Your estate is small. The cost of settling what it means grows with it.

Definitions multiply faster than people do

Somebody defines a signup in the product database. Somebody defines it again in the dashboard, because the first definition was hard to find. A third version goes into the deck. Nobody was careless: there was nowhere for the first one to live where the others would have seen it.

A small team absorbs this for a while, because the person who wrote each version is still there and can be asked. That stops being true at exactly the point you are least able to spend a quarter on it.

And every assistant you add works the meaning out from raw schemas, which means the ambiguity you have been carrying quietly starts answering questions for customers.

What you can do

Three entries in the catalog do the work a team this size does not have the people for.

• Get the modelling as a draft

  Domains, models and definitions are drafted across your catalog rather than assembled by hand, so the part that would take an analyst months arrives as something to review.

• Work where you already work

  The semantic core is reachable from the assistant in your editor, so nobody has to open a governance tool to find out what a metric means.

• Grow into dbt rather than out of it

  The semantic models and metrics your dbt project defines are held to the same governance as the rest of the estate, so the transformation layer and the semantic layer do not drift apart as both grow.