Skip to main content
Guides & How-To

MDM: master data management, not mobile device management

MDM makes sure customers, suppliers and products are consistent and correct across your systems. Here is when it is worth tackling, and how little machinery is often needed at the start.

Author: Agent360 editorial team
Published: 28 October 2026
Last updated: 9 October 2026
Reading time: 9 min read

In short: MDM (Master Data Management) is the discipline and the tools that make sure your most important shared data (customers, suppliers, products and locations) is consistent, correct and has an owner across all systems. Here MDM means master data, not mobile device management, which uses the same abbreviation. The goal is one true version of each customer and each product. Many smaller businesses get a long way by naming one system as the source for each type of master data and introducing fixed rules, before buying a separate MDM