schema-markup（模式标记）

Schema Markup Implementation

Production-grade structured data implementation covering 20+ schema types, rich result eligibility rules, AI search optimization, and CMS-specific deployment patterns. Handles auditing existing markup, implementing new schema, and fixing validation errors.

Table of Contents

• Operating Modes
• Schema Type Selection Matrix
• Implementation Patterns
• Rich Result Eligibility Rules
• AI Search Optimization
• Knowledge Graph Strategy
• CMS Deployment Guide
• Validation and Testing
• Common Errors and Fixes
• Audit Framework
• Output Artifacts
• Related Skills

Operating Modes

Mode 1: Audit Existing Markup

1. Extract all JSON-LD blocks from the page source
2. Validate required vs recommended fields per schema type
3. Cross-reference with Google's current rich result requirements
4. Score completeness 0-100 per schema block
5. Deliver prioritized fix list with corrected JSON-LD

Mode 2: Implement New Schema

1. Identify page type and matching schema types
2. Select primary + supporting schema combination
3. Generate complete, copy-paste-ready JSON-LD populated with page content
4. Advise on placement method (inline head, CMS plugin, server-side rendering)
5. Test before deployment

Mode 3: Fix Validation Errors

1. Map Google Search Console errors to specific fields
2. Identify root cause (missing field, wrong format, content mismatch)
3. Deliver corrected JSON-LD with change log
4. Explain the fix to prevent recurrence

Schema Type Selection Matrix

Primary Schema by Page Type

Stacking Rules

Always add:

• BreadcrumbList to any non-homepage if breadcrumbs exist on the page
• Organization to the homepage (site-wide identity)

Common valid stacks:

• Article + BreadcrumbList + Person + ImageObject (blog posts)
• Product + Offer + AggregateRating + BreadcrumbList (product pages)
• LocalBusiness + OpeningHoursSpecification + GeoCoordinates + Review (local pages)
• HowTo + Article + BreadcrumbList + ImageObject (guides)

Never combine:

• Product on a page that does not sell a product (Google penalizes misuse)
• Multiple Organization blocks for the same entity (combine into one)
• FAQPage on pages where the Q&A is not visible to users

Implementation Patterns

JSON-LD Format (Always Use This)

JSON-LD is the only format worth implementing. Google recommends it, it lives in the <head>, and it does not touch your HTML markup. Microdata and RDFa are legacy -- do not use them for new implementations.

Placement

<head>
  <script type="application/ld+json">
  {
    "@context": "https://schema.org",
    "@type": "Article",
    "headline": "Your Article Title",
    "author": {
      "@type": "Person",
      "name": "Author Name",
      "url": "https://example.com/authors/name",
      "sameAs": ["https://linkedin.com/in/name", "https://twitter.com/name"]
    },
    "datePublished": "2026-01-15",
    "dateModified": "2026-03-01",
    "image": "https://example.com/images/article-hero.jpg",
    "publisher": {
      "@type": "Organization",
      "name": "Company Name",
      "logo": {
        "@type": "ImageObject",
        "url": "https://example.com/logo.png"
      }
    }
  }
  </script>
</head>

Multiple <script type="application/ld+json"> blocks per page are valid. Use separate blocks for unrelated schema types. Nest related types within one block.

Scope Rules

Rich Result Eligibility Rules

Google does not give rich results for all valid schema. These are the current requirements (as of 2026):

Article Rich Result

Product Rich Result

FAQPage Rich Result

HowTo Rich Result

AI Search Optimization

AI search systems (Google AI Overviews, Perplexity, ChatGPT Search, Bing Copilot) use structured data for content understanding, citation decisions, and entity recognition.

Why Schema Matters for AI Search

1. Content type classification -- AI systems use @type to determine if content is a how-to, product listing, FAQ, or opinion piece
2. Citation eligibility -- FAQPage and HowTo schema increase citation likelihood because AI systems can extract structured Q&A and step-by-step content directly
3. Freshness signals -- datePublished and dateModified help AI systems filter by recency
4. Authority signals -- author with sameAs links to known profiles boosts entity recognition
5. Entity connection -- Organization with sameAs links to Wikidata, LinkedIn, and social profiles strengthens entity resolution

AI Search Schema Playbook

Knowledge Graph Strategy

Getting into Google's Knowledge Graph means your entity (person, company, product) is recognized and displayed in panels, AI answers, and cross-referenced searches.

