Executive Summary
Every MDM product can match records in a demonstration. The program succeeds or fails on who owns each domain and whether the systems that use the data accept the governed record.
Master data management (MDM) software keeps one trusted record for each customer, product, supplier, location or other shared business entity, and sends that record to the applications that use it. Informatica describes MDM as the process of creating and maintaining a single, trusted record for customers, products, suppliers and more across your business. SAP names data consolidation, data governance and data quality management as the three key pillars of effective MDM. To choose MDM software, start with the domain you must master first, decide how the hub will relate to your source systems, and test whether the people who own the data can run the stewardship work themselves.
Item master data management is the product domain of the same market. In this guide, item master data means the product record that ERP, supply chain and commerce systems share: identifiers, attributes, units and specifications. Syndigo lists unifying product attributes, product identifiers, metadata and specifications into a single source of truth. Stibo Systems describes PIM as a subset of MDM: MDM covers various data domains, whereas PIM focuses only on product data. If the item master is your first domain, test product-first platforms and multidomain suites on the same item records.
What Does Master Data Management Software Do, and Which Domain Comes First?
Informatica calls the categories into which master data is classified domains, and lists customer, product, supplier, reference data, location, asset and employee data among the common ones. Most programs start with one domain whose duplicate or conflicting records already cause visible errors: customer records split between CRM and billing, or item records that differ between the ERP and the online catalog.
SAP describes master data management as the discipline of creating one trusted master reference for all important business data, such as product data, customer data, asset data and finance data. SAP states that master data integration does not change the quality of master data, but always distributes master data in its current state. Keep that distinction in a vendor demonstration. Moving records between systems is integration. Matching, correcting and approving them is master data management.
Platform changes often put MDM on the agenda. SAP states that the new, more automated processes implemented after a move to SAP S/4HANA will rely more heavily on clean and correct master data. Plan the master data work before the ERP cutover, since migrated records carry their duplicates with them.
Which Implementation Style Fits: Registry, Consolidation, Coexistence or Centralized?
The implementation style decides where the golden record lives and whether the hub changes data in the source systems. Stibo Systems lists four common master data management implementation styles: registry, consolidation, coexistence and transaction or centralized. Stibo Systems states that the registry style provides a read-only view of data without modifying master data. Stibo Systems states that the coexistence style builds a golden record as the consolidation style does, but stores master data in the central MDM system and updates it in its source systems. Stibo Systems states that the centralized style also lets you create master data, making the MDM system the system of record.
A registry or consolidation hub changes no source system, so it needs less agreement to start. Coexistence and centralized styles change how records are created and updated, so each source system owner has to agree to the change. Many programs choose a style per domain: a consolidated customer hub for reporting, and a centralized item master where new products are created.
Building a hub in-house means writing matching, survivorship rules, stewardship screens, audit history and integrations, all of which the products in this guide include. Building makes sense only when one system owns the data and the need is deduplication inside that system.
| Style | Where the golden record lives | What changes in the source systems |
|---|---|---|
| Registry | An index of matched records; the data stays in the sources | Nothing; sources are read, not written |
| Consolidation | In the hub, built from source data | Nothing; sources keep their own versions |
| Coexistence | In the hub, which also sends updates back | Records are updated from the hub |
| Centralized | In the hub, which also creates new records | Sources receive records from the hub, the system of record |
How do you evaluate Master Data Management (MDM) Software?
Vendors describe the same core functions. Informatica advises prioritizing data integration across the existing tech stack, data governance including automated deduplication and cleansing, and cloud-native scalability when selecting an MDM tool. The products differ in how those functions work on your own records, run by your own stewards.
Matching is where the methods differ most. Reltio states that it matches and merges records with pretrained LLMs, extensible rules and dynamic survivorship, with no model training required. IBM describes IBM Master Data Management as a cloud-native platform that unifies and governs data across domains and uses machine learning for entity resolution and relationship discovery. Ataccama states that AI agents triage match proposals in bulk, with clear cases merged automatically and the rest sent to stewards for review. Ask each vendor to show why two records merged, and how a steward reverses a wrong merge.
