Transloadit

What is an Asset Taxonomy?

An asset taxonomy is a structured classification scheme of categories, terms, and relationships used to organize media consistently. It provides a shared vocabulary for describing and locating assets.

How Asset Taxonomies work

A taxonomy defines controlled concepts and the broader, narrower, or related relationships used to classify assets. It differs from free-form tagging because accepted terms have managed identities, labels, and placement in a hierarchy or facet. Indexers attach those concepts to asset records, and search interfaces use them for navigation and filters. Governance handles synonyms, retired terms, and structural changes so integrations retain consistent meaning.

Key facts

  1. Store a term’s identifier on asset records, never its display label. The W3C’s Simple Knowledge Organization System (SKOS) models a concept as a resource with its own identifier, at most one preferred label per language, and alternative labels for synonyms, so a rename or translation leaves existing assignments untouched. Check whether your system’s identifier is itself a path: an Adobe Experience Manager TagID encodes the namespace and the position in the hierarchy, so moving a tag creates a new ID and old references keep resolving only through a redirect the system maintains.
  2. Polyhierarchy places one concept under more than one parent; SKOS allows it, as when a dog concept sits under both mammals and domesticated animals. Consumers then reach the same term by several paths and must deduplicate by identifier, not by path or label. Where a system identifies terms by their path, a second parent means a second identifier, so confirm polyhierarchy support before you model categories.
  3. Whether filtering by a parent term also returns assets tagged only with its children depends on the system. SKOS deliberately leaves skos:broader non-transitive and provides skos:broaderTransitive for uses such as query expansion, while Adobe Experience Manager treats content tagged fruit/apple as also tagged fruit. Decide which behavior your filters and counts need, then test it.
  4. Retiring a term is a data migration, not a label deletion, and systems differ on what deletion does. Drupal core’s issue tracker documents reference fields that keep pointing at deleted entities such as taxonomy terms, WordPress’s wp_delete_term detaches the term from every object that carried it and moves child terms up to its parent, and Adobe Experience Manager’s merge keeps the merged tag as a hidden pointer so content tagged with it resolves to the surviving tag. Before you retire a term, inspect its assignments and child terms: re-parent or retire the children, and re-map or merge assignments where a suitable replacement exists, otherwise remove them deliberately. Then verify the outcome you chose: after a merge that keeps a redirect, old references should resolve to the surviving term; before a physical deletion, nothing should still reference the term.

When Asset Taxonomies matter

Teams design a taxonomy to support filtering, navigation, tagging rules, and predictable retrieval. Categories must remain stable enough for automation while accommodating new content types.

How Transloadit helps with Asset Taxonomies

Transloadit’s /image/describe Robot can supply candidate keywords for classification. It recognizes objects in images and returns them as English words, optionally with a confidence percentage for each label. The recognition provider is chosen automatically unless you set provider to aws or gcp. Treat that output as free-form input, not as terms from your vocabulary: map each label you accept to one of your term identifiers in your own application before storing it on an asset record.

Explore Transloadit’s media management capabilities

Sources

Turn media knowledge into a working pipeline

Connect uploads, processing, AI, storage, and delivery through one declarative API — with the encoding stack, scaling, and format churn handled for you.

Try Transloadit for free