Search semantics
Structured data: making a site's meaning machine-readable
Structured data does not improve a page. It describes one — stating the entities and relationships a page already contains in a form machines can read without inference.
Search engines infer a great deal from ordinary HTML, and they are good at it. Structured data exists to remove the inference — to state, unambiguously, that this organisation published this article about this service, and that the page sits here in the hierarchy.
Entities and relationships, not tags
The useful mental model is a graph rather than a checklist. One organisation publishes services and articles; each page sits somewhere in a hierarchy; each article has an author and a date. Marked up correctly, those facts connect.
Organisation, service, article, breadcrumb
Which types belong where
Which schema belongs on which page
| Page | Types | Why |
|---|---|---|
| Home | Organization, WebSite | Establishes who publishes everything else |
| Service page | Service, BreadcrumbList | States the offer and its position in the structure |
| Case study | Article, BreadcrumbList | Dated, authored evidence |
| Insight | Article, FAQPage where visible | Editorial content with real questions on the page |
Generated, not pasted
Hand-pasted JSON-LD is accurate on the day it is added and wrong shortly afterwards. A headline is edited, a date changes, a page is renamed — and the markup still describes the previous version.
When schema is emitted from the same content model that renders the page, that class of error disappears. Publishing an article produces its Article markup, its breadcrumb trail and its sitemap entry from one declaration — the same principle described in internal linking architecture.
Structured data is the last layer of a search architecture, and the one that only pays off when the layers beneath it are already sound. It is built in as part of SEO and content architecture.
Frequently asked questions
- What is structured data?
A standardised vocabulary, usually schema.org expressed as JSON-LD, that states explicitly what a page is about: which organisation published it, what service it describes, where it sits in the site hierarchy.
- Does structured data improve rankings?
Not directly. It improves how accurately a page is understood and makes it eligible for richer presentation in results. Accuracy of understanding is worth having on its own terms.
- Which schema types does a business site need?
Usually Organization and WebSite sitewide, Service on service pages, Article on editorial content, and BreadcrumbList on anything nested. FAQ markup only where the questions are genuinely visible on the page.
- Can structured data describe content that is not on the page?
No. Marking up content a visitor cannot see is a guidelines violation and can remove a site's eligibility for rich results entirely.
- Should schema be added by a plugin?
It is better generated from the same content model that renders the page, so it cannot drift when the page is edited. That is the approach behind SEO and content architecture.