intermediate
1 min answer
Finance, sales and product each define "active customer" differently, and all three are correct for their purpose. How should the glossary handle this?
Show the full answer Hide the answer
The mistake: forcing a single definition
Attempting to make three teams agree on one definition produces a long negotiation ending in either a compromise that serves nobody or a decision that one team ignores. The definitions differ because the decisions differ — a revenue recognition question and a product engagement question are genuinely not the same question.
The workable approach
Name them distinctly and define each precisely. "Billing-active customer", "engagement-active customer", "contract-active customer" — three terms, three definitions, three owners, each traceable to the query that computes it.
Then:
- Ban the ambiguous term in reporting. If a dashboard says "active customers" with no qualifier, it is a defect.
- Record the relationship between them, so someone comparing two numbers can see why they differ instead of filing a data quality issue.
- Bind each definition to executable logic in the semantic layer, so the definition and the computation cannot drift apart. A glossary maintained separately from the code is a glossary that is wrong.
- Assign an owner per term, empowered to decide, because definitional disputes without an arbiter run forever.
The general principle
A glossary's job is to eliminate ambiguity, not to eliminate variety. Multiple precise definitions are healthy; one vague definition everyone interprets differently is the actual problem.