Skip to main content
Version: v3.3 print this page

Introduction

Master Data Management (MDM) is one of the Amorphic Applications and a comprehensive data management system designed to streamline, govern, and master organizational data. It relies on Amorphic as the central data backbone, ingesting raw data from Amorphic datasets and writing mastered unified master records back to Amorphic. It acts as a central hub that gathers, cleans, and unifies your data. Its primary goal is to create a "Single Source of Truth"—ensuring that you have one accurate, complete, and trusted profile for every customer, product, or partner, regardless of where that data originated.

Why we need MDM?​

In today’s business environment, data often lives in different places—sales spreadsheets, marketing platforms, customer support systems, and shipping logs. This often leads to fragmented information, where a single customer might appear as three different people across your systems. Duplicate and inconsistent records make it hard to run accurate analytics, personalize experiences, or trust reports.

MDM solves this by acting as a central layer that collects data from your various sources, standardizes it, identifies which records refer to the same real-world entity (person, product, or partner), and merges them into a single, trusted unified master record. That unified master record becomes the one profile your organization uses for that entity everywhere. As a result, you get one accurate, complete profile per customer, product, or partner; analytics and reports can be trusted; and campaigns and operations can rely on a single source of truth instead of conflicting data.

How it works?​

  1. Ingestion: It takes in records from all your different business sources.
  2. Standardization: It fixes common formatting issues (e.g., ensuring "Street" and "St." are treated the same way).
  3. Unification: It identifies when different records actually belong to the same person or entity and merges them into one unified master record.

High-Level User Flow (The Journey)​

The MDM user journey is tightly integrated with the Amorphic ecosystem. From defining your data model to publishing unified master records, you move through three main phases.

Phase 1: Configuration & Modeling​

Goal: Define the data structure and how your source data maps to it.

  • Select domain: Define a custom schema mapping that matches your business object.
  • Select dataset: The system shows only the Amorphic datasets you have access to, so you pick the right source for your schema mapping.
  • Map schema: Map columns from your chosen dataset to the standardized input fields (e.g., which column is "name", "address", or "identifier"). This tells MDM how to read and interpret your data.
  • Prepare for enrichment: The system is designed to support normalization, standardization, and cleansing rules. In the initial phase these may act as pass-throughs, but the structure is in place so you can add rules later without changing the overall flow.

Phase 2: Orchestration​

Goal: Run a workflow that reads from Amorphic, matches and resolves records, and produces preliminary unified master records.

  • Create workflow: Start a new workflow from the MDM application and give it a clear purpose (e.g., "Customer deduplication – Q1").
  • Select data: Choose the input dataset from Amorphic; the list is filtered by your permissions so you only see datasets you can use.
  • Configure matching: Decide how the system should find duplicates—rule-based matching for exact or near-exact matches, or fuzzy matching for probabilistic, ML-based matching when spellings or formats differ.
  • Run the workflow: Execute the workflow. The system processes the data, applies your matching logic, and outputs a set of preliminary unified master records for review.

Phase 3: Governance & Stewardship​

Goal: Review, correct, and finalize the unified master records before publishing.

  • Open Unified Record Hub: Data stewards open the processed data in the Unified Record Hub to see match results, clusters, and record details.
  • Steward the data: Perform merges (combine duplicate records into one), unmerges (split incorrectly merged records), or overrides (correct or enrich specific fields) so the unified master records reflect the right, trusted view. Merge/Unmerge are only available for workflows. These are done on the record details page and update the source record that contributed the winning value to the unified record.
  • Finalize and publish: Once stewardship is complete, the final unified master records are written back to a designated Amorphic dataset. That target dataset can be created as part of the workflow, so your mastered data is available in Amorphic for reporting, analytics, and downstream applications.