Knowledge Graph Entry Requirements

1. Wikidata entry -- Create or claim your entity on Wikidata.org
2. Wikipedia presence -- A Wikipedia article dramatically increases KG entry probability
3. Consistent NAP -- Name, Address, Phone must be identical across all citations
4. sameAs network -- Organization schema must link to all official profiles

sameAs Best Practices

{
  "@type": "Organization",
  "name": "Your Company",
  "url": "https://yourcompany.com",
  "sameAs": [
    "https://www.wikidata.org/wiki/Q12345678",
    "https://en.wikipedia.org/wiki/Your_Company",
    "https://www.linkedin.com/company/yourcompany",
    "https://twitter.com/yourcompany",
    "https://www.crunchbase.com/organization/yourcompany",
    "https://github.com/yourcompany"
  ]
}

Order of importance for sameAs links:

1. Wikidata (strongest entity signal)
2. Wikipedia
3. LinkedIn
4. Official social profiles
5. Industry directories (Crunchbase, G2, Capterra)

CMS Deployment Guide

WordPress

• Yoast SEO / Rank Math: Auto-generate Article, Organization, BreadcrumbList. Add custom schema via their blocks for HowTo and FAQPage.
• Custom schema: Add via wp_head action hook or a custom plugin.
• Avoid: Plugins that inject schema via JavaScript (Google may not render it).

Webflow

• Per-page: Add custom code in page settings > Custom Code > Head
• CMS-driven: Use Webflow CMS to generate dynamic JSON-LD via embedded code blocks with CMS field references
• Site-wide: Add Organization schema in Project Settings > Custom Code > Head

Shopify

• Product schema: Auto-generated by most themes. Verify it includes Offer and AggregateRating.
• Article/Blog schema: Usually missing -- add manually via theme.liquid or a schema app.
• Organization: Add to theme.liquid <head> section.

Next.js / Custom React

• Server-side rendering: Generate JSON-LD in the page component and render in <Head>.
• next-seo library: Provides schema components for common types.
• Dynamic pages: Generate schema from your data layer, not hardcoded.

Static sites (Hugo, Jekyll, Gatsby)

• Template-level: Add JSON-LD to layout templates using template variables.
• Per-page: Use frontmatter data to populate schema fields dynamically.

Google Tag Manager (GTM)

• Warning: GTM-injected schema is often NOT indexed by Google because it renders client-side after JavaScript execution.
• Use only when: No other option exists (no CMS access, no dev resources).
• Better alternative: Server-side injection via CMS or template engine.

Validation and Testing

Three-Layer Validation

Test every schema implementation with all three tools before deployment:

Validation Workflow

1. Pre-deployment: Rich Results Test + Schema.org Validator on the rendered HTML
2. Post-deployment (day 1): Check page is crawled via URL Inspection tool in GSC
3. Post-deployment (week 2-4): Check Enhancements section in GSC for errors at scale
4. Ongoing (monthly): Monitor GSC Enhancements for new errors from content updates

Common Errors and Fixes

Audit Framework

Audit Scorecard (0-100)

Priority Classification

Output Artifacts

Related Skills

• seo-audit -- For full technical and content SEO audits spanning beyond structured data. Use when the problem is broader than schema.
• site-architecture -- For URL structure and navigation. Use when architecture is the root cause, not schema.
• programmatic-seo -- For sites with thousands of pages that need schema at scale. Schema patterns feed into pSEO template design.
• content-creator -- For content creation. Use before implementing Article schema to ensure content quality.

Troubleshooting

Success Criteria

• Rich result eligibility: 100% of content pages with appropriate schema types eligible for rich results per Google Rich Results Test
• Validation pass rate: Zero errors in Google Search Console Enhancements reports across all schema types
• Rich result CTR boost: Structured data pages achieving 20-35% higher CTR than non-structured pages (2026 benchmark from SearchPilot testing)
• AI citation impact: FAQPage and HowTo schema present on all informational content pages to maximize AI extraction
• Entity recognition: Organization schema with 5+ sameAs links deployed site-wide; brand appearing in Knowledge Graph
• Coverage breadth: Schema implemented on 95%+ of indexable pages (BreadcrumbList minimum, content-specific types on relevant pages)
• Freshness accuracy: dateModified updated within 24 hours of every content change across all Article schema

Scope & Limitations

In scope:

• Schema type selection and JSON-LD implementation for 20+ schema types
• Rich result eligibility verification and optimization
• AI search visibility through structured data
• Knowledge Graph entity optimization
• Schema validation, testing, and error resolution
• CMS-specific deployment guidance (WordPress, Webflow, Shopify, Next.js)

Out of scope:

• Content creation for schema-eligible pages (use Content Production)
• Technical SEO beyond structured data (use SEO Audit)
• Microdata or RDFa implementations (JSON-LD only — Google's recommendation)
• Custom schema.org extensions or vocabulary proposals
• Server-side rendering implementation
• CMS plugin development

Known limitations:

• Google does not guarantee rich results even with valid schema — authority and content quality also factor in
• GTM-injected schema is frequently not indexed — server-side deployment is required for reliability
• Schema.org spec updates faster than Google's support — not all valid types generate rich results
• Rich result types can be deprecated with minimal notice (e.g., HowTo rich results were restricted in 2023)
• Structured data CTR impact varies by industry and SERP features present

Scripts

# Validate JSON-LD schema from a file or URL
python scripts/schema_validator.py --file schema.json --json

# Audit a page for schema coverage and completeness
python scripts/schema_validator.py --html page.html --verbose

# Generate JSON-LD templates for common page types
python scripts/schema_generator.py --type Article --title "My Post" --author "Jane" --json