Reference Glossary
Structured metadata
Metadata stored in discrete, typed fields — a date field, a controlled picklist, a number — that software can reliably query, filter, and validate, as opposed to free-text description.
Why it matters in a DAM
Structured metadata is what makes faceted filtering in a DAM work at all: a filter for "Rights: Cleared AND Format: Vector" only functions if "Rights" is a defined field with a fixed set of possible values, not a sentence someone wrote. The moment a field becomes free text, that filter breaks silently — "cleared for use," "OK to use," and "rights cleared" all mean the same thing to a person but are three unrelated strings to a database query.
A worked example
Common mistake
Teams migrate legacy free-text captions into a schema field labeled as structured without actually mapping the values into that field's defined option set — the field looks structured in the interface, but it still holds inconsistent free-text strings underneath, so filtering on it silently returns incomplete results.
Whether a metadata field counts as structured depends on the field definition, not the DAM platform generally — the same system can hold a rigorously structured “Usage Rights” dropdown alongside a completely free-text “Notes” field. Structure is a property of the individual field: a defined data type (date, number, single-select, multi-select) with, ideally, a controlled vocabulary constraining what values are allowed.
The payoff for structure is machine reliability: a structured date field can be sorted chronologically and filtered by range without ambiguity, while a free-text date typed as “March 2026,” “03/2026,” or “early spring 2026” defeats any of that. This is why DAM implementations spend disproportionate effort structuring a small number of high-value fields — rights status, expiration date, approval state — even when the bulk of an asset’s description stays as free-text caption or notes.
Structured metadata is also what most downstream integrations depend on. A PIM, CMS, or e-commerce platform pulling assets via API typically expects specific structured fields in specific formats; unstructured description, however rich, generally can’t be consumed programmatically without extraction — which is where structured and unstructured metadata increasingly meet through AI-assisted tagging.
Frequently asked
What makes metadata structured rather than free text?
Structured metadata is stored in discrete, typed fields — a date field, a controlled picklist, a number — that software can reliably query, filter, and validate, as opposed to a sentence of free-text description.
Why does faceted filtering depend on structured metadata specifically?
A filter like 'Rights: Cleared AND Format: Vector' only works if 'Rights' is a defined field with a fixed set of possible values, not a sentence someone wrote. The moment the field becomes free text, that filter breaks silently.
Why can free-text values silently break a filter?
Because 'cleared for use,' 'OK to use,' and 'rights cleared' all mean the same thing to a person but are three unrelated strings to a database query — a structured field with a fixed value avoids that fragmentation entirely.
What's the risk when migrating free-text captions into a field labeled structured?
Teams sometimes migrate legacy free-text captions into a schema field labeled as structured without actually mapping the values into that field's defined option set. The field looks structured in the interface, but it still holds inconsistent free-text strings underneath, so filtering on it silently returns incomplete results.
Does labeling a field structured in the interface guarantee the data inside it is structured?
No — as above, a field can be presented as structured while still holding unmapped free-text values underneath, because the migration never converted the actual data into the field's defined options.
What's an example of a structured field commonly used for DAM filtering?
A rights-status field limited to a fixed set of values like 'Cleared,' 'Restricted,' or 'Expired' is a typical example — its value set is controlled, so a search filter on it returns reliable, complete results every time.