72/100 Site Score: The Astro Framework Was Right, But the Content Depth Was Missing
A GEO-SEO audit of a GEO service provider reveals the critical gap between perfect technical infrastructure and inadequate content depth — and why great tech alone won't earn AI citations.
Photo by Lukas Blazek on Pexels
Case Background
In June 2026, we performed a full GEO-SEO diagnostic audit on a GEO service provider’s own official website. Yes — a company that builds GEO-optimized websites for clients had their own site put under the microscope.
Client Profile:
- Business Type: GEO / AI marketing and website development services
- Tech Stack: Astro v4.16.19 (static site generation)
- Server: nginx 1.24.0 / Ubuntu
- Core Strength: Cutting-edge frontend framework, solid technical SEO fundamentals
The Core Question: Why does a technologically modern, well-architected website only score 72 out of 100 on GEO readiness?
This case study is particularly instructive because the site got nearly everything right on the technical side — and still fell short. If a GEO service provider’s own website has gaps, imagine what a typical industrial B2B site looks like.
Diagnostic Results: A “Good” 72/100
Using our GEO-SEO diagnostic scoring system, this site achieved a composite score of 72/100, which falls into the “Good — room for improvement” band.
Here is the breakdown across all evaluated dimensions:
| Dimension | Weight | Score | Percentage |
|---|---|---|---|
| AI Citability & Visibility | 25% | 18/25 | 72% |
| Brand Authority Signals | 20% | 8/20 | 40% |
| Content Quality & E-E-A-T | 20% | 12/20 | 60% |
| Technical Foundation | 15% | 10/15 | 67% |
| Structured Data | 10% | 9/10 | 90% |
| Platform Optimization | 10% | 7/10 | 70% |
| Composite GEO Score | 100% | 72/100 | 🟡 Good |
The 72/100 score is respectable — far better than the typical sub-30 scores we see from traditional industrial websites. But for a company that sells GEO optimization services, it reveals a critical blind spot: content depth.
Win #1: Technical Foundation Was Nearly Flawless
Robots.txt Configuration — Textbook Perfect
The site’s robots.txt was configured exactly as GEO best practices dictate. Every major AI crawler was explicitly allowed:
| Crawler | Status |
|---|---|
| GPTBot | ✅ ALLOWED |
| OAI-SearchBot | ✅ ALLOWED |
| ChatGPT-User | ✅ ALLOWED |
| ClaudeBot | ✅ ALLOWED |
| PerplexityBot | ✅ ALLOWED |
| Google-Extended | ✅ ALLOWED |
| Bytespider | ✅ ALLOWED |
| … (14 AI crawlers total) | ✅ All permitted |
Score: 25/25 ✅
This is a textbook configuration. Most enterprise websites either lack a robots.txt entirely or inadvertently block AI crawlers. This site got it right.
llms.txt — Fully Deployed
The site deployed both /llms.txt and /llms-full.txt, providing AI crawlers with a structured content manifest. The files included:
- ✅ Homepage, services, blog, and about page links
- ✅ Free diagnostic tool link
- ✅ Case study showcase links
- ✅ Company overview
- ✅ Tech stack description (Astro, Tailwind, FastAPI, PostgreSQL)
- ✅ GEO optimization notes
- ✅ Contact information
- ✅ Certifications and credentials
Score: 25/25 ✅
Why This Matters
A shocking number of business websites either lack a robots.txt file or misconfigure it in ways that block AI crawlers from accessing content. Even more have never heard of the llms.txt protocol.
This site achieved industry best practice on both fronts — the AI crawler access layer was completely solved.
