网页结构化数据 Skill

SEO On-Page: Schema / Structured Data

Guides implementation of Schema.org structured data (JSON-LD) for rich snippets, enhanced search results, and Generative Engine Optimization (GEO).

When invoking: On first use, if helpful, open with 1–2 sentences on what this skill covers and why it matters, then provide the main output. On subsequent use or when the user asks to skip, go directly to the main output.

Scope (On-Page SEO)

• Schema markup: Schema.org types for rich results, AI search visibility, and machine-readable content
• Schema.org vs. search engines: Schema.org defines 800+ types; each search engine supports only a subset for rich results

Schema.org vs. Search Engine Support

Schema.org and Google Structured Data are not fully aligned. Schema.org is an open vocabulary (800+ types); Google, Bing, and other engines each support only a curated subset for rich results.

Engine	Support	Notes
Google	Subset only	Only types in Google's search gallery generate rich results. Valid Schema.org markup not in Google's list won't produce enhanced snippets—even if technically correct.
Bing	Subset; different	Supports JSON-LD, Microdata, RDFa, Open Graph. Some types (e.g., Product, Offer) have format-specific support. Check Bing Webmaster docs.
Other engines	Varies	Yandex, DuckDuckGo, AI search tools (Perplexity, etc.) may use Schema.org for understanding even when they don't display rich results.

Practical implication: Implement Schema.org markup for your content type. If Google doesn't show rich results for that type, Bing or AI systems may still use it. Always verify against Google's developer docs for Google-specific rich result eligibility.

Rich Results: Google Support (2026)

High-impact types: Product, Review snippets, Article/News, Video, Recipe, LocalBusiness, Event, Breadcrumb, Sitelinks searchbox, JobPosting.

Limited or context-dependent: FAQ (government/health sites for many queries), Education Q&A, Course, SoftwareApplication, Speakable (news), DiscussionForumPosting.

Deprecated: HowTo (fully removed for all devices as of Sep 2023), COVID data panels, some AMP-only formats, data-vocabulary.org.

Implementation: JSON-LD preferred; include @context, @type, stable @id; ISO 8601 dates; match structured data to visible content. Validate with Rich Results Test. Rich results can increase CTR up to ~35% and improve AI citation.

Schema ↔ SERP Features ↔ Rich Results (Strongly Related)

Schema, SERP features, and rich results are strongly related. Schema is the necessary condition for most rich results. When targeting a SERP feature, implement the corresponding schema type. See serp-features for the full SERP feature list and optimization.

Rich Results vs Featured Snippets

• Rich results: Schema-powered enhancements to standard listings (stars, breadcrumbs, FAQ dropdowns, product info). Appear within organic positions; do not require top-10 rank.
• Featured snippets: Google-extracted answer boxes at position zero. No schema required; content structure matters. Schema (FAQPage, HowTo, Article) can support extraction.

Schema Type	SERP Feature / Rich Result	Notes
FAQPage	PAA, Featured Snippet	FAQ dropdown; Q&A-style snippet. Eligibility restricted for many sites (e.g. government/health)
BreadcrumbList	Breadcrumbs	Path display in result
AggregateRating, Review	Reviews / Stars	Star ratings
HowTo	Formerly rich results (deprecated Sep 2023)	No longer generates Google rich results; Bing/AI may still use
Article	In-Depth Articles, Snippet	Article rich result
VideoObject	Video	Video thumbnail; see video-optimization
Product, Offer	Shopping, Product	Product/shopping results
Recipe	Recipe	Recipe rich result
JobPosting	Google Jobs	Job listings
Event	Event	Event rich result
WebSite + SearchAction	Sitelinks searchbox	Site links for brand queries
Organization, Person	Knowledge Panel	Entity info; see entity-seo

Workflow: 1) Use serp-features to identify target SERP feature; 2) Look up schema type in this table; 3) Implement and validate with Rich Results Test.

