Master Data Management

The master data model is the formal, unified schema that defines the structure, attributes, relationships, and business rules for all critical data entities (e.g., Customer, Product, Supplier). It serves as the single blueprint for the golden record, ensuring semantic consistency across the enterprise. Key elements include:

• Entity Definitions: Core business objects and their mandatory/optional attributes.
• Hierarchies and Relationships: Defines how entities relate (e.g., organizational structures, product catalogs).
• Business Rules and Validation Logic: Encodes domain-specific constraints and data quality rules.

Identity resolution is the core process of disambiguating and linking records from disparate source systems that refer to the same real-world entity. It employs deterministic and probabilistic matching algorithms to create a unified identifier. This process is critical for building the golden record.

• Deterministic Matching: Uses exact or rule-based comparisons (e.g., matching on Tax ID).
• Probabilistic Matching: Uses statistical models and fuzzy logic to match on non-exact data (e.g., name and address variations).
• Survivorship Rules: Determines which source system's attribute values 'survive' into the golden record when conflicts exist.

Data governance within MDM establishes the policies, standards, and organizational roles for managing master data throughout its lifecycle. Data stewards are business or technical experts assigned accountability for the quality and definition of specific data domains.

• Policy Management: Defines data ownership, access controls, and compliance requirements.
• Stewardship Workflows: Operationalizes governance through review/approval processes for data changes.
• Issue Management: Tracks and resolves data quality exceptions and definitional disputes.

The integration hub is the middleware layer responsible for bidirectional data flow between the MDM system and all connected source and consuming applications (spoke systems). It ensures the golden record is propagated and kept consistent across the enterprise.

• Publish/Subscribe Model: Broadcasts mastered data updates to subscribing systems.
• Change Data Capture (CDC): Listens for updates in source systems to trigger mastering processes.
• Conflict Handling: Manages synchronization conflicts when the same record is updated in multiple systems simultaneously.

Data quality management components are embedded within the MDM platform to profile, cleanse, standardize, and monitor master data. This ensures the integrity of the golden record before and after its creation.

• Data Profiling: Analyzes source data to understand content, structure, and quality issues.
• Cleansing & Standardization: Applies rules (e.g., address normalization, phone number formatting) to raw data.
• Quality Dashboards & Monitoring: Provides real-time metrics on completeness, validity, accuracy, and consistency of master data.

MDM can be deployed using different architectural implementation styles, each with distinct trade-offs between latency, system impact, and complexity. The choice depends on business requirements and existing IT landscape.

• Registry Style: Provides a lightweight, read-only index of matched records; source systems retain original data.
• Consolidation Style: Creates a physical golden record in a central repository, primarily for reporting and analytics.
• Coexistence Style: Maintains a central golden record while allowing source systems to update their local copies, which are synchronized.
• Transactional (Centralized) Style: The central MDM hub is the sole system of entry and authority for master data; all other systems are consumers.

The Customer domain manages all information about individuals or organizations that purchase goods or services. This includes:

• Party Data: Names, contact details, and communication preferences.
• Relationship Data: Hierarchies (e.g., parent-subsidiary) and household linkages.
• Behavioral Data: Aggregated transaction history and engagement scores.

A Golden Record for a customer eliminates duplicates from CRM, billing, and support systems, enabling a 360-degree view for personalized marketing, unified service, and accurate lifetime value calculation.

The Product domain defines all goods and services an enterprise sells or manages. Key components include:

• Descriptive Attributes: SKUs, specifications, dimensions, and packaging.
• Hierarchical Data: Categories, families, and bill-of-materials structures.
• Digital Assets: Images, manuals, and compliance certificates.

Consistent product data is critical for e-commerce catalogs, supply chain planning, and regulatory compliance (e.g., SDS sheets). MDM synchronizes data across ERP, PIM, and e-commerce platforms.

This domain manages entities that provide goods, services, or infrastructure to the organization. It encompasses:

• Legal Entity Data: Registered names, tax IDs, and compliance certifications.
• Financial Data: Banking details, payment terms, and credit ratings.
• Performance Data: Quality scores, on-time delivery rates, and risk assessments.

