Information organized in a machine-readable format with explicitly identified properties and relationships.
Structured data is information arranged into identifiable fields and relationships that software can process. Rather than relying on a paragraph alone, a record can distinguish an event’s name, date and location as separate properties. The underlying idea is broader than any one search engine or markup language.
In website SEO, structured data often describes the page’s content using Schema.org terms. JSON-LD, Microdata and RDFa are formats Google supports for that purpose, subject to feature-specific documentation. The content being described, its vocabulary and its technical format are related choices but should not be confused.
Good structure does not compensate for bad facts. A well-formed record with the wrong date is still wrong. Plan how information will remain aligned with the page as the business changes it, especially where the same value appears in both visible content and machine-readable fields.
Takeaway: Treat structured data as another representation of real information that needs ownership and maintenance. Review the data model and its accuracy before concentrating on a validation score.
Example
A workshop page shows a new event date, but its structured record still contains the previous date because someone edited only the visible text. Both versions may be technically readable while contradicting each other. A reliable publishing workflow updates the shared source or checks both representations together.
What to check
Ask what each property means, where its value comes from and who updates it. Use appropriate types rather than stuffing keywords into fields. Passing a syntax test is one check; eligibility for a search feature and actual display are separate questions.
For more detail, see Getting started with Schema.org.