MDM styles, matching, enrichment and canonical models
The MDM toolkit an architect compares: hub versus linking styles, a canonical model that maps each source once, deterministic and probabilistic matching with weights and thresholds, steward review for uncertain matches, enrichment from external reference data, and alternate hierarchies over one set of golden records.
Compare and contrast the various techniques, approaches and considerations for implementing master data management (MDM) Solutions (for example, MDM implementation styles; harmonizing and consolidating data from multiple sources; establishing data survivorship rules, thresholds and weights; leveraging external reference data for enrichment; canonical modeling techniques, hierarchy management).
~6 min read · 6 practice questions ·
What the exam expects
- Compare MDM implementation styles (a central hub that writes a golden record versus linking that leaves sources intact) and name the source of truth per entity.
- Harmonize many sources through a canonical model instead of pairwise mappings.
- Explain deterministic and probabilistic matching, match weights and thresholds, and when uncertain matches go to a data steward.
- Recommend enrichment with external reference data and alternate hierarchies over one master.
Key terms: golden record (the consolidated "best" record per entity); canonical model (one common schema every source maps to once); match threshold (the score a pair must reach to match); enrichment (adding values internal systems lack); alternate hierarchy (another roll-up of the same master records).
Key facts
- Detail
- Golden record overwrites sources; unified profile links and keeps lineage
- Detail
- Name 50%, number 20%, suffix 15%, suite 15%; name and number alone (70) aren't a match
- Detail
- AI model with a 0.7 confidence threshold; not on Account fields
- Detail
- Less precise methods raise consolidation and undesirable matches
- Detail
- Flagged for steward review, not merged automatically
- Detail
- Postal authority data; industry classification codes
- Detail
- All trace back to the same golden records
- Why does a canonical model scale better than pairwise mappings?
- When should a probable duplicate go to a steward instead of merging automatically?
- Why can two streets sharing a number and name still fail standard matching?
- What fills a gap such as industry that no internal system holds?
The questions are free to read. Answers with their sources and self-grading are included with the pack.
Keep reading with the Platform Data Architect pack
- Core concepts
- Exam traps
- Worked scenario
The pack includes:
- All 21 lessons of the Platform Data Architect guide, with reading progress
- Interactive decision walks and step-by-step sequence steppers
- Inline practice questions and one-click topic practice (6 questions for this lesson)
- An official source cited for every claim
- An AI tutor that answers from the guide
- Unlimited timed practice exams