Master Data Management (MDM)

Master Data Management (MDM) is the discipline, the governance framework and the tooling for managing an organisation's shared reference data — the customer, supplier, material, employee and product records that drive operational transactions across multiple systems. MDM addresses the perennial problem of mid-market and enterprise IT: the same customer exists in CRM, ERP, e-commerce, warehouse system and ten spreadsheets, with subtle differences that compound into reporting errors, double mailing and operational chaos.

Master-data domains

Customer — name, address, billing-and-shipping addresses, tax IDs, payment terms, classifications, credit limit
Supplier — name, address, payment details, bank account, tax IDs, supplier-class ratings, contracts
Material / item — SKU, attributes, units, weights, dimensions, hazardous classifications, intrastat codes
Product — marketing-grade attributes (overlap with PIM)
Employee — personal data, organisational assignment, payroll classification, access roles
Account / cost-centre — chart of accounts, controlling structures, group consolidation hierarchies

Leading MDM platforms

Enterprise: SAP Master Data Governance, Informatica MDM, Stibo STEP, Profisee, IBM InfoSphere MDM, Reltio — 200,000 to 2,000,000 EUR per year. Mid-market: Semarchy xDM, Pimcore MDM, EnterWorks (Winshuttle), VisionR — 30,000 to 200,000 EUR per year. Embedded in ERP: SAP S/4HANA includes MDG (Master Data Governance) as add-on; Microsoft Dynamics 365 F&O includes Common Data Service / Dataverse for master-data sharing; Oracle Cloud has Product Hub and Customer Data Management Cloud. For mid-market in Germany, Switzerland and Austria, MDM is often implemented via the ERP's built-in capabilities plus a lightweight workflow tool, rather than a dedicated MDM platform.

MDM implementation patterns

Four architectural patterns. Consolidation MDM: MDM passively collects and matches records from source systems without changing them — cheap to implement, no operational impact, but limited cleanup value. Registry MDM: MDM creates an index of golden records that points to source records — good for reporting and analytics, but transactions still happen against unclean sources. Centralised MDM: MDM is the system of record, all changes happen there, source systems consume via subscription — high implementation effort, biggest long-term benefit. Coexistence MDM: bidirectional sync between MDM and key source systems — most common in practice for ERP-centric organisations.

Governance and roles

MDM is mostly a governance problem dressed as a technology problem. Successful MDM programmes establish clear ownership: Data Domain Owner (business head accountable for a domain — e.g. Head of Sales owns customer data), Data Stewards (operational role for daily quality control and exceptions), Data Custodians (IT role for technical maintenance). Without these roles, MDM tools quickly become an expensive duplicate of the same data quality problems. Mid-market in DACH typically combines the stewardship role with existing operational managers rather than creating dedicated FTEs.

Frequently Asked Questions

Do we need MDM if the ERP is our single source of truth?

If the ERP truly is the only system holding master data, you do not need separate MDM. In practice, almost no mid-market company achieves that — CRM, e-commerce, marketing automation, warehouse system and BI all hold slightly different versions. The question becomes whether the ERP-side master-data quality is good enough, or whether you need an MDM layer to enforce it.

MDM versus PIM — how do they relate?

PIM is a specialised MDM for product master data with extra emphasis on marketing-grade attributes, multilingual content and channel publishing. MDM is broader, covering customer, supplier and employee master data alongside product. Companies often run both: MDM for operational master data, PIM for marketing-and-commerce product data, with bidirectional sync between them.

What is a realistic MDM payback for a mid-market manufacturer?

For a 300-employee manufacturer with five-plus systems holding overlapping customer and supplier data, MDM payback typically comes through: 30-50% reduction in customer-data inconsistency errors, 5-15% reduction in supplier duplicates with corresponding payment-term optimisation, faster customer onboarding (days instead of weeks), and cleaner reporting. Payback period: 18-36 months on a 200,000-500,000 EUR implementation.