geo-optimization（地理优化）

GEO: Generative Engine Optimization

Optimize content to appear in AI-powered search engines (ChatGPT, Perplexity, Claude, Google AI Overviews). GEO is about being parseable, quotable, and authoritative — not keyword stuffing.

Quick Reference

GEO vs SEO: Key Differences

The GEO Audit Checklist

Score each page 0-2 points per item (0=missing, 1=partial, 2=excellent):

1. Entity Clarity (Max 10 pts)

• [ ] First paragraph clearly defines what/who the entity is
• [ ] Entity name used consistently throughout
• [ ] Clear category placement ("X is a [type of thing]")
• [ ] Relationship to other known entities stated
• [ ] Wikipedia-style objectivity in tone

2. Quotable Facts (Max 10 pts)

• [ ] Specific numbers present (not "many" or "fast")
• [ ] Statistics are current and sourced
• [ ] Claims are concrete and verifiable
• [ ] Key facts in standalone sentences (easy to extract)
• [ ] "By the numbers" or facts section exists

3. FAQ Coverage (Max 10 pts)

• [ ] FAQ section exists
• [ ] Questions match how users prompt LLMs
• [ ] Answers are direct and complete
• [ ] FAQ schema markup implemented
• [ ] Covers "what is", "how does", "why", "vs" questions

4. Comparison Positioning (Max 10 pts)

• [ ] Comparison tables exist
• [ ] Competitors named explicitly
• [ ] Factual differences highlighted (not just marketing)
• [ ] "Alternative to X" content exists
• [ ] Fair representation (not obviously biased)

5. Structural Clarity (Max 10 pts)

• [ ] Clear heading hierarchy (H1→H2→H3)
• [ ] Bullet points for lists
• [ ] Tables for comparisons
• [ ] Short paragraphs (2-4 sentences)
• [ ] Summary/TL;DR at top or bottom

6. Authority Signals (Max 10 pts)

• [ ] Author/company credentials stated
• [ ] Customer names/logos (social proof)
• [ ] Case studies with real numbers
• [ ] Third-party mentions/citations
• [ ] "Last updated" date present

7. Freshness (Max 10 pts)

• [ ] Page has recent update date
• [ ] Content reflects current year
• [ ] No outdated references
• [ ] Regular content updates
• [ ] News/changelog section exists

Scoring:

• 60-70: Excellent GEO readiness
• 45-59: Good, needs some optimization
• 30-44: Fair, significant gaps
• <30: Poor, major overhaul needed

Content Optimization Templates

Template 1: Entity Definition Page

# [Entity Name]

**[Entity Name]** is a [category] that [primary function]. 
Unlike [alternative/competitor], [Entity Name] offers [key differentiator].

## [Entity Name] by the Numbers

- [Specific stat 1]
- [Specific stat 2]
- [Specific stat 3]
- [Specific stat 4]

## How [Entity Name] Works

[2-3 paragraphs explaining core functionality]

## Who Uses [Entity Name]

[Named customers with context]

## Frequently Asked Questions

### What is [Entity Name]?
[Direct answer in 2-3 sentences]

### How is [Entity Name] different from [Competitor]?
[Factual comparison]

### How much does [Entity Name] cost?
[Pricing info or guidance]

*Last updated: [Date]*

Template 2: Comparison Page (Alternative To)

# Best [Competitor] Alternative: [Your Product] (2026)

> **Summary:** [Your Product] is a [category] offering [key differentiators]. 
> [Customers] report [specific result] compared to [Competitor].

