A metadata taxonomy is the hierarchical classification system that organizes an organization's custom metadata into related groups. Where metadata is the descriptive information and a controlled vocabulary is the approved values for a field, a taxonomy is the larger structure that decides which fields exist and how they relate. It is the framework that makes search, filtering, and workflows behave consistently.
Why it matters
Without structure, metadata becomes more noise than signal: too many fields, overlapping terms, and no logic a user can predict. A good taxonomy is the difference between metadata that produces results and metadata that produces frustration, and it is the second most important foundation of a healthy program after people.
How it shows up in practice
Building one starts with a snapshot: export the metadata already attached to your priority assets and audit your asset types and use cases. A grocery brand might land on top-level categories like Product, Brand, and Campaign, each with child fields. Stakeholders from each department weigh in so the structure reflects how they search, and the team agrees on single terms ("Blazer," not three variants). The taxonomy is then mapped into the platform as custom fields tied to IPTC where possible, with required fields set so a baseline is always applied. Aprimo and Orange Logic Cortex are examples of platforms built around deep, configurable taxonomies.
Common mistakes
- Trying to "boil the ocean" by covering every possible use case before shipping anything.
- Designing the taxonomy in isolation from the people who will search it.
- Skipping required fields, so the taxonomy is optional in practice.
- Enriching the entire archive instead of the priority assets that deliver value now.
Stacks explains the build in building order from chaos.