Scenarios 19–24
19 — One dashboard for executives and analysts
Stem: executives receive raw exception rows and no scorecard/trends.
Best answer: Data Quality Response / reporting design.
Roles: DQ team, consumers, Governance.
Best action: role-appropriate scorecards/trends, retaining detail for analysts.
Weaker: raw defects only because “more detail is transparent”; not decision-useful.
Changed fact: executive needs SLA breach decision → include service-level conformance/escalation status.
Source: p. 450.
20 — Name cleanup
Stem: free-form PERSON_NAME must be split before enterprise formatting.
Best answer: Parsing first, then standardization.
Roles: DQ/integration specialist.
Best action: pattern-based component recognition, then transform to target representation.
Weaker: call first step enrichment.
Changed fact: add latitude/longitude after address standardization → enrichment.
Source: pp. 451–453.
21 — Confidence-based correction
Stem: identity-matching software proposes merges with confidence; uncertain cases require Steward approval.
Best answer: manually-directed correction.
Supporting: MDM.
Roles: Steward, DQ/MDM operations.
Best action: confidence thresholds, permitted automation, human routing, audit.
Weaker: auto-commit every merge.
Changed fact: rules deterministic and risk low → fully automated may fit.
Source: pp. 458–459.
22 — Organization knows pain but not data
Stem: leaders complain about bad reports; no profiling or quantified actual state.
Best answer: readiness — build objective actual-state knowledge first.
Roles: DQ lead, Stewards, analysts.
Best action: profile/measure critical data, link findings to pain/risk, then roadmap.
Weaker: launch full enterprise program based only on complaints.
Changed fact: actual state, risk, and support already well understood → broader strategic program may be ready.
Source: pp. 460–461.
23 — Hero spreadsheet
Stem: one employee manually fixes regulatory report monthly; process opaque and single-person dependent.
Best answer: cultural / hero-culture risk.
Supporting: process documentation, Governance.
Roles: process owner, Governance, DQ team.
Best action: document requirements/process, control/automate appropriately, spread knowledge, measure objectively.
Weaker: reward hero and preserve dependency.
Changed fact: approved controlled exception with documentation and backup ownership → risk lower.
Source: pp. 461–462.
24 — Quality Metadata missing
Stem: DQ process validates addresses, but pass/fail results are discarded and consumers recheck repeatedly.
Best answer: Metadata Management relationship.
Why: DQ rules/results/context are Metadata consumers can use to judge fitness.
Roles: Metadata team, DQ team, Stewards.
Best action: persist quality rule/results/context and expose to consumers.
Weaker: add repeated downstream revalidation everywhere.
Changed fact: stored DQ result becomes stale after source change → currency/versioning of DQ Metadata must be managed.
Source: pp. 462–463.