01 / Align & prepare
Agree on what the platform needs to do.
Define the initial data domain, the questions it should support, and the source systems involved. Establish the environment, connections, modeling conventions, and review responsibilities before development begins.
- Decisions we make together
- Business priorities, source access, target platform, naming standards, and ownership of shared models.
- What we put in place
- A scoped first increment, a shared vocabulary, and the setup and development guidance the team will use.
02 / Profile & model
Understand the data before generating the model.
Profile source tables and identify business keys, relationships, and descriptive attributes. Map those concepts to existing enterprise entities, and agree how new ones fit. Review reference data, changing relationships, and history requirements.
- Decisions we make together
- What identifies each business concept, how keys remain distinct across systems, and which history and sensitive attributes need special handling.
- What we put in place
- A reviewed source model and logical Data Vault design, including hubs, links, satellites, and source mappings.
03 / Generate & extend
Turn standards into consistent implementation.
Use metadata, conversion rules, and templates to carry the logical design into physical models, loading structures, and deployment output. Extend the automation where your source behavior or architecture requires more than the standard pattern.
- Decisions we make together
- Incremental watermarks, lookback and retention settings, historical backfills, and the exceptions that need custom logic.
- What we put in place
- Generated assets, documented configuration, and reviewed extensions that can be reused as more sources are added.
See extension examples ↗
04 / Validate & release
Coordinate changes across models and jobs.
Develop changes in controlled increments and coordinate work on shared entities. Select the changed objects for deployment, review import logs, and test generated loading behavior before merging the approved design into the shared baseline.
- Decisions we make together
- Release scope, shared-entity coordination, job dependencies, refresh timing, and validation responsibilities.
- What we put in place
- Reviewed deployment groups, tested jobs, and an updated model baseline. Loading is sequenced so parameters, source loads, staging, and dependent work run in the right order.
05 / Deliver information
Put useful data in front of the business.
Create an initial analytical view so stakeholders can validate the shape and meaning of the data. Identify business rules that should be shared or historically tracked, then incorporate their results into the data products.
- Decisions we make together
- Facts, dimensions, business calculations, rule ownership, traceability needs, and the definitions that reports should use.
- What we put in place
- Information marts, reusable business-rule outputs where needed, and a foundation for governed semantic models, reporting, and AI.
06 / Operate & enable
Leave a platform your team can extend.
Validate refresh and reload behavior, measure analytical performance, and tune the structures that need it. Share the modeling, deployment, and support patterns so your team can onboard the next source and maintain the implementation.
- Decisions we make together
- Performance expectations, operating responsibilities, retention and reload policies, and how template changes are reviewed during upgrades.
- What we put in place
- Documented workflows, configuration guidance, knowledge transfer, and a repeatable path for the next delivery increment.