Governance decides who may add or merge classes, or the taxonomy inflates without order

What it is

If anyone can “add a section” in the CMS, the tree grows by departmental politics and this week’s campaign. Governance names who may propose, who may approve add or merge, on what grounds, and how often review happens. Without it, inflation is not a natural result of content growth; it is an open permission. Governance is there to stop unordered node proliferation, not to stop all change. Change should come in through one door.

“Let’s discuss it” is not governance. Governance is enforceable permission plus a record that can later answer “who added this class, on which day, why.”

Why it happens

Adding a class has large local benefit (my campaign has an entrance) and global cost spread across every lookup (one more unpredictable neighbour). Individual rationality inflates the tree. Merging’s benefit is global and its cost sits with the team that loses a dedicated entrance, so merges are almost never proposed. Governance puts both operations on one table, judged by the same standards (volume, distinctiveness, duplication of an existing dimension), rather than first-to-mint keeping the slot.

If permission only locks add and not rename or reparent, inflation changes costume: the same classes, privately reattached. Governance covers add, merge, rename, reparent, retire.

Studying it

Treat “who can change the tree” as a variable and watch inflation and predictability.

• Paradigms: compare an open-minting period with an approval period on net new classes, duplication, and tree-test success; audit how many current classes have an approval record. Interview whether reasons for refusal are public.
• Independent variables: approver role, hard rules such as minimum volume, review cadence, whether emergency temporary classes with expiry are allowed.
• Dependent variables: net new classes per year, share of undocumented classes, count of near-synonym pairs in a layer, findability over time.
• Methodological note: inflation can look like “richer nav” in traffic logs. Pair it with predictability; do not treat click spread as health. Governance studies must include refused proposals, not only accepted ones. Missing refusal logs means governance did not happen.

Where it stops holding

Personal workspaces, drafts, and team-private spaces can have looser local classes if they do not leak into global nav and official filters. Emergencies that need a same-day entrance should have a temporary-class path with expiry, not a permanently open mint. Multi-brand or multi-tenant trees are governed inside the tenant; forcing one tree across tenants turns governance into another unordered merge. The controlled-vocabulary owner may be the same person as nav governance, but approval standards for vocabulary nodes and nav nodes differ; do not share one “you can build anything” form.

Applying it

• Write a permission matrix: propose, approve, implement, emergency temporary class. Turn off unreviewed “new section” in the CMS.
• Add/merge requests must state: volume or expected volume, difference from existing classes, which old paths will move.
• Every approval leaves a record. Temporary classes expire into a merge or a takedown.
• Verify by sampling ten nav classes for approver and date. If they cannot be named, they came in through the open door. Then look at a year of net change: if adds dwarf merges and findability is falling, the gate is not working.

Related

• Within the group: G1.11.1 A taxonomy should leave room for future content, not fill up around what is known today · G1.11.2 Premature classes for tiny sets lose meaning as content grows · G1.11.3 The cost of changing architecture rises with inbound links, bookmarks, and settled mental models · G1.11.4 Structural changes need redirects so old paths do not break · G1.11.6 Imbalanced class use is a signal to restructure; do not rely on intuition alone
• Adjacent: G1.10 Controlled vocabularies and synonyms · G1.07 Content inventory and audit · T3.04 Content strategy and information governance
• Search terms: taxonomy governance · category minting · information architecture stewardship