A controlled vocabulary is a fixed, predefined set of approved terms used to fill a given metadata field, so the same concept is always described the same way. It is not the same as keywords in general; it is the governed list that keywords are drawn from. By limiting what users can enter into a field, it keeps inconsistent or random values out.
Why it matters
A full DAM that nobody can search is almost always a language problem, not a content problem. When one person tags an image "Happy" and another tags a near-identical image "Excited," a search for "Happy" never surfaces the second image, so assets go invisible. A controlled vocabulary fixes one agreed term per concept and ends the guessing.
How it shows up in practice
An apparel brand decides that a tailored jacket is always "Blazer," never "Sport Coat" or "Suit Jacket," and puts that term in a drop-down so uploaders cannot improvise. A sports team agrees that the action field uses "Pass," never "Throw," so every quarterback image is findable the same way. A stock-photo company keeps both "Flowers" and "Tulips" because its users search both ways. Most DAM platforms, from Bynder to Acquia DAM, support drop-down or restricted fields specifically so a controlled vocabulary can be enforced at upload.
Common mistakes
- Building the vocabulary from what librarians prefer instead of how users actually search.
- Letting the list grow stale as the organization's language changes.
- Over-tagging, so searches return everything and nothing stands out.
- Duplicating a value that another field already captures.
The Stacks controlled vocabulary guide includes a starter exercise.