des-requirements-and-kpis

To turn business questions into measurable data engineering requirements, you transform approved questions into functional and non-functional requirements, candidate KPIs, and metric definitions with explicit acceptance criteria, traceability, and SLA freshness.

Defining SLA freshness and metric ownership requires specifying explicit acceptance criteria and ownership decisions within your KPI catalog, stopping the process to request user decisions when formulas, grain, or SLAs cannot be safely inferred.

Create a requirements and KPI catalog by mapping approved business questions to candidate KPIs and non-functional requirements, ensuring traceability, tracking evidence status, and explicitly listing unresolved items to prevent downstream physical design.

You should halt the requirements gathering process when formulas, grain, owners, SLAs, freshness, acceptance criteria, or conflicts cannot be safely inferred, applying HALT-safe decision handling to request explicit user decisions.

Validate KPI readiness for the next data engineering phase by applying evidence mapping and a Done Gate to decide whether the phase is Pass, Pass with risks, Fail, or Blocked based on explicit, measurable requirement definitions.

Yes, you can prevent premature physical design by producing requirement and KPI catalogs that focus on functional metrics, evidence status, and traceability, ensuring downstream phases design data products from evidence rather than assumptions.