Matching
Matching helps you identify and consolidate duplicate records that represent the same real-world entity across your data sources, ensuring a single, accurate view of your master data.
|
Throughout this guide, matching refers to the complete process of identifying and consolidating duplicate records unless otherwise specified. When we need to distinguish the technical steps, we will explicitly refer to matching (identifying and linking records) and merging (consolidating them into golden records). For details, see How matching works: Overview. |
What is matching?
Matching is a crucial process in master data management used to identify and handle duplicate records representing the same business object (for example, a customer) across different data sets coming from different primary systems, which is typical for consolidation and coexistence MDM implementation styles. It involves comparing new or existing data against stored records using rules to detect similarities and discrepancies.
For example, imagine your organization has these records from different systems:
| Source | System A (CRM) | System B (E-commerce) | System C (Support) |
|---|---|---|---|
Name |
Cameron Smith |
C. Smith |
Cam Smith |
cameron.smith@email.com |
cameron.smith@email.com |
cameron.smith@email.com |
|
Phone |
(555) 123-4567 |
+1-555-123-4567 |
555.123.4567 |
While these appear to be different customers at first glance, matching rules can identify that they likely represent the same person based on similar email addresses, phone numbers (despite different formatting), and name variations. The matching process would group these records together and create a single, comprehensive golden record for Cameron Smith.
Why matching matters
As effective matching ensures you maintain a single, accurate view of each entity in your system, it helps you address a number of challenges your organization might be facing:
-
Duplicate records leading to inconsistent communication and poor customer experience.
-
Inaccurate reporting and analytics due to fragmented data.
-
Compliance issues when data is scattered across multiple records.
-
Wasted resources from redundant data processing and storage.
How matching works: Overview
The matching process follows these fundamental steps:
-
Records enter the system - New or updated records from your various data sources.
-
Records are cleansed and standardized - Data is prepared by removing inconsistencies, normalizing formats, and addressing quality issues to optimize comparison.
-
MDM compares records using rules - Records are compared against existing data using your configured matching rules.
-
Records are grouped with master ID - When matches are found, records are grouped together and assigned a unique identifier.
Once matching identifies and groups related records, the next step is merging. This is where these grouped records are consolidated into golden records.
The entire matching process is executed within a matching plan through the Matching step component. To learn more about the inner workings of the Matching step, see How the Matching Step Works.
Key concepts in matching
Before looking into how matching works in ONE MDM, it’s helpful to understand these key terms that you’ll encounter throughout the matching process.
- Master ID
-
A unique identifier assigned to each group of matched records. Master IDs are designed to remain stable as data evolves, though significant data changes might occasionally require ID updates.
- Golden or master record
-
The single, consolidated view of an entity created by combining information from all matched records in a group. This represents the best or most complete version of the data.
- Silver record
-
A contextual master record representing the consolidated view of an entity within a specific scope. These records are created through deduplication of related records rather than matching, for example, deduplicating addresses within each party and address type.
- Instance or source record
-
The original, unconsolidated records from source systems that serve as input to the matching process. They might contain duplicate or partial information and must first go through cleansing and standardization before matching.
Matching approaches
ONE MDM provides three approaches to handle matching and merging, depending on your data quality and business requirements. These work together and can be used in combination:
-
Automated approach: The application automatically links records through matching rules and merges them based on your consolidation rules, which is ideal for clean, consistent data. This is the process handled by the Matching step.
-
Manual approach: You uncover potential links through investigation or reports, then personally review and resolve each case through manual matching and merging decisions.
-
Proposal-based approach: A hybrid where the application suggests potential matches that require your approval before any merging occurs.
How ONE MDM handles matching
ONE MDM is designed to maintain stable, consistent identities:
-
Identities stay stable by default: Once records are matched and assigned a master ID, that identity typically remains unchanged even as the underlying data evolves.
-
Most merges require approval: Two different master IDs generally require manual approval to be merged, preventing accidental consolidation. However, automatic regrouping can occur when matching rules cause previously separate groups to merge.
-
IDs are preserved: When identities are merged, the resulting master ID is always one of the IDs that were previously assigned to the participating groups rather than creating entirely new identifiers.
These core principles ensure data consistency and traceability in ONE MDM. To explore how these principles work in practice and their underlying rationale, see Matching Architecture.
Next steps
Explore matching further:
-
Working with Matching - Practical aspects of using matching day-to-day, including the three matching approaches, managing identities over time, and different configuration layers and options.
-
Matching Architecture - Implementation details and design principles behind ONE MDM matching approach.
-
Matching Configuration - Comprehensive setup guide for implementing matching, from initial configuration to best practices.
Was this page useful?