A JSON column stores flexible structured data inside one database field, letting a product add nested fields without waiting for a schema migration. It is a release valve, not proof that the data lacks structure: its purpose depends on whether it holds user answers, integration data, or something else. section 7.6
Read a JSON column as a product decision:
- Use responses for a booker's custom form answers when different forms would otherwise require repeated migrations. section 7.6
- Use metadata for engineer-facing, integration-specific data rather than presenting it as the booking's primary business fields. section 7.6
- Ask what the nested data represents before calling the design clever or risky; the surrounding column names reveal its intended boundary. section 7.6
