Search Everywhere Optimization now
- Target YouTube, LinkedIn, Reddit, not just Google
- Use AI to create platform-specific content variants

**Combined Strategy:**
1. Build content cluster on "AI marketing automation"
2. Create YouTube tutorials (ranks in Google + YouTube)
3. Repurpose to LinkedIn thought leadership
4. Use Reddit for community validation signals

Example 3: Community-Led Growth (Greg Isenberg)

Query:

"How do I build community around my AI productivity tool?"

Expected Response:

**Greg Isenberg's Community Framework:**

[Retrieved from NotebookLM: Community-building videos]

**Phase 1: Vibe Check (Month 1-2)**
- Find where your ideal users already hang out
- Join conversations, don't pitch
- Identify the "energy" — what gets them excited?

**Phase 2: Minimum Viable Community (Month 3)**
- Discord or Slack, not Facebook
- 1 channel: #wins (social proof flywheel)
- You post first 20 wins yourself if needed

**Phase 3: Community-Led Product**
- Ship features the community requests
- Give power users early access
- Make them feel like co-creators

**Greg's Key Insight:** "Community isn't a channel, it's a moat"

Common Patterns

Pattern 1: Multi-Stage Funnel Question

# Question spans multiple domains
"I need help with my SaaS go-to-market: offer, SEO, and community"

# Response includes:
# - Alex Hormozi: Offer positioning
# - Neil Patel: SEO strategy  
# - Greg Isenberg: Community layer

Pattern 2: Framework Deep-Dive

# Request specific framework
"Explain Alex Hormozi's Value Equation for my use case"

# Response:
# - Retrieves original explanation from NotebookLM
# - Maps framework to your specific scenario
# - Includes anti-patterns from KB

Pattern 3: Comparative Analysis

# Compare expert approaches
"How would Dan Koe vs Alex Hormozi approach pricing a course?"

# Response:
# - Dan Koe: Personal brand, premium positioning, audience relationship
# - Alex Hormozi: Value equation, enterprise pricing, sales frameworks
# - Synthesis: When to use each approach

Troubleshooting

Issue: Generic or Vague Response

Problem: Response doesn't sound like the expert

Solution:

• Check if question is in expert's domain
• Request named expert explicitly
• Ask for specific framework by name

# Instead of: "How do I market?"
# Try: "Ask Alex Hormozi: How should I position my offer using the Value Equation?"

Issue: No Retrieval Evidence

Problem: Response lacks [Retrieved from NotebookLM] markers

Solution:

• Expert may not have covered this topic
• Reframe question to match expert's known content areas
• Check expert domain table above

Issue: Multi-Expert Overload

Problem: Too many perspectives for a simple question

Solution:

• Ask for single expert
• Rephrase as focused question

# Instead of: "How do I grow?"
# Try: "What's Neil Patel's SEO strategy for [specific use case]?"

Issue: Outdated Framework

Problem: Expert's content is from 2023-2024

Solution:

• Ask for principles, not tactics
• Request ⚠️ extrapolation for 2026 context

"What would Neil Patel's SEO principles be for 2026, given AI Overviews?"

Advanced Usage

Combine with Your Context

# Provide your specific situation
"Here's my SaaS: [details]. Ask Alex Hormozi how to price it."

# Attach data
"My conversion rate is 2%. Ask Neil Patel to audit my funnel."

Sequential Expert Consultation

# Step 1: Offer with Hormozi
"Alex Hormozi: Review my offer"

# Step 2: Traffic with Neil Patel  
"Neil Patel: Now how do I drive traffic to this offer?"

# Step 3: Community with Greg Isenberg
"Greg Isenberg: Should I add a community layer?"

Export Expert Advice

# Generate structured output
"Create a marketing strategy doc consulting:
- Alex Hormozi for offer
- Neil Patel for SEO
- Greg Isenberg for community"

# Output: Markdown doc with all expert perspectives organized

Tools Used Internally

The panel is built with:

• yt-dlp: YouTube URL batch collection
• notebooklm-py: Programmatic NotebookLM access
• NotebookLM Pro: 300 sources/notebook indexing
• Claude Code Skills: Persona Protocol + dual-layer fusion

You don't need to install these separately — they're embedded in the skill.

Best Practices

1. Be specific: "How do I price?" → "How do I price a B2B SaaS at $10k ACV?"
2. Name the expert: When you know who you want
3. Provide context: Share your industry, stage, constraints
4. Request frameworks: Ask for specific models by name
5. Iterate: Start with one expert, then consult others

License

MIT — Free to use, modify, and distribute.