SEO Structure Architect
Overview
Structure is how search engines and AI systems parse your content. Clear hierarchy, complete schema, and logical internal linking determine whether your content gets indexed, cited, and ranked. 65% of pages cited by Google AI Mode include structured data.
Schema Markup
Always use JSON-LD (in <head>). Google recommends it, new features ship to JSON-LD first, and it has the lowest error rate.
Priority Schema Types (2026)
CANONICAL OWNER. This table and the Deprecated table below are the single
source of truth for schema status across the SEO/ecommerce family.
ai-search-optimizer and seo-serp-optimizer rank schema for their purpose but
must not restate status here — they point at this file.
Rationale: three partial copies of these tables previously existed; FAQPage
drifted into three different verdicts, and the copies still disagree by omission
(one lacked WebSite and MerchantReturnPolicy entirely). Verify with
python3 ~/.claude/skills/_meta/gates.py G_CLAIM_FRESHNESS --claim-mode strict.
| Schema |
Where |
Impact |
Notes |
Product + Offer |
Every product page |
CRITICAL |
Include: name, image, description, brand, SKU, price, availability, reviews. Products with complete schema are preferentially cited by AI shopping |
Organization |
Site-wide (once) |
CRITICAL |
sameAs to all official profiles. Entity foundation |
LocalBusiness/ComputerStore |
Site-wide (once) |
HIGH |
NAP, geo, areaServed, openingHours |
BreadcrumbList |
All pages |
HIGH |
Up to 30% CTR boost on desktop (removed from mobile SERPs Jan 2025) |
WebSite |
Site-wide (once) |
HIGH |
Controls site name display in SERPs |
AggregateRating |
Product pages |
HIGH |
Star ratings boost CTR 20-35% |
FAQPage |
Product/category/guide pages |
LOW |
Rich result fully removed May 2026 (no site qualifies). Still valid schema, still valuable for AI parsing |
MerchantReturnPolicy |
Site-wide |
MEDIUM |
Requires returnPolicyCountry. Trust signal |
Deprecated — Do NOT Implement for Rich Results
| Schema |
Status |
HowTo |
Fully deprecated (all sites) |
FAQPage |
Fully deprecated 7 May 2026 — no rich result for any site (was gov/health-only from Aug 2023). Search Console reporting removed Jun 2026, API Aug 2026. Still valid schema — keep for AI parsing, expect zero SERP real estate |
Q&A |
Deprecated Jan 2026 |
SitelinksSearchBox |
Deprecated Jan 2026 |
Dataset (for search) |
Deprecated Jan 2026 |
SpecialAnnouncement |
Deprecated Jan 2026 |
PracticeProblem |
Deprecated Jan 2026 |
Schema Rules
- Never implement empty schema — worse than no schema at all
- Populate ALL relevant properties before adding
- Validate with Google Rich Results Test before deployment
sameAs URLs must be valid and accessible — 404s hurt entity recognition- For custom-built PCs without GTINs: use
mpn with your SKU system, set identifier_exists: false
Header Hierarchy
- One H1 per page matching the primary topic/entity
- H2s for main sections — question-format headings boost AI extraction and PAA eligibility
- H3s for subsections with related terms
- Never skip levels (H1 → H3 directly)
- AI systems parse content into chunks using headings — logical hierarchy = better AI citation
Internal Linking
Topic clusters remain the gold standard. Zyppy's 23M internal link study shows pages with 40-44 internal links get ~4x more traffic than pages with 0-4.
Hub-and-Spoke Model
- 1 pillar page (hub) links to 6-12 spoke pages
- Each spoke links back to hub + cross-links to 2-3 related spokes
- Each spoke should have 3-6 internal links total
- Fewer than 6 spokes = lacks coverage breadth; more than 12 = dilutes focus
Linking Rules
- Contextual links in body content carry more weight than nav/footer links
- Descriptive anchor text (never "click here" or "learn more")
- Diversify anchor text pointing to same page
- 2-5 contextual links per 1,000 words
- Total links per page under 150
Site Architecture
- 3-click max from homepage to any important page
- Homepage → Category → Subcategory → Product
- Click depth directly correlates with crawl efficiency — pages beyond 3 clicks see reduced crawl rates
- Flat URLs (
/product-name) are flexible; moderate hierarchy (/category/product-name) is practical for e-commerce
- Faceted/filter URLs need explicit crawl control — parameter combinations can explode crawl paths and create thin duplicates. See
woocommerce-faceted-navigation for the index-worthy-vs-thin allowlist and the canonical/noindex/robots decision matrix.
