Scenarios 13–18
13 — Prioritize improvements
Stem: 25 low-severity issues vs one rare defect that can cause regulatory misreporting.
Best answer: prioritize by business impact/criticality/risk, not defect count.
Roles: Data Owner, Governance, DQ team.
Best action: prioritize regulatory-risk defect and document rationale/cost-benefit.
Weaker: fix largest number of rows first.
Changed fact: rare defect has no meaningful impact while widespread defect blocks customers → priority can reverse.
Source: pp. 428–429, 444.
14 — Quick hit vs strategic fix
Stem: bad postal codes can be corrected now; root cause is application accepting invalid codes.
Best answer: quick correction + root-cause/preventive change.
Roles: process owner, system owner, Steward.
Best action: remediate existing records and implement upstream validation.
Weaker: cleanup only.
Changed fact: source cannot be changed and controlled midstream standardization is cheaper/safer → ongoing correction may be justified.
Source: pp. 444–445, 457–459.
15 — Rule drift
Stem: DQ rule rejects a value that became valid after business policy changed.
Best answer: Manage DQ Rules.
Supporting: rule lifecycle/change control.
Roles: business owner, Steward, DQ analyst.
Best action: review/approve/version rule and update implementation.
Weaker: suppress alerts without correcting stale expectation.
Changed fact: policy never changed and value truly invalid → keep rule, remediate data/process.
Source: pp. 438, 445–446.
16 — Daily threshold breach
Stem: validity score falls below threshold every morning, then recovers after manual fix.
Best answer: monitoring detects recurring issue; RCA/prevention is missing.
Roles: DQ operations, process/system owner.
Best action: locate morning introduction point, prevent recurrence, continue monitoring.
Weaker: celebrate successful daily cleansing as sufficient control.
Changed fact: source cannot change and controlled correction is accepted operating design → focus on SLA/performance of correction.
Source: pp. 446–459.
17 — Unassigned incident
Stem: severe data defect logged but unassigned for days.
Best answer: issue-management assignment failure.
Supporting: incident management, SLA.
Roles: DQ operations manager, expert owner, Governance.
Best action: assign by expertise, track aging, escalate by impact/urgency/SLA.
Weaker: create another profiling rule; detection is not the workflow gap.
Changed fact: ownership assigned but diagnosis stalls because lineage missing → lineage/technical investigation becomes obstacle.
Source: pp. 447–450.
18 — Supplier handoff
Stem: vendor sends critical Reference Data; contract lacks quality rules, thresholds, response, escalation.
Best answer: establish DQ SLA / supplier quality requirements.
Supporting: Reference Data, contracts.
Roles: Data Owner, vendor/procurement manager, Steward, DQ team.
Best action: define covered data, rules/measures, thresholds, notification, remediation deadline, escalation.
Weaker: rely only on downstream cleansing.
Changed fact: SLA exists but monitoring absent → implement inspection/measurement at boundary.
Source: pp. 449–450.