Win #2: Structured Data Excellence
Schema Detection Results
The site deployed three types of structured data markup, all correctly implemented:
| Schema Type | Status | Details |
|---|---|---|
| Organization | ✅ | Name, logo, URL, description, address, contact info, social links |
| WebSite | ✅ | Name, URL, description, SearchAction (site search) |
| ProfessionalService | ✅ | Name, description, phone, email, address, geo-coordinates, hours, price range |
Score: 9/10 ✅
Technical SEO Checklist — All Green
| Check Item | Status |
|---|---|
| Canonical tag | ✅ Present |
| H1 tag | ✅ Present, semantically correct |
| OG tags | ✅ Complete (og:type/url/title/description/image/site_name/locale) |
| Twitter Card | ✅ summary_large_image |
| Viewport | ✅ Responsive configuration |
| Semantic HTML | ✅ Built-in via Astro |
| Responsive design | ✅ Default Astro behavior |
Why Astro Was the Right Choice
Choosing Astro as the frontend framework was a smart architectural decision:
Performance advantages:
- Static site generation (SSG) delivers near-instant load times
- Standard HTML payloads with gzip compression enabled
- ETag + Last-Modified caching strategy
SEO advantages:
- Server-side rendered (SSR) output, so content is fully visible to crawlers
- Semantic HTML generated automatically
- View transition support for modern UX
Compared to traditional WordPress:
- WordPress: Bloated, plugin-dependent, poor performance baselines
- Astro: Lightweight, native performance, SEO-friendly architecture
This is the kind of technical foundation every modern business website should have.
Problem #1: Content Depth Was Insufficient — AI Citation Became Difficult
Citability Score: 25/100 (Grade F)
| Metric | Value |
|---|---|
| Total content blocks | 1 block |
| Average citation score | 25/100 |
| Optimal-length blocks (134–167 words) | 0 blocks |
| Rating | F — Difficult to cite |
Root Cause Analysis
The homepage contained only 507 words, and the primary content block was a mere 25 words — far below the optimal citation length of 134–167 words that AI engines prefer.
Why This Matters for GEO:
When AI engines (ChatGPT, Claude, Perplexity, Gemini) generate answers, they preferentially cite content that:
- Contains specific data and verifiable claims
- Falls within the 134–167 word block range — long enough to be substantive, short enough to be directly quotable
- Is well-structured with clear topic sentences
This site had the technical infrastructure to be crawled and indexed by AI engines — but once the AI found the content, there was nothing substantial enough to cite.
The paradox: A GEO service provider’s site was perfectly set up to be visited by AI crawlers, but had nothing worth quoting once they arrived.
Recommended Fixes (Priority 1 — 1 to 4 Weeks)
- Add 2–3 citable paragraphs (134–167 words each) to the homepage
- Expand each service section to 800–1,000 words with substantive detail
- Reference industry data from credible sources (Ahrefs, Gartner, SEMrush) to increase information density
- Lead each paragraph with a direct answer to the core question users ask
Problem #2: Security Headers Were Missing
Security Audit Results
| Security Header | Status |
|---|---|
| Strict-Transport-Security (HSTS) | ❌ Missing |
| Content-Security-Policy | ❌ Missing |
| X-Frame-Options | ❌ Missing |
| X-Content-Type-Options | ❌ Missing |
| Referrer-Policy | ❌ Missing |
| Permissions-Policy | ❌ Missing |
Why Security Headers Matter
While security headers don’t directly impact GEO scores, they serve as trust signals — especially for a company that sells technical services:
- HSTS prevents downgrade attacks
- CSP mitigates XSS vulnerabilities
- X-Frame-Options prevents clickjacking
- X-Content-Type-Options prevents MIME-type sniffing
For a technology service provider, missing security headers send a subtle but damaging signal: “We don’t follow our own advice.”
