Metadata is commonly defined as 'data about data,' but that's too loopy for my brain.
I like to think of metadata as meaning the 'story of data' which can tell things like where the data comes from, how old it is, who made it, when it was last updated, and so on.
The story can get pretty complex, and it can be understood by people and machines with different degrees of meaning through controlled dictionaries, vocabularies, glossaries, thesauri, hierarchies, taxonomies, ontologies, knowledge graphs, and the like.
Metadata gives data a context. It summarizes basic information about data assets which makes data easier to live with, and it helps identify, organize, archive, and preserve digital assets.
Meaning and value depends on context.
Profiles data or data content for determining what the data can be used for. Powers the data lifecycle process for assessing where to store and how long to keep the data (hot, warm, cold).
Used for Identification (terms, languages, author, subject, keyword, title, publisher, type, definition), Quality, Validation, Reliability, Lineage (provenance), and Metrics (freshness, duplicates)
Data about the containers of data; order/organization information, like page order, tables, columns, keys, indexes, and relationships.
Information describing Rights Management (intellectual property, ownership, source, access, digital rights, privacy) and Preservation (created date, file type, file version, and retention)
Models and rules.