Centralized supplier data prevents fraud, optimizes procurement, ensures contract compliance, and mitigates supply chain risk by providing a single view of vendor relationships and performance.

The Location domain provides a standardized model for all physical and logical places relevant to the business. This includes:

• Geographic Sites: Stores, warehouses, offices, and manufacturing plants with precise addresses and geo-coordinates.
• Internal Locations: Specific aisles, bins, or rooms within a facility.
• Sales Territories: Defined regions, districts, and routes for go-to-market operations.

Accurate location data is foundational for logistics optimization, asset tracking, territory management, and regulatory reporting (e.g., for tax jurisdictions).

This domain covers physical and non-physical items of value owned or leased by the enterprise. It includes:

• Fixed Assets: Machinery, vehicles, IT hardware, and real estate.
• Digital Assets: Software licenses, patents, trademarks, and digital content.
• Financial Assets: Instruments like stocks or bonds.

MDM for assets ensures accurate financial depreciation, supports maintenance scheduling, manages warranty and lease terms, and provides a complete register for insurance and audit purposes.

This domain manages data about people employed by the organization and its internal structure. Key elements are:

• Employee Master: Role, department, manager, cost center, and skills.
• Organizational Hierarchy: Reporting structures, matrix relationships, and legal entity roll-ups.
• Reference Data: Job codes, pay grades, and cost center mappings.

A mastered organizational chart integrated with HR, finance, and project management systems is essential for headcount planning, financial consolidation, access management, and enterprise reporting.

An architectural framework that uses a knowledge graph as a unifying semantic layer to provide integrated, contextualized, and governed access to enterprise data across disparate sources. Unlike traditional MDM, which focuses on core entities, a semantic fabric provides meaning and relationships for all data, enabling complex queries and AI-ready data structures.

• Core Function: Creates a business-understandable, conceptual model over raw data.
• Key Difference from MDM: MDM manages the 'golden record' of key entities; a Semantic Fabric provides the 'meaning' and connections between all entities and data points.

A decentralized sociotechnical architecture that organizes data ownership by business domain. It treats data as a product, with domain teams responsible for their own data products. MDM in a Data Mesh context shifts from a centralized custodial model to a federated governance model, where domains are responsible for the quality and definition of their master entities, which are then shared as products.

• Core Principle: Domain-oriented decentralization and data-as-a-product.
• Relation to MDM: Defines how master data is owned, produced, and consumed in a scalable, modern organization.

The single, authoritative, and consolidated version of truth for a core business entity (e.g., customer, product, supplier). It is the primary output of an MDM process, created by merging, cleansing, and deduplicating data from multiple source systems. The golden record serves as the definitive source for that entity across all enterprise systems.

• Creation Process: Involves entity resolution, data quality rules, and survivorship rules.
• Purpose: Eliminates conflicting versions of the same entity, ensuring consistency in reporting and operations.

The core computational process within MDM for identifying, linking, and merging records that refer to the same real-world entity across different data sources. It uses algorithms to analyze attributes, detect matches, and resolve conflicts.

• Techniques: Includes deterministic rule-based matching, probabilistic/fuzzy matching, and machine learning-based models.
• Challenge: Dealing with incomplete, inconsistent, or erroneous data (e.g., 'Jon Doe Corp' vs. 'John Doe Corporation').

An abstraction that sits between physical data stores and consuming applications (like BI tools). It provides a business-friendly conceptual model—often defined by ontologies, taxonomies, and business metrics—so users query data in their own terms. An MDM system often feeds cleansed golden records into a semantic layer.

• Key Benefit: Decouples business logic from underlying database schemas.
• Example: Mapping technical column names like CUST_ID and PRD_CD to business concepts like 'Customer' and 'Product' with defined hierarchies.

A data integration technique that provides a unified, abstracted view of data from multiple disparate sources in real-time, without requiring physical data movement or replication. It can be used to create a logical, integrated view of master data that remains in its source systems, complementing a physical MDM hub.

• Core Advantage: Provides real-time access to the most current source data.
• Use with MDM: Can federate queries to both the MDM hub (for golden records) and operational systems (for transactional context).