Scenario Lab 13–18
13 — Dirty source values
Stem: load discovers malformed customer-state values.
- Primary: DQ defect/root cause.
- Supporting: Integration, Metadata, Governance, source stewardship.
- Roles: Steward, DQ specialist, source owner, integration team.
- Best: remediate/correct source where possible; define controlled warehouse exception handling and monitor recurrence.
- Weaker: hide every defect via warehouse transformation; preserves upstream root cause.
- Changed: valid values converted per business rule, e.g. currency/units → transformation.
- Source: pp. 375–379.
14 — Unknown dimension arrives late
Stem: fact arrives before dimension; business accepts controlled provisional handling.
- Primary: load-order exception with allowed provisional member.
- Supporting: Dimensional modeling, population, Metadata/audit.
- Roles: DW designer, integration engineer, Steward as needed.
- Best: optimistic load—create provisional/unknown dimension member, load fact, reconcile correctly later.
- Weaker: reject despite explicitly allowed temporary handling.
- Changed: policy forbids unresolved facts → pessimistic recycle.
- Source: pp. 377–379.
15 — Fact cannot resolve to dimension
Stem: policy forbids inventing a member.
- Primary: referential/load exception under strict rule.
- Best: pessimistic recycle/reject with alert, correction and later reload.
- Weaker: random/placeholder member contrary to policy.
- Changed: controlled provisional members allowed → optimistic may fit.
- Source: pp. 377–379.
16 — One BI tool for everyone
Stem: executives, power users and front-line staff are forced into one interface.
- Primary: BI portfolio/user segmentation.
- Supporting: Requirements, support, security, cost/governance.
- Roles: BI product owner, architects, CoE/support, user representatives.
- Best: segment users and match capabilities/tools to needs while rationalizing portfolio.
- Weaker: one interface solely for admin simplicity.
- Changed: users truly share similar workflows/capabilities → common tool can be appropriate.
- Source: pp. 379–384.
17 — Users build dashboards themselves
Stem: business users explore governed warehouse data and share dashboards.
- Primary: self-service BI governance.
- Supporting: Metadata, Security, DQ, Governance, support.
- Roles: analysts, BI CoE, Stewards, security/governance.
- Best: governed self-service on trusted reusable data, shared definitions, lineage, permissions, training/support, publication controls.
- Weaker: assume every user-created dashboard is successful self-service.
- Changed: no governed data/Metadata/security → unmanaged analysis; establish guardrails.
- Source: pp. 382–386, 389–392.
18 — Where did this KPI come from?
Stem: auditor wants origin and transformation path.
- Primary: provenance/traceability.
- Supporting: Metadata, Governance, Integration, audit.
- Roles: Metadata owner, Steward, integration/DW team, auditor.
- Best: end-to-end lineage supported by source-target and transformation Metadata.
- Weaker: provide glossary definition only; explains meaning, not provenance.
- Changed: question is only “what does this field mean?” → dictionary/glossary first.
- Source: pp. 363–364, 385–391.