Generative Engine Optimization (GEO)

GEO = optimizing content so AI systems (Google AI Overviews, Perplexity, ChatGPT, Gemini) choose, cite, and quote your content in generated answers. Structured data makes content machine-readable; AI engines extract and cite more accurately. Key schema types for GEO: Organization, Person/Author, WebSite, WebPage, FAQPage, HowTo, Article, Product, AggregateRating. See generative-engine-optimization for full GEO strategy.

Initial Assessment

Check for project context first: If .claude/project-context.md or .cursor/project-context.md exists, read it for product type and content.

Identify:

1. Page type: Article, Product, FAQ, Organization, JobPosting, Event, etc.
2. Content: What entities to describe
3. Goal: Rich snippets, AI Overview visibility, Knowledge Panel

Schema Type Classification

Core Types (General Use)

Type	Use case
Organization	Site-wide; company info, logo, sameAs; see placement below
WebSite	Site-wide; search action, site name; pair with Organization on homepage
Article	Blog posts, news, tool intros
BreadcrumbList	Breadcrumb navigation
FAQPage	FAQ sections; triggers PAA-style results
Person	Author info; pairs with Article
ImageObject	Image metadata for rich results
HowTo	Tutorials, step-by-step guides. Note: Google fully deprecated HowTo rich results for all devices (Sep 2023); Bing/AI may still consume HowTo schema

Exclusive Types (Specific Scenarios)

Type	Use case
JobPosting	Recruitment sites, AI Job Matching
Product	E-commerce product pages
Event	Event pages, ticketing (not general blogs)
SoftwareApplication	App pages, tool pages
LocalBusiness	Local business pages
Dataset	Data platforms, datasets
DiscussionForumPosting	Forums, community posts
Quiz	Education, flashcards
MathSolver	Math tools
CaseStudy	Case study pages
Recipe	Recipes, meal plans, cooking instructions

Rule: Use core types for most sites. Use exclusive types only when page content matches (e.g., don't use Event on a blog; don't use JobPosting on a product page).

Organization & WebSite Schema Placement

Where	Organization	WebSite	Notes
Homepage	Minimum	Minimum	Add both Organization and WebSite to homepage at least. Organization describes the entity that owns the site; WebSite enables sitelinks searchbox and site identity.
Root layout / global	Optimal	Optimal	Place in site-wide layout (e.g. layout.tsx, _document, global header/footer) so schema appears on every page. Google uses the first instance found; one instance per site is sufficient.
About page	No	No	About page uses AboutPage schema (page-specific: headline, description, author, about). Organization is entity-level, not page-level—do not confine it to About. See about-page-generator.

Implementation: JSON-LD in <head>; use @id (e.g. https://example.com/#organization) to link Organization ↔ WebSite ↔ WebPage for entity graph. See entity-seo for @id and Knowledge Panel.

Action: Website/Product Type → Schema Mapping

Use this table to recommend which exclusive schema types fit a site. Match the site's content and product type to the most relevant schema. When in doubt, start with core types (Organization, WebSite, Article); add exclusive types only when content clearly matches.

Website / Product type	Recommended exclusive schema	Why
AI meal planner, recipe site, food blog, cooking app	Recipe	Ingredients, instructions, cook time, servings—highly relevant for food/meal content. Google supports Recipe rich results.
Job board, recruitment site, careers page	JobPosting	Title, company, location, salary, employment type. Required for Google Jobs.
Event platform, ticketing, webinar, conference	Event	Date, location, price. Use only on actual event pages.
SaaS, app, Chrome extension, tool, software product page	SoftwareApplication	App name, category, rating, price, OS. Fits product/feature pages.
E-commerce product page	Product	Price, availability, brand, reviews. Use with Offer, AggregateRating.
Forum, community, Reddit-style, Q&A	DiscussionForumPosting	Post content, author, comments. For user-generated discussion.
Data platform, dataset repository, Scale AI / Surge AI	Dataset	Dataset name, creator, license, distribution format. For data catalog pages.
Education site, flashcards, Quizlet-style	Quiz	Question-answer pairs. For educational Q&A content.
Math solver, calculator, equation tool	MathSolver	Math problem input, solution output. For math tools.
Restaurant, local service, store locator	LocalBusiness	Address, hours, NAP. For local SEO.
Case study, customer story page	CaseStudy	Client, outcome, methodology. For B2B case studies.
FAQ page, product FAQ, support FAQ	FAQPage	Question + acceptedAnswer pairs. Triggers PAA-style results.
Tutorial, how-to guide, step-by-step	HowTo	Steps, tools, time. Note: Google fully deprecated HowTo rich results (Sep 2023); Bing/AI may still consume the schema. Consider FAQPage as alternative
News article, press release	NewsArticle	Use instead of Article for news.
Video page, podcast episode	VideoObject / PodcastEpisode	For video/audio content. See video-optimization for VideoObject, thumbnail, key moments.