Stewardship is the daily work after go-live. IBM lists built-in stewardship tools with audit trails and ML guidance for match decisions. SAP lists pre-built change request workflows, matching and consolidation, and continuous data quality monitoring for stewards. During the trial, ask a business steward, not an administrator, to process a change request and resolve a duplicate.
The data model decides how quickly a new attribute or domain arrives. EBX states that what you model is what you get, with fully configurable applications generated on the fly. EBX lists managing classifications and hierarchies, such as postal codes, cost centers and financial hierarchies, once and distributing them everywhere. Ask whether adding an attribute is configuration or a services project.
The governed record has to reach the systems that use it. Informatica describes headless MDM as decoupling master data capabilities from the UI and making them available as APIs and MCPs. Semarchy describes a golden record as certified data, and an intelligent data product as that record wrapped in the business context, semantics, lineage, ownership and generated interfaces needed to consume it. Reltio lists prebuilt connectors for 1,000+ systems. Count the connectors your own systems need, and ask which are in the price.
| Capability | What it does | What to test |
|---|---|---|
| Matching and merging | Finds records that describe the same entity | Match results on your own duplicates, with the reason for each merge |
| Survivorship | Picks the winning value for each attribute | Rules per attribute and per source, and who may change them |
| Stewardship workflow | Routes changes and doubtful matches to people | A business steward completing a change request unaided |
| Data model | Defines domains, attributes and relationships | Adding an attribute and a domain during the trial |
| Hierarchies and reference data | Manages code lists and parent-child structures | Versioning and effective dates for a hierarchy change |
| Data quality | Validates and standardizes values | Rules for your formats, run before and after matching |
| Integration and write-back | Moves records in and out of source systems | Write-back to one source, including a rejected update |
| Audit and lineage | Records who changed what and where values came from | The history of one record from source to golden record |
Vendors in Master Data Management (MDM) Software
Representative vendors include Ataccama, EBX – Cloud Software Group, Informatica MDM, Profisee, Reltio, SAP Master Data Governance, Semarchy and Stibo Systems. The list is alphabetical, is not a ranking, and is not for sale. How we choose vendors
The approaches below follow two questions: which domain the product was built around, and which data platform it expects to sit beside. Most vendors now sell several domains, and the approach describes the design each one starts from.
Several vendors attach MDM to a wider data platform. Informatica describes its 360 Applications as domain-specific solutions built on MDM that give a pre-configured view of customers, products, suppliers and more. IBM lists flexible deployment (SaaS, on-premises, hybrid), low-code interfaces and open APIs for IBM MDM. SAP describes SAP Master Data Governance as the governance layer of a business data fabric, unifying master data, policy and metadata. Semarchy states that its foundational MDM capabilities provide entity resolution, quality, enrichment and AI-in-loop stewardship to deliver certified golden records. Semarchy states that customers can choose SaaS or self-host and deploy on Snowflake, Microsoft, AWS or on-prem.
Others are built as cloud services or around one vendor's data stack. Reltio describes its agent-enabled, cloud-native SaaS MDM platform as including LLM-augmented entity resolution, real-time data integration, continuous data quality management, reference data management and graph-based relationship views. Reltio states that its cloud-native offering delivers SaaS economics with zero-downtime upgrades. Profisee describes itself as the trusted choice for Copilot, Fabric and Microsoft-centric customers.
Product-first, model-driven and quality-led platforms complete the field. Stibo Systems describes STEP as its platform that turns fragmented enterprise data into governed master records. Syndigo states that its multi-domain MDM unifies, governs and synchronizes data across systems. Syndigo lists consolidating, standardizing and managing product data across PIM, ERP, e-commerce and retail systems. EBX, formerly TIBCO EBX, is described by its publisher as enabling organizations to model, manage and govern master, reference and metadata across domains. ON EBX states that it operates as a dedicated business unit under Cloud Software Group. Ataccama states that Ataccama Master Data Management consolidates fragmented, duplicate data into golden records, and that users can work in the UI or headless. Ataccama describes achieving reliable matching with validated, enriched data using 300+ built-in quality functions.
