The Silent Breaking Change Problem
In most data organisations, there is no formal agreement between data producers and consumers. An upstream team renames a column or changes a data type, and downstream pipelines break silently — often only noticed when a business dashboard shows wrong numbers.
What a Data Contract Defines
- Schema: columns, types, nullability
- Semantics: business meaning of each field
- SLA: freshness, availability, and volume expectations
- Ownership: who is responsible for maintaining the contract
Enforcement
Tools like Soda, Great Expectations, and custom CI/CD hooks can automatically validate that data contract terms are met before data is published to consumers.