Examples:

• AI meal planner (e.g., generates weekly meal plans with recipes) → Add Recipe schema to each recipe/meal page; Article or WebPage for landing pages
• AI writing tool → SoftwareApplication on product page; Article on blog
• Recruitment SaaS → JobPosting on job listing pages; SoftwareApplication on product page
• Recipe blog → Recipe on each recipe post; Article for non-recipe posts

Output: When recommending schema, state: (1) which exclusive types fit the site/product, (2) which page types get which schema, (3) core types to add site-wide (Organization, WebSite, BreadcrumbList).

Article / BlogPosting / NewsArticle: Type Selection & Implementation

Choose the most specific type that matches content:

Type	Use case
BlogPosting	Informal blog posts; individual authors; regularly updated
Article	Formal, evergreen content; tool intros; encyclopedic
NewsArticle	Time-sensitive news; recognized publishers

Required properties: headline (max 110 chars), image (min 1200px wide; absolute URL), datePublished (ISO 8601), author (Person or Organization), publisher (Organization with logo).

Recommended: dateModified, description, mainEntityOfPage (canonical URL).

Date display for CTR: Google recommends showing only one date on the page. If both datePublished and dateModified are visible, Google may pick the wrong date for SERP display—Search Engine Land saw ~22% CTR drop. Best practice: show dateModified if it exists, otherwise datePublished. Keep both in JSON-LD; the rule applies to visible date only.

JSON-LD example (BlogPosting):

{
  "@context": "https://schema.org",
  "@type": "BlogPosting",
  "headline": "The Ultimate SEO Checklist for 2025",
  "description": "A complete guide to optimizing blog posts for search and AI.",
  "image": "https://example.com/image.jpg",
  "datePublished": "2025-01-15T09:00:00Z",
  "dateModified": "2025-02-01T14:30:00Z",
  "author": { "@type": "Person", "name": "Jane Doe", "url": "https://example.com/author/jane" },
  "publisher": { "@type": "Organization", "name": "Example", "logo": { "@type": "ImageObject", "url": "https://example.com/logo.png" } }
}

Place in <head> via <script type="application/ld+json">. For article pages, use og:type: article with og:article:published_time, og:article:modified_time, og:article:author. See article-page-generator, open-graph.

BreadcrumbList

For breadcrumb navigation. Schema must match visible breadcrumbs exactly. See breadcrumb-generator for UI, placement, and semantic HTML.

Requirement	Guideline
Format	JSON-LD in <script type="application/ld+json">
URLs	Absolute URLs with https:// for each item
Position	Sequential integers starting from 1
Match	Schema must match visible breadcrumbs exactly

JSON-LD example:

{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://example.com/" },
    { "@type": "ListItem", "position": 2, "name": "Category", "item": "https://example.com/category/" },
    { "@type": "ListItem", "position": 3, "name": "Current Page", "item": "https://example.com/category/current-page/" }
  ]
}

Multiple paths: Google supports multiple BreadcrumbList objects on the same page when a page is reachable via multiple paths (e.g., product in multiple categories). Use an array of BreadcrumbList objects.

