metadata is data that describes other data — the labels, fields, and properties that say what a piece of content is, where it came from, and how it should be found, handled, or displayed.
The word joins the Greek-derived prefix meta-, used here in the sense of "about," to data, and it is usually glossed as "data about data." Philip Bagley is credited with coining it in his 1968 work Extension of Programming Language Concepts, where it described data about the structure of other data. It has since spread across computing, library science, publishing, and web search.
Practitioners commonly sort it into kinds. Descriptive metadata — a title, an author, keywords, tags — helps someone find and identify a thing. Structural metadata records how parts fit together, such as the chapters of a book or the fields of a database. Administrative metadata covers provenance, rights, and access. A photo's capture date and camera settings, a library catalog record, and the title and description a web page hands to search engines are all metadata: statements about the content rather than the content itself.
The line between metadata and data is relative rather than fixed. One system's metadata is another system's data — a tag index is metadata to the memos it labels and raw material to the tool that analyzes it. In 📝Search Engine Optimization (SEO), the word usually narrows to the title and description a page exposes to search results, which shape whether a searcher clicks but not what the page actually says.
In 📝MythOS, metadata is everything about a memo that isn't its body — title, tags, agent-facing description, SEO title and description, visibility — and the memo's template owns it. Our SEO runs taught us that restructuring the body moves rankings; metadata-only edits are hygiene.