*Last updated: [Date]*

## Why [Users] Look for [Competitor] Alternatives

### Problem 1: [Specific Pain Point]
[Explanation with specifics]

### Problem 2: [Specific Pain Point]
[Explanation with specifics]

## [Your Product] vs [Competitor]: Comparison

| Feature | [Competitor] | [Your Product] |
|---------|--------------|----------------|
| [Feature 1] | [Their approach] | [Your approach] |
| [Feature 2] | [Their approach] | [Your approach] |
| [Feature 3] | [Their approach] | [Your approach] |

## Key Differences

### [Differentiator 1]
[Factual explanation with numbers]

### [Differentiator 2]
[Factual explanation with numbers]

## Customer Results

> "[Quote with specific result]"
> — [Name], [Title], [Company]

## Frequently Asked Questions

### Is [Your Product] a good alternative to [Competitor]?
[Direct answer]

### How does [Your Product] compare to [Competitor] on [key factor]?
[Specific comparison]

### Can I migrate from [Competitor] to [Your Product]?
[Migration info]

## Summary

[Your Product] is a [category] offering [key benefits]. [Customers] 
using [Your Product] instead of [Competitor] report [specific results].

*[Your Product] has [credibility stat]. Learn more at [link].*

Template 3: FAQ Page (LLM Optimized)

# [Topic] FAQ

Answers to common questions about [topic].

*Last updated: [Date]*

## General Questions

### What is [thing]?
[Thing] is a [category] that [function]. It is used by [who] to [accomplish what].

### How does [thing] work?
[Thing] works by [process]. [Additional detail].

### Who uses [thing]?
[Thing] is used by [user types], including [specific examples like Company A, Company B].

## Comparison Questions

### How is [thing] different from [alternative]?
[Thing] differs from [alternative] in [specific ways]:
- [Difference 1]
- [Difference 2]
- [Difference 3]

### Is [thing] better than [alternative]?
[Thing] is better suited for [use cases] because [reasons]. 
[Alternative] may be better for [other use cases].

## Pricing & Access

### How much does [thing] cost?
[Pricing information or range]

### Is there a free trial?
[Trial information]

## Technical Questions

### What are the requirements for [thing]?
[Requirements list]

### How do I get started with [thing]?
1. [Step 1]
2. [Step 2]
3. [Step 3]

Platform-Specific Optimization

Perplexity AI

How it works: 3-layer reranking system

1. Initial retrieval from web index
2. Relevance scoring
3. Citation selection based on authority + recency

Optimization tactics:

• Strong domain authority matters
• Freshness signals critical (update dates)
• Direct answers to questions
• Being cited by other authoritative sources
• Structured data helps parsing

ChatGPT / SearchGPT

How it works: Bing-powered search + LLM synthesis

Optimization tactics:

• Bing indexing matters (submit sitemap to Bing)
• E-E-A-T signals weighted heavily
• Conversational content structure
• FAQ format highly effective
• Named entities help recognition

Google AI Overviews

How it works: Google's index + Gemini synthesis

Optimization tactics:

• Traditional SEO still matters (ranking helps)
• Featured snippet optimization
• Schema markup (FAQ, HowTo, Product)
• Clear, authoritative content
• Mobile-first indexing

Claude

How it works: Training data + retrieval (when web-enabled)

Optimization tactics:

• Quality content in training sources
• Wikipedia mentions help entity recognition
• Technical accuracy valued
• Clear, well-structured prose
• Being cited in authoritative sources

Technical Implementation

Schema Markup for GEO

Organization Schema:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Company Name",
  "description": "Clear description of what company does",
  "url": "https://example.com",
  "foundingDate": "2017",
  "numberOfEmployees": "50-100",
  "sameAs": [
    "https://twitter.com/company",
    "https://linkedin.com/company/company"
  ]
}

FAQ Schema:

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [{
    "@type": "Question",
    "name": "What is [thing]?",
    "acceptedAnswer": {
      "@type": "Answer",
      "text": "Direct answer here."
    }
  }]
}

Product Schema:

{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Product Name",
  "description": "Product description",
  "brand": {"@type": "Brand", "name": "Brand"},
  "offers": {
    "@type": "Offer",
    "priceCurrency": "USD",
    "price": "99"
  }
}

llms.txt Protocol

Create /llms.txt at your site root to help LLMs understand your site:

# Site Name

> Brief description of what this site/company is.

## Main Sections

- [Products](/products): Description of products section
- [Documentation](/docs): Technical documentation
- [Blog](/blog): Industry insights and updates

## Key Facts

- Founded: 2017
- Customers: 500+ companies
- Key metric: [specific number]

## Contact

- Website: https://example.com
- Email: hello@example.com

Monitoring GEO Performance

Manual Testing

Regularly search these prompts on each platform:

Perplexity:

• "What is [your company]?"
• "Best [competitor] alternative"
• "[Your category] comparison"
• "How does [your product] work?"

