All insights
[ Engineering ]12 December 2024·6 min read

The cost of a database mistake (and how to avoid making one)

Schemas are decisions you have to live with for years. A few hours of modelling up front saves months of rework later.

We see it often: a startup ships fast, the database evolves organically, and three years later the team is paying a tax on every feature — because reporting is impossible, queries are slow, and nobody dares change the schema.

The three modelling mistakes we see most

  • Overloaded “status” columns that try to encode lifecycle, type and priority into a single string.
  • JSON blobs used as a substitute for proper relations — fast to ship, painful to query.
  • Implicit ownership: data that belongs to a customer, but nothing in the schema enforces it.

A few hours, well spent

Two days of domain modelling at the start of a project will save you months of unpicking down the line. It’s the highest-leverage investment in any new build.

Written by Axia Design Studio. Have a project where this might apply? Get in touch.

[ Let's build something ]

Bring us a brief, a sketch, or a vague idea on a napkin.

A free 30-minute call with a senior designer or engineer — no sales script, just useful answers.