Schema.org

The vocabulary is organized as a hierarchy of types with multiple inheritance, anchored by the root type Thing. Each type inherits properties from its parent types, creating a rich, extensible model. For example, a LocalBusiness inherits from both Organization and Place, combining properties like address and openingHours with name and telephone.

• Root Type: Thing
• Core Branches: CreativeWork, Event, Organization, Person, Place, Product, Action, Intangible
• Inheritance: Types can inherit from multiple super-types
• Extensions: Hosted extensions exist for specialized domains like auto, bib, and health-lifesci

Schema.org vocabulary is syntax-agnostic, but three primary serialization formats are supported for embedding structured data in web pages. JSON-LD is Google's recommended format because it is injected as a standalone script block without interleaving with HTML markup, making it easier to manage and less prone to template breakage.

• JSON-LD (Recommended): A JavaScript object in a <script type="application/ld+json"> tag, decoupled from HTML
• Microdata: HTML attributes (itemscope, itemprop) woven directly into existing markup
• RDFa: Attribute-based syntax extending HTML5 for richer Linked Data expressions
• Key Advantage of JSON-LD: Can be dynamically generated and injected without altering the DOM structure of the page

Schema.org enables a shift from keyword-based indexing to entity-based understanding. By marking up a page about a specific product, you are not just providing strings; you are declaring a Product entity with properties like sku, brand, offers, and aggregateRating. This allows search engines to disambiguate entities and populate Knowledge Graph panels with factual, structured data.

• Entity Disambiguation: Use sameAs to link to authoritative Wikidata or Wikipedia entries
• Rich Results: Powers product carousels, recipe cards, job postings, and event listings in SERPs
• Fact Extraction: Enables search engines to answer specific queries directly from your markup
• Example: A Book entity with isbn and author properties confirms the canonical identity of the work

Beyond static description, Schema.org defines an Actions vocabulary that allows a website to describe the potential actions a user can take. This transforms a web page from a passive document into an interactive capability declaration. An Action type connects an agent to an object, enabling search engines to surface deep-links directly to specific tasks.

• Core Action Types: SearchAction, OrderAction, ReserveAction, WatchAction
• Entry Point: Defines the URL endpoint where the action can be performed
• Potential Action: Declares what a user can do directly from the search result
• Example: A recipe site declaring a CookAction with the recipe as the object, enabling voice assistant integration

Implementing Schema.org markup is the prerequisite for achieving rich results in Google Search—visually enhanced listings that include stars, images, carousels, and other interactive elements. However, markup alone does not guarantee a rich result; the structured data must pass validation and the page must meet Google's quality thresholds.

• Validation Tool: Google's Rich Results Test validates both syntax and eligibility
• General Validator: Schema.org's own validator checks pure schema compliance
• Common Rich Result Types: Product, Recipe, Article, Event, FAQ, HowTo, JobPosting, Course
• Critical Rule: Markup must be a truthful representation of the visible page content; hidden or misleading markup is a violation

Before JSON-LD became dominant, structured data was embedded directly into HTML attributes. Microdata uses itemscope, itemtype, and itemprop attributes to annotate content within its existing markup. RDFa (Resource Description Framework in Attributes) extends this with richer attribute vocabulary.

• Microdata: Tightly couples data to visible HTML, making maintenance harder
• RDFa: Supports more complex vocabularies beyond Schema.org, including Dublin Core and FOAF
• Legacy Status: Still parsed by search engines but rarely used in new implementations

Google's Knowledge Graph is the machine-readable encyclopedia that consumes Schema.org markup to build entity understanding. It maps relationships between people, places, organizations, and things, powering Knowledge Panels and disambiguating queries.

• Entity Reconciliation: Schema.org markup helps Google match your entities to its canonical Knowledge Graph IDs
• sameAs Property: Use this to explicitly link your entity to authoritative sources like Wikidata, Wikipedia, or official social profiles
• Impact: Strong entity alignment improves eligibility for Knowledge Panels and branded search features

Beyond traditional search, Schema.org extends into voice assistants and conversational AI. The Speakable type identifies content sections suitable for text-to-speech playback, while Action types define how digital assistants can interact with your services.

• Speakable: Mark up key passages for Google Assistant to read aloud from news articles
• EntryPoint: Define how an action can be invoked via voice commands
• PotentialAction: Specify actions users can take directly from search results, like booking or ordering