An organisation deploys a data catalogue and nobody uses it. What makes a catalogue useful rather than an inventory?
Show the full answer Hide the answer
Why it is unused
- It answers a question nobody asked. A list of tables is not what an analyst needs; they need to know which table answers their question and whether they can trust it.
- It is out of date, because it was populated manually and the estate changed.
- It is not where the work happens. An analyst writing a query does not visit a separate tool first.
What makes it useful
- Populated automatically from the systems themselves — schemas, lineage, freshness, usage — rather than by hand. A manually maintained catalogue is wrong within a quarter.
- Usage data: which tables are actually queried, by whom, how often. This is the single most useful signal for an analyst deciding which of four similar tables to use — the one everyone else uses is probably the right one.
- Lineage, so a number can be traced to its inputs and the effect of an upstream change can be assessed.
- Ownership, with a real person, so a question has somewhere to go.
- Freshness and quality signals inline, so trustworthiness is visible at the point of choosing.
- Integrated into the query tool, so it appears where the work happens rather than requiring a detour.
The signal that distinguishes value from inventory
Whether it answers "can I trust this and who do I ask". A catalogue that lists tables and their columns provides what the database already provides. A catalogue that says this table is owned by that person, refreshed hourly, used by forty analysts, derived from these sources, and last failed a quality check on Tuesday is providing something the database does not.
The governance connection
A catalogue is the precondition for governance rather than an outcome of it. Retention policies, sensitivity classification, access decisions and deletion propagation all require knowing what data exists and where it flows — and you cannot apply a policy to data whose existence is unknown.
The derived copies are where the unknowns live, which is why automated lineage matters more than the inventory of primary tables.