Structured data (JSON-LD)
Machine-readable page metadata (schema.org) embedded in pages so crawlers and LLMs can reliably interpret content types, entities, and relationships.
Last updated: December 12, 2025
Definition
Machine-readable page metadata (schema.org) embedded in pages so crawlers and LLMs can reliably interpret content types, entities, and relationships.
In practice
- Keep JSON-LD aligned with what the page actually says (no drift between UI and structured data).
- Use consistent headings and section structure so both humans and machines can scan reliably.
Common mistakes
- Adding structured data that doesn’t match page content (hurts trust and parsing).
- Forgetting to include basics (breadcrumbs, publish/updated dates, canonical URL).
Related terms
- Programmatic content (in GTM30) — A structured system (not a pile of long-tail pages) that turns your website into: a discoverability layer, a single source of truth for your narrative, and a data layer for automation, CRM, and LLMs.
- Canonical page / canonical link — The single page you want threads, content, ambassadors, and internal tools to link to for a topic — so knowledge compacts instead of fragmenting across repeated answers.
- Template spine — A consistent page structure for each page type (use case, glossary, KB, etc.) so users and machines know where to find definitions, steps, mistakes, and related links.
- LLM readiness / LLM visibility — Designing pages so LLMs can retrieve and summarize accurate answers from canonical, structured sources (clear headings, consistent sections, and entity-keyed content).