Recommended Fixes (Priority 1 — 1 to 4 Weeks)
Add the following security headers to the nginx configuration:
Strict-Transport-Security: max-age=31536000; includeSubDomainsContent-Security-Policy(appropriately scoped)X-Frame-Options: DENYX-Content-Type-Options: nosniff
Problem #3: Brand Authority Was Underdeveloped
Brand Mention Audit
| Platform | Mentions | Notes |
|---|---|---|
| Wikipedia | ❌ None | New brand, not yet listed |
| Baidu Baike | ❌ None | Not established |
| Zhihu (Chinese Q&A) | ⚠️ Unknown | Needs confirmation |
| ❌ None | No overseas presence | |
| B2B platforms | ⚠️ Few | Early-stage brand |
Score: 8/20 🟡
Why Brand Authority Matters for GEO
Our research across multiple audits has found a correlation coefficient of 0.664 between brand mentions and AI citation rates. This is a strong positive relationship.
Even with perfect technical infrastructure, an unknown brand will be deprioritized by AI engines when selecting sources. The AI doesn’t just look at whether content is accessible — it weighs source authority.
The gap: The site was technically accessible, structurally perfect, but lacked the external brand signals that AI engines use as trust proxies.
Recommended Fixes (Priority 2 — 1 to 3 Months)
- Publish technical articles on Zhihu and industry communities
- Create a Baidu Baike entry for brand recognition in the Chinese market
- Encourage clients to share case studies, creating organic word-of-mouth
- Add a client testimonial wall with specific metrics and outcomes
- Showcase core team credentials, certifications, and industry experience
The Gap Between Technical Infrastructure and Content Authority
This 72/100 case study reveals a critical lesson for the GEO era:
Technical infrastructure and content authority are two separate mountains — and you need to climb both.
What Technical Infrastructure Gives You
- ✅ Crawlability — AI bots can find your site
- ✅ Indexability — AI bots can parse your content
- ✅ Structure — AI bots can understand your content hierarchy
- ✅ Performance — AI bots get fast responses
What Technical Infrastructure Does NOT Give You
- ❌ Citability — AI bots won’t quote thin content
- ❌ Authority — AI bots won’t trust an unknown brand
- ❌ Depth — AI bots won’t learn from shallow pages
- ❌ Differentiation — AI bots won’t choose your content over competitors
The site in this case study had the first column entirely solved. But the second column — content authority — was barely started.
The 17 vs. 72 Comparison
Comparing this site to the 17/100 industrial valve site we audited:
| Dimension | 17-Point Site (Valve Manufacturer) | 72-Point Site (GEO Service) |
|---|---|---|
| Tech stack | ❌ Legacy static HTML | ✅ Modern Astro framework |
| robots.txt | ❌ Missing | ✅ Perfect configuration |
| llms.txt | ❌ Not present | ✅ Fully deployed |
| Schema markup | ❌ None | ✅ 3 schema types |
| Subpages | ❌ All returning 404 | ✅ Complete |
| Content depth | ❌ Under 200 words | ⚠️ 507 words (still insufficient) |
| Security headers | ❌ All missing | ❌ All missing |
| Brand authority | ❌ Zero presence | ⚠️ Early stage |
Key Takeaways
Takeaway #1: Tech stack determines your floor.
- The 17-point site chose the wrong architecture and was crippled from the start.
- The 72-point site chose Astro, nailed the technical layer, and set a high floor.
Takeaway #2: Content depth determines your ceiling.
- The 17-point site had virtually no content — its ceiling was bricked shut.
- The 72-point site had technically correct but shallow content — its ceiling was capped well below its potential.
Takeaway #3: Brand authority is a long game.
- Both sites need brand-building work.
- This is a sustained effort, not a one-time fix.
The Right Path for Enterprise Website Construction
These two case studies — the 17-point valve manufacturer and the 72-point GEO service provider — together define the correct path for building a GEO-ready website.
