DefinedTerm · Glossary
What is JSON-LD
JSON-LD (JavaScript Object Notation for Linked Data) is a W3C standard serialisation format for structured data. It embeds machine-readable facts inside a webpage as a self-contained JSON script block — separate from the visible HTML — so crawlers and AI agents can extract entity information without parsing prose. Google, Bing, and the major AI assistants give JSON-LD the highest reliability rating among the three structured-data formats (JSON-LD, Microdata, RDFa). For UK trade businesses, JSON-LD is the delivery mechanism for LocalBusiness, Service, FAQPage, and Review schema types that power rich results and AI-generated answers.
Full definition
JSON-LD stands for JavaScript Object Notation for Linked Data. It is a W3C Recommendation (current version 1.1, published July 2020) that defines how to express Linked Data using standard JSON syntax. The key design decision is the @context keyword, which maps plain JSON keys to URLs in a shared vocabulary — most commonly https://schema.org — so that any software knowing the vocabulary can interpret the data without custom parsers.
In web deployment, JSON-LD lives inside a <script type="application/ld+json"> element, almost always in the page <head>. Because it is a separate script block and not interleaved with visible content, it can be generated, updated, and validated independently of the page's HTML and CSS. This separation makes JSON-LD the easiest format to maintain in a CMS or headless architecture.
A complete JSON-LD block for a plumbing business might describe multiple nested types — LocalBusiness, PostalAddress, GeoCoordinates, OpeningHoursSpecification, and AggregateRating — all in a single script element.
Why it matters in 2026
Google has explicitly recommended JSON-LD as its preferred structured-data format for several years. The recommendation is rooted in reliability: Microdata and RDFa are interspersed with HTML, so a template change that moves or removes a DOM element can silently break markup. JSON-LD is isolated from layout code.
AI answer engines — including Google AI Overviews, Bing Copilot, and Perplexity — consume structured data at the indexing stage to build entity records. A well-formed JSON-LD block means the AI can read the business name, address, service area, ratings, and opening hours in milliseconds, without any natural-language interpretation. For zero-click answers ("Is there a Gas Safe engineer in Leeds open on Saturday?"), structured data is the primary — sometimes the only — source of the answer.
Rich results enabled directly by JSON-LD types include: Local Pack entries, FAQ accordions in SERPs, Review stars, Sitelinks, and Event listings.
How it works
The minimal JSON-LD structure has three elements:
{
"@context": "https://schema.org",
"@type": "TypeName",
"property": "value"
}
@context declares the vocabulary. @type identifies the entity kind. Every remaining key-value pair is a property defined for that type in the vocabulary. Values can be strings, numbers, booleans, nested objects (using @type again), or arrays.
Multiple entities can share one script block using the @graph array, which is especially useful for pages that describe both a LocalBusiness and a WebPage or BreadcrumbList.
Validation workflow: write the block, run it through validator.schema.org for vocabulary correctness, then through Google Rich Results Test to confirm eligibility for specific rich results.
Difference from Microdata and RDFa
| Attribute | JSON-LD | Microdata | RDFa |
|---|---|---|---|
| Location in page | Separate script block | Inline HTML attributes | Inline HTML attributes |
| Coupling to layout | None | Tightly coupled | Tightly coupled |
| Google preference | Recommended | Supported | Supported |
| Maintainability | High (template-independent) | Low (breaks with HTML edits) | Low (breaks with HTML edits) |
| Multiple entities per page | Easy via @graph | Verbose, error-prone | Verbose, error-prone |
| CMS / headless generation | Straightforward | Complex | Complex |
Related terms
LocalBusiness schema, sameAs, E-E-A-T.
Fuentes
Términos relacionados
- localbusiness-schema
- sameas
- e-e-a-t