Master Data / Golden Record & Entity Resolution
free10 data-quality rules for Master Data Management (MDM). Covers duplicate detection, golden record uniqueness, cross-source linkage, orphan master detection, survivorship rules, match confidence, conflicting attributes, key standardization, merge/split audit trails, and golden record freshness.
Checks included (10)
Duplicate Candidate Detection
Detect probable duplicate entities using normalized name + email/address/phone similarity above a threshold. Surfaces unmerged duplicates.
One Golden Record per Entity
Each resolved entity must have exactly one golden/master record — duplicate master_ids defeat the purpose of MDM.
Cross-Source Key Linkage Complete
Every source-system record must be linked to a master_id in the crosswalk/xref table — no unlinked source rows.
No Master Without Source
A golden record must be backed by at least one surviving source record — orphan masters indicate a broken merge or delete.
Survivorship Rule Applied
Within each entity cluster exactly one source record is marked source-of-truth / is_golden = true for each survived attribute.
Conflicting Attribute Flag
Where two source records for the same entity disagree on a critical attribute (DOB, tax id, legal name), the conflict must be flagged for stewardship.
Match Confidence Above Threshold
Auto-merged records must carry a match-confidence score at or above the auto-merge threshold; lower scores require manual review.
Key Attributes Standardized Before Match
Match keys (email, phone, name, postal code) must be normalized (lowercased, trimmed, canonical format) before matching to avoid false non-matches.
Merge / Split Audit Recorded
Every merge or split must be logged with timestamp, operator/rule, and the affected record ids for reversibility and audit.
Golden Record Refreshed Within SLA
Golden records should be re-resolved within the refresh SLA after upstream source changes; stale masters drift from reality.