| Vendor | Approach | Where it fits |
|---|---|---|
| Ataccama | Data quality platforms with MDM | Golden records and stewardship within a data quality and catalog platform |
| SAP Master Data Governance | Governance in the ERP stack | Change-request workflows for master data within SAP Business Data Cloud |
| Profisee | MDM aligned to one cloud data stack | Volume-based pricing and an upgrade path from Microsoft Master Data Services |
| Semarchy | MDM as data products | Golden records published as governed data products, SaaS or self-hosted |
| EBX – Cloud Software Group | Model-driven master and reference data | Applications generated from the data model, with hierarchies managed once |
| Informatica MDM | Multidomain suites in a data platform | Prebuilt 360 applications for customers, products and suppliers on one MDM |
| Stibo Systems | Product-first MDM | The STEP platform for product master data, extended to other domains |
| Reltio | SaaS MDM hubs | Real-time SaaS MDM with LLM-based matching and graph relationship views |
One vendor is named for each of the eight approaches; IBM Master Data Management and Syndigo appear in the sections above, and the wider field runs to about twenty. The approaches were written first and the vendors selected for each afterward, and a row here cannot be bought. MDM vendors in this table or outside it can present their own products in the Spotlight below.
How much should you budget for Master Data Management (MDM) Software?
MDM vendors meter different things: records or domains under management, consumption units, named users or platform tiers. The meter decides how the bill grows when you add a domain or a source system. Implementation partners and the stewards who run the hub add costs that no license covers.
Informatica states that Informatica Processing Units (IPUs) give customers access to all eligible cloud services, and that customers can also access all Informatica master data management on a per-domain records basis. Profisee states that its domain-agnostic, volume-based pricing means customers only pay for the data they need. Profisee offers a SaaS option as a fully managed service, or deployment via PaaS in any cloud or on-prem. Ataccama states that three dimensions determine the Ataccama ONE pricing tier: named users, data objects managed in the platform and active DQ configurations, and that read-only consumer users are unlimited at every tier. Ataccama states that exceeding the threshold will never disrupt operations or result in retroactive charges.
The pages cited here publish meters, not prices. Ask for a quote at your current record count and at the count you expect after the next two domains, so that growth is priced before you sign.
| Meter | You are charged for | Grows with | Where it goes wrong |
|---|---|---|---|
| Records under management | Mastered records, sometimes counted per domain | New domains and sources | Counting source records before matching rather than golden records |
| Consumption units | Units drawn by each cloud service used | Jobs, volumes and services switched on | Matching and integration runs drawing on the same pool |
| Named users and tiers | Users with full access, plus tier thresholds | Steward and developer headcount | Business stewards priced as full users |
| Domains and environments | Each domain and each non-production environment | Domains added after the first | Test environments priced like production |
| Services and staff | Implementation partner and internal stewards | Domains, sources and write-back targets | Stewardship staffing left out of the business case |
How long does implementation take for Master Data Management (MDM) Software?
Roll out one domain at a time. Prove that stewards can run the first domain and that consuming systems use its records before adding the next.
Choose the first domain. Name a business owner and stewards, list the systems that create and use its records, and agree the survivorship rules: which source wins for each attribute.
Load real records from the main sources. Tune match rules until stewards accept the results, and have a steward resolve change requests and wrong merges without vendor help.
Connect consuming systems through APIs, events or batch files. For a coexistence or centralized style, switch on write-back to one source system first and check the records it receives.
Track duplicates, open change requests and failed data quality rules each week. Add the next domain when the first runs without project staff.
What should you ask vendors about Master Data Management (MDM) Software?
Ask each vendor to answer in writing and to demonstrate on a sample of your own records.
-
Is product or item data the first domain?Yes Shortlist product-first platforms and a multidomain suite, and test both on the same item records.No Start from multidomain suites, SaaS hubs and data quality platforms.
-
Does your ERP create most of the records you must master?Yes Include the ERP vendor's governance option and test change requests inside your ERP process.No Weight connectors to the systems that create your records.
-
Must corrected records flow back into the source systems?Yes Plan a coexistence or centralized style, and test write-back during the trial.No A registry or consolidation hub may be enough to start.
Master Data Management (MDM) RFP package: 122 questions, 10 demo scenarios and a vendor scorecard
RFP packages are a separate paid product. Vendors do not pay to appear in buyer guides.