Frame
Phase Goal
Establish whether a microproduct is worth building and feasible to deliver before you enter execution.
What Must Be True Before You Build
- A repeated workflow pain exists for a specific user segment.
- You can access enough data to create clear utility, and every material external source has documented usage rights. See Data Licensing.
- The value proposition is concrete enough to test quickly.
- A baseline architecture is feasible within team and budget constraints, and every material data product has a justified data architecture: sources, freshness, storage, serving, and complexity. See Modern Data Stack.
Required Outputs
- Problem brief with target user, problem, data inputs, output utility, and success metric.
- Usage rights for each material external data input: license, terms reviewed, and whether commercial use, redistribution, attribution, and share-alike apply.
- Go/no-go criteria for moving into implementation.
architecture.mddata_architecturerecord for every material data product: sources, volume, freshness, ingestion, storage, transformation, orchestration, serving, reliability, quality, cost, and complexity justification.