Best Practices

Principle	Guideline
Accuracy	Data must match visible page content; never add invisible or misleading data
Completeness	Include all required properties per type
Most specific type	Use NewsArticle over Article when applicable
JSON-LD	Preferred format; place in <script type="application/ld+json">
@id for entities	Use @id for Organization, Person to enable entity linking; see entity-seo
Phased implementation	Add required properties first; then optional for optimization
Validation	Test with Rich Results Test and Schema Markup Validator
inLanguage (multilingual)	Add "inLanguage": "en-US" (IETF BCP 47) to match hreflang; localize names, descriptions, FAQs for rich snippets per locale

Multilingual Schema (inLanguage)

For multilingual sites, add inLanguage to JSON-LD to reinforce language targeting. Align with hreflang values (e.g. "inLanguage": "zh-CN" with hreflang="zh-CN").

Localize schema data: Translate structured data fields (name, description, FAQ acceptedAnswer, etc.) for each locale to improve rich snippet CTR in that language.

Types that support inLanguage: Article, BlogPosting, WebApplication, FAQPage, HowTo, Product, Organization.

Implementation Workflow

1. Analyze page type and content; choose matching Schema type
2. Select format — JSON-LD recommended (Google, Bing, AI tools support it)
3. Write structured data; start with required properties
4. Validate with Rich Results Test, Schema Markup Validator
5. Deploy and monitor via Search Console enhanced reports

Common Errors and Fixes

Error	Fix
Data doesn't match visible content	Schema must describe only what users see
Missing required properties	Check Google/Schema.org docs for each type
Wrong type for page	Don't use Event on non-event pages; don't use JobPosting on product pages
Format/syntax errors	Validate JSON-LD; check quotes, brackets, commas
Over-markup	Mark only relevant content; avoid stuffing unrelated types

Implementation

Next.js (metadata)

export const metadata = {
  other: {
    'script:ld+json': JSON.stringify({
      "@context": "https://schema.org",
      "@type": "Article",
      "headline": "...",
      "description": "...",
      "inLanguage": "en-US",
      "image": "https://example.com/image.jpg",
      "datePublished": "2024-01-01T00:00:00Z",
      "dateModified": "2024-01-15T00:00:00Z",
      "author": { "@type": "Person", "name": "..." },
      "publisher": { "@type": "Organization", "name": "...", "logo": { "@type": "ImageObject", "url": "..." } }
    }),
  },
};

HTML (generic)

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "...",
  "description": "...",
  "inLanguage": "en-US",
  "author": { "@type": "Person", "name": "..." },
  "publisher": { "@type": "Organization", "name": "...", "logo": { "@type": "ImageObject", "url": "..." } }
}
</script>

Validation Tools

Tool	Purpose
Google Rich Results Test	Check if Google can generate rich results
Schema Markup Validator	Validate against Schema.org spec
Search Console	Enhanced reports; monitor validity over time

Output Format

• Action first: Use the Website/Product Type → Schema Mapping table to recommend which exclusive schema fits the site (e.g., AI meal planner → Recipe; SaaS tool → SoftwareApplication)
• Schema type recommendation (core vs. exclusive)
• Page-level mapping: Which pages get which schema
• JSON-LD structure with required properties
• Validation steps
• References: Schema.org, Google Structured Data, Bing Markup

Related Skills

• article-page-generator: Article structure; Article/BlogPosting/NewsArticle schema; date display
• serp-features: Strongly related—schema maps to SERP features; see mapping table above
• faq-page-generator: FAQPage schema; FAQ content structure
• howto-section-generator: HowTo section component (steps, JSON-LD); HowTo vs FAQPage
• breadcrumb-generator: BreadcrumbList schema implementation
• featured-snippet: FAQPage, HowTo for snippets
• video-optimization: VideoObject, video sitemap, thumbnail, key moments
• entity-seo: Organization, Person for entity recognition; @id; Knowledge Panel
• homepage-generator: Organization + WebSite schema on homepage or root layout
• indexing: Google Indexing API for JobPosting, BroadcastEvent