schema.org

The collaborative structured-data vocabulary for the web — a shared set of types and properties for describing entities, relationships, and actions. Founded by Google, Microsoft, Yahoo, and Yandex and run through an open W3C-style community process. Effectively a lightweight, mass-adopted web ontology: the pragmatic, deployed end of knowledge-representation.

What it is

Encodings: RDFa, Microdata, and JSON-LD — the last being the dominant form on the modern web and an rdf serialization, which is how schema.org sits on the RDF foundation rather than beside it.
Scale: 45M+ domains and 450B+ objects — by deployment, the most successful piece of the semantic-web project by a wide margin.
Primary payoff: search rich-results / structured data eligibility and feeding search engines’ knowledge graphs — the concrete incentive that drove adoption where heavier ontologies stalled.

Why it matters here

schema.org is the spoke’s evidence for the central thesis (synthesis): the semantic web succeeded not through rich OWL reasoning but through a shallow, shared vocabulary with an obvious payoff and low authoring cost. It’s the deployment counterweight to rdf‘s “powerful but heavy” critique.

What deployment looks like up close (2026-07-28)

merchant-listing-structured-data is the first source here showing the vocabulary at the point of use, and two things stand out.

The required surface is tiny. To qualify a product page for Google’s shopping surfaces you need five properties: name, image and offers on Product, then price and priceCurrency on the nested Offer. Everything else — ratings, brand, sku/mpn/gtin, availability, itemCondition, shippingDetails, hasMerchantReturnPolicy, strikethrough and member pricing — is recommended, each unlocking a specific visual treatment. Publishers adopt the floor because it’s nearly free and climb only as far as a visible payoff justifies. That gradient is the low-authoring-cost thesis made literal.

The vocabulary also isn’t where the rules are. Eligibility comes with conditions schema.org never mentions: the page must sell the product, rich results “only support pages that focus on a single product,” each currency needs its own URL, the page must be crawlable, and content policy excludes regulated goods. schema.org defines what you can say; the consumer defines what saying it gets you, and publishers implement against the second document. A shared vocabulary with an unshared payout is governed by whoever pays.

The consumer is changing (2026-07-28)

Two sources arriving the same day bet that the next consumer of this vocabulary is an AI system rather than a search crawler, and that the delivery model has to change for it: yoast-schema-aggregation merges a site’s per-page graph into one deduplicated REST endpoint, and entitymap proposes a separate entitymap.json file borrowing schema.org’s own @type and sameAs. Both move the description out of the page and to a predictable URL. Google’s first-party position is that structured data “isn’t required for generative AI search” — see structured-data-for-ai for the tension.

Reflexive note

schema.org is the @type vocabulary this entire hub runs on — every spoke’s pages carry schema.org-typed JSON-LD frontmatter. Documenting it here means the wiki describes its own typing model. (Cross-cutting adjacency: schema markup is also the canonical structured-data signal the search-marketing spoke’s SEO/GEO thread cares about.)

Tier

T1 — primary source (the vocabulary’s own site). url-only ingest.

knowledge-representation · rdf · ontology-dev-tools · knowledge-graph · merchant-listing-structured-data · structured-data-for-ai · yoast-schema-aggregation · entitymap · synthesis