Core Web Vitals
| Metric |
Good |
Needs Improvement |
Poor |
| LCP (Largest Contentful Paint) |
< 2.5s |
2.5-4.0s |
> 4.0s |
| INP (Interaction to Next Paint) |
< 200ms |
200-500ms |
> 500ms |
| CLS (Cumulative Layout Shift) |
< 0.1 |
0.1-0.25 |
> 0.25 |
- INP replaced FID in March 2024 — measures full interaction responsiveness, not just input delay
- CWV is a tiebreaker in competitive niches, not a primary factor
- Pages at position 1 are 10% more likely to pass CWV thresholds than position 9
- Every 100ms of latency costs ~1% in sales; 53% of mobile users abandon pages >3 seconds
- LCP >3s = 23% more traffic loss in December 2025 core update
XML Sitemaps
- Max 50,000 URLs per file, 50MB uncompressed. Use sitemap index for larger sites
- Google ignores
<priority> and <changefreq> completely — only <lastmod> matters
<lastmod> must accurately reflect genuine content changes — fake dates hurt- Only include indexable, canonical, 200-status URLs
Anti-Patterns
| Don't |
Why |
| Use Microdata/RDFa for new implementations |
JSON-LD is simpler, lower error rate |
| Implement HowTo schema expecting rich results |
Fully deprecated |
| Implement FAQ schema expecting rich results |
Fully deprecated May 2026 — no rich result for any site |
| Build deep architecture (4+ clicks) |
Kills crawlability |
Set <priority>/<changefreq> in sitemaps |
Google ignores them completely |
| Use strict silos with zero cross-linking |
Outdated; contextual cross-links between related topics are beneficial |
| Have thin content anywhere on site |
Helpful Content is site-wide — weak sections hurt everything |
1---2name: seo-structure-architect3description: Use when implementing schema markup, optimizing header hierarchy, building internal linking structures, planning site architecture, or checking Core Web Vitals. Covers JSON-LD schema, topic clusters, breadcrumbs, sitemaps, and page structure for both traditional and AI search.4---56# SEO Structure Architect78## Overview910Structure is how search engines and AI systems parse your content. Clear hierarchy, complete schema, and logical internal linking determine whether your content gets indexed, cited, and ranked. 65% of pages cited by Google AI Mode include structured data.1112## Schema Markup1314**Always use JSON-LD** (in `<head>`). Google recommends it, new features ship to JSON-LD first, and it has the lowest error rate.1516### Priority Schema Types (2026)1718> **CANONICAL OWNER.** This table and the Deprecated table below are the **single19> source of truth for schema status** across the SEO/ecommerce family.20> `ai-search-optimizer` and `seo-serp-optimizer` rank schema for *their* purpose but21> **must not restate status here** — they point at this file.22> Rationale: three partial copies of these tables previously existed; `FAQPage`23> drifted into three different verdicts, and the copies still disagree by omission24> (one lacked `WebSite` and `MerchantReturnPolicy` entirely). Verify with25> `python3 ~/.claude/skills/_meta/gates.py G_CLAIM_FRESHNESS --claim-mode strict`.2627| Schema | Where | Impact | Notes |28|--------|-------|--------|-------|29| `Product` + `Offer` | Every product page | CRITICAL | Include: name, image, description, brand, SKU, price, availability, reviews. Products with complete schema are preferentially cited by AI shopping |30| `Organization` | Site-wide (once) | CRITICAL | sameAs to all official profiles. Entity foundation |31| `LocalBusiness`/`ComputerStore` | Site-wide (once) | HIGH | NAP, geo, areaServed, openingHours |32| `BreadcrumbList` | All pages | HIGH | Up to 30% CTR boost on desktop (removed from mobile SERPs Jan 2025) |33| `WebSite` | Site-wide (once) | HIGH | Controls site name display in SERPs |34| `AggregateRating` | Product pages | HIGH | Star ratings boost CTR 20-35% |35| `FAQPage` | Product/category/guide pages | LOW | Rich result **fully removed May 2026** (no site qualifies). Still valid schema, still valuable for AI parsing |36| `MerchantReturnPolicy` | Site-wide | MEDIUM | Requires returnPolicyCountry. Trust signal |3738### Deprecated — Do NOT Implement for Rich Results3940| Schema | Status |41|--------|--------|42| `HowTo` | Fully deprecated (all sites) |43| `FAQPage` | **Fully deprecated 7 May 2026** — no rich result for