Schema describes; it does not invent
Structured data gives machines an explicit description of information already present on a page. It can reduce ambiguity around an organisation, article, service or breadcrumb, but it does not make unsupported claims true or guarantee a search enhancement.
Begin with a stable organisation identity: name, canonical URL, logo and verified profiles. Use a suitable subtype only when the business genuinely fits it, and avoid listing offices, ratings or service areas that cannot be substantiated.
Choose types that match the page
Apply WebSite and WebPage context across the site, BreadcrumbList to visible navigation paths, Article to editorial content and Service where a page describes a real service. FAQPage should reflect questions and answers users can read on the page and current search policies.
Connect entities using consistent identifiers rather than recreating slightly different organisations on every template. Dates, authors, images and URLs should match the canonical page and update when the visible content changes.
- Create one consistent Organization entity.
- Select the most specific truthful page type.
- Keep marked-up information visible to users.
- Test JSON-LD and canonical URLs in production.
Manage markup as production code
Generate schema from controlled data, validate syntax in automated tests and inspect rendered production pages. Template changes can create duplicate blocks, invalid URLs or mismatches even when the source looked correct.
Review enhancement and indexing reports, but judge success by clearer understanding and eligible presentation rather than the number of schema types installed. More markup is not automatically better markup.