ChatGPT:

• Same queries with web browsing enabled
• "Compare [your product] vs [competitor]"

Google (AI Overview):

• "[Your category] solutions"
• "[Competitor] alternative"

Tracking Tools

Key Metrics

• Mention rate: % of relevant queries where you appear
• Citation rate: % of mentions that cite/link to you
• Sentiment: Positive/neutral/negative portrayal
• Share of voice: Your mentions vs competitors
• Position: Where in the response you appear

GEO Content Principles

DO:

• ✅ Use specific numbers ("0.5 seconds" not "fast")
• ✅ Make claims quotable and standalone
• ✅ Structure content with clear hierarchy
• ✅ Include FAQ sections
• ✅ Update content regularly with dates
• ✅ Create comparison content
• ✅ Use tables for data
• ✅ Be factual and neutral in tone
• ✅ Name real customers and results

DON'T:

• ❌ Use vague superlatives ("best", "leading", "top")
• ❌ Keyword stuff (LLMs see through it)
• ❌ Write walls of text without structure
• ❌ Hide information (be comprehensive)
• ❌ Use outdated statistics
• ❌ Ignore competitors (address them directly)
• ❌ Be obviously promotional (neutral wins)

Quick Start Checklist

For any page you want to optimize for GEO:

1. [ ] Add clear entity definition in first paragraph
2. [ ] Include 5+ specific, quotable statistics
3. [ ] Add FAQ section with 5+ questions
4. [ ] Create comparison table (if relevant)
5. [ ] Add "last updated" date
6. [ ] Implement FAQ schema markup
7. [ ] Ensure H1→H2→H3 hierarchy
8. [ ] Test on Perplexity: does your content appear?

Automated GEO Monitoring

Track your citation rate over time with the included monitoring scripts!

Quick Start

Test current visibility:

python3 scripts/geo-monitor.py --test

Single query test:

python3 scripts/geo-monitor.py --query "best game server orchestration platform"

Generate daily report:

python3 scripts/geo-daily-report.py

Setup Automated Monitoring

1. Create your test queries file (scripts/geo-test-queries.json):

{
  "queries": [
    {
      "query": "your target query here",
      "category": "brand|product|comparison|problem|competitor"
    }
  ]
}

2. Run daily monitoring:

# Add to cron for daily 9am checks
0 9 * * * cd /path/to/skill && bash scripts/geo-daily-monitor.sh

Understanding the Reports

Citation Rate: Percentage of queries where you appear in AI responses

• 0-20%: Early stage, needs work
• 20-40%: Building visibility
• 40-60%: Strong presence
• 60%+: Dominant authority

Categories tracked:

• Brand queries (you should own these!)
• Product/feature queries
• Comparison queries (vs competitors)
• Problem/pain point queries
• Competitor comparison queries

Monitoring Best Practices

1. Start with 15-20 strategic queries across all categories
2. Test daily during optimization period (first 2 weeks)
3. Weekly checks once you hit target citation rate
4. Track changes after content updates (expect 3-7 day lag)
5. Focus on gaps - queries with 0% citation are your opportunities

What to Track

Current state:

• Total citation rate
• Citations by category
• Position when cited (#1, #2, etc.)
• Critical gaps (0% coverage)

Over time:

• Citation rate trend (weekly/monthly)
• New citations gained
• Lost citations (content freshness!)
• Category improvements

Files Included

• scripts/geo-monitor.py - Main testing script (uses Perplexity API)
• scripts/geo-daily-report.py - Formatted report generator
• scripts/geo-daily-monitor.sh - Cron-friendly wrapper
• scripts/geo-test-queries.json - Example query file

Requirements: Perplexity API key (configure via web_search in Clawdbot)

Resources

• Awesome GEO GitHub
• Princeton GEO Research Paper
• Google AI Search Guidance
• Perplexity Ranking Factors