any site (was gov/health-only from Aug 2023). Search Console reporting removed Jun 2026, API Aug 2026. Still valid schema — keep for AI parsing, expect zero SERP real estate |44| `Q&A` | Deprecated Jan 2026 |45| `SitelinksSearchBox` | Deprecated Jan 2026 |46| `Dataset` (for search) | Deprecated Jan 2026 |47| `SpecialAnnouncement` | Deprecated Jan 2026 |48| `PracticeProblem` | Deprecated Jan 2026 |4950### Schema Rules5152- **Never implement empty schema** — worse than no schema at all53- Populate ALL relevant properties before adding54- Validate with Google Rich Results Test before deployment55- `sameAs` URLs must be valid and accessible — 404s hurt entity recognition56- For custom-built PCs without GTINs: use `mpn` with your SKU system, set `identifier_exists: false`5758## Header Hierarchy5960- **One H1 per page** matching the primary topic/entity61- **H2s for main sections** — question-format headings boost AI extraction and PAA eligibility62- **H3s for subsections** with related terms63- **Never skip levels** (H1 → H3 directly)64- AI systems parse content into chunks using headings — logical hierarchy = better AI citation6566## Internal Linking6768**Topic clusters remain the gold standard.** Zyppy's 23M internal link study shows pages with 40-44 internal links get ~4x more traffic than pages with 0-4.6970### Hub-and-Spoke Model7172- 1 pillar page (hub) links to 6-12 spoke pages73- Each spoke links back to hub + cross-links to 2-3 related spokes74- Each spoke should have 3-6 internal links total75- Fewer than 6 spokes = lacks coverage breadth; more than 12 = dilutes focus7677### Linking Rules7879- Contextual links in body content carry more weight than nav/footer links80- Descriptive anchor text (never "click here" or "learn more")81- Diversify anchor text pointing to same page82- 2-5 contextual links per 1,000 words83- Total links per page under 1508485## Site Architecture8687- **3-click max** from homepage to any important page88- Homepage → Category → Subcategory → Product89- Click depth directly correlates with crawl efficiency — pages beyond 3 clicks see reduced crawl rates90- Flat URLs (`/product-name`) are flexible; moderate hierarchy (`/category/product-name`) is practical for e-commerce91- Faceted/filter URLs need explicit crawl control — parameter combinations can explode crawl paths and create thin duplicates. See `woocommerce-faceted-navigation` for the index-worthy-vs-thin allowlist and the canonical/noindex/robots decision matrix.9293## Core Web Vitals9495| Metric | Good | Needs Improvement | Poor |96|--------|------|-------------------|------|97| **LCP** (Largest Contentful Paint) | < 2.5s | 2.5-4.0s | > 4.0s |98| **INP** (Interaction to Next Paint) | < 200ms | 200-500ms | > 500ms |99| **CLS** (Cumulative Layout Shift) | < 0.1 | 0.1-0.25 | > 0.25 |100101- INP replaced FID in March 2024 — measures full interaction responsiveness, not just input delay102- CWV is a **tiebreaker** in competitive niches, not a primary factor103- Pages at position 1 are 10% more likely to pass CWV thresholds than position 9104- Every 100ms of latency costs ~1% in sales; 53% of mobile users abandon pages >3 seconds105- LCP >3s = 23% more traffic loss in December 2025 core update106107## XML Sitemaps108109- Max 50,000 URLs per file, 50MB uncompressed. Use sitemap index for larger sites110- **Google ignores `<priority>` and `<changefreq>`** completely — only `<lastmod>` matters111- `<lastmod>` must accurately reflect genuine content changes — fake dates hurt112- Only include indexable, canonical, 200-status URLs113114## Anti-Patterns115116| Don't | Why |117|-------|-----|118| Use Microdata/RDFa for new implementations | JSON-LD is simpler, lower error rate |119| Implement HowTo schema expecting rich results | Fully deprecated |120| Implement FAQ schema expecting rich results | Fully deprecated May 2026 — no rich result for any site |121| Build deep architecture (4+ clicks) | Kills crawlability |122| Set `<priority>`/`<changefreq>` in sitemaps | Google ignores them completely |123| Use strict silos with zero cross-linking | Outdated; contextual cross-links between related topics are beneficial |124| Have thin content anywhere on site | Helpful Content is site-wide — weak sections hurt everything |
joogy06