Step 1: Choose the Right Tech Stack (Sets the Floor)
Right choices:
- ✅ Astro / Next.js / Nuxt.js (modern frameworks)
- ✅ Static site generation (performance-first)
- ✅ Server-side rendering (SEO-first)
Wrong choices:
- ❌ WordPress (bloated, poor baseline performance)
- ❌ Pure client-side rendering (SEO-hostile)
- ❌ Template-based site builders (technical debt)
Step 2: Nail Technical SEO Basics (Non-Negotiable)
Mandatory items:
robots.txt— explicitly allow AI crawlersllms.txt— AI content protocol for structured discovery- Schema markup — structured data for all entity types
- Meta tags — Description, OG, Twitter Card
- Canonical URLs — prevent duplicate content issues
Step 3: Build Content Depth (Determines the Ceiling)
Content requirements:
- Homepage: 1,000–1,500 words of substantive content
- Each service section: 800–1,000 words with specific detail
- Citable paragraphs: 134–167 words each, data-backed
- Industry references: cite specific studies, statistics, and benchmarks
- Case studies: concrete results with measurable outcomes
Step 4: Invest in Brand Authority (Long-Term Compound)
Building channels:
- Industry Q&A platforms (Zhihu, Quora, relevant communities)
- Encyclopedia entries (Wikipedia, Baidu Baike)
- B2B platforms with verified profiles
- Client testimonial wall with specific metrics
- Team credentials and experience showcase
Our Recommended Service Tiers for This Client
Based on the audit findings, we proposed three service tiers to close the gaps.
Tier 1: Astro-Powered Ultimate Website (Digital Infrastructure)
Already completed by the client:
- ✅ Astro framework
- ✅ Static site generation
- ✅ Responsive design
Still needs optimization:
- Security header configuration
- Performance fine-tuning
Tier 2: Traditional Google SEO (Search Traffic)
Target outcomes:
- 5–10 high-intent long-tail keywords
- Top-5 rankings within 3–6 months
Service scope:
- Keyword research and content mapping
- Content strategy development
- Backlink acquisition
- Rank tracking and iterative optimization
Tier 3: AI Engine GEO Optimization (AI Citation)
Already completed by the client:
- ✅ Perfect robots.txt configuration
- ✅ Full llms.txt deployment
- ✅ Structured data schema
Still needs optimization:
- Content depth expansion (the core gap)
- AI-optimized citability paragraphs
- Brand authority building
Conclusion: Tech Is the Prerequisite, Content Is the Differentiator
This 72-point website case study reveals an uncomfortable truth for technically-minded founders:
Getting the tech right is table stakes. Content depth is the real moat.
This site had already achieved industry best practice in the technical layer:
- Astro — a modern, SEO-friendly framework
- Perfect
robots.txt— every AI crawler welcomed - Full
llms.txtdeployment — AI content protocol enabled - Three schema types — structured data comprehensively implemented
But with only 507 words of homepage content and a citability score of 25/100, the AI engines had nothing worth quoting.
Website construction for the GEO era is a system with four layers:
- Tech stack determines the floor
- Technical SEO ensures the foundation
- Content depth defines the ceiling
- Brand authority is the long-term compounder
If you are evaluating your own site’s GEO readiness, start with the technical layer — but do not stop there. The sites that win in the AI citation era will be those that combine technical excellence with genuine content depth.
The GEO service provider in this case study is already ahead of 90% of business websites on the technical side. Closing the content gap is the difference between a 72 and a 90+.
Free Diagnostic
We offer a free GEO-SEO diagnostic service. Enter your website URL to receive:
- GEO visibility score
- Technical SEO issue checklist
- Schema analysis report
- Priority-ranked optimization recommendations
Originally published at 2000m.net · Contact for reproduction permission.
Related Articles
72分网站案例:技术选型正确,但内容深度不足
某GEO技术服务商网站诊断报告,展示Astro建站的技术优势,同时揭示内容深度对AI引用的影响。
企业网站诊断案例:17分网站背后的致命缺陷
某宁波电磁阀制造商网站GEO-SEO诊断报告,揭示企业网站建设的常见误区和技术盲区。
Website Autopsy: The Fatal Flaws Behind a 17/100 Score
A deep-dive GEO-SEO audit of a Ningbo solenoid valve manufacturer reveals the most common and devastating mistakes in industrial B2B website construction — and how to fix them.
首发于 2000m.net
转载请注明出处