How a data team is organised affects what it delivers.
Common Roles
- Data engineers: build pipelines and platforms.
- Analytics engineers: model data for analysis, often with SQL and dbt.
- Data analysts: answer business questions and build reporting.
- Data scientists: statistical modelling, experiments and machine learning.
- ML engineers: productionise models.
- AI engineers: build applications on language models.
- Data product managers and governance leads.
Organisational Models
- Centralised: one data team serving the organisation. Consistent standards; can become a bottleneck.
- Embedded: data people sit within business teams. Close to needs; risk of inconsistency.
- Hub and spoke: a central platform and standards team with embedded specialists. A common compromise.
Success Factors
- Clear ownership of data products.
- Shared definitions and platforms.
- Career paths for specialists.
- Close collaboration with business teams.
Evolve
The right structure changes as organisations mature.