# Structured Data (JSON-LD) Schema.org JSON-LD fuels both Google rich results AND AI search citation. ## Why JSON-LD over Microdata/RDFa - Decoupled from markup — easier to maintain. - Single block per page in `` or ``. - Google's preferred format. - Easier to lint and validate. ## Common types per surface | Surface | Primary type | Secondary | |---|---|---| | Product page | `Product` | `Offer`, `AggregateRating`, `Review` | | Article / blog post | `Article` (or `BlogPosting`, `NewsArticle`) | `Person` (author), `Organization` (publisher) | | Landing page | `WebPage` + `Organization` | `BreadcrumbList` | | FAQ section | `FAQPage` | each Q as `Question` + `Answer` | | How-to | `HowTo` | each step as `HowToStep` | | Software / app | `SoftwareApplication` | `Offer`, `AggregateRating` | | Local business | `LocalBusiness` (specific subtype) | `PostalAddress`, `OpeningHours` | | Recipe | `Recipe` | `NutritionInformation` | | Event | `Event` | `Place`, `Offer` | | Video | `VideoObject` | `Clip` | | Comparison page | `Article` + multiple `Product` | `ComparisonTable` (no official type — use `Table` | ## Template: Product ```html ``` ## Template: Article ```html ``` ## Template: Organization (site-wide) Place once, on the homepage: ```html ``` ## Template: FAQPage ```html ``` > FAQPage is heavily used by AI engines for direct answers. High ROI for any product or doc page with common questions. ## Template: SoftwareApplication ```html ``` ## Validation - [Google Rich Results Test](https://search.google.com/test/rich-results) - [Schema.org Validator](https://validator.schema.org) - Lint in CI: `schema-dts` (TypeScript), `pyld` (Python). ## Common pitfalls - Schema doesn't match visible content → Google ignores or penalizes. - Empty fields → invalid; omit instead. - Reviews without real reviews → manipulation flag. - Duplicate schema blocks → use `@graph` to bundle. - Missing required properties for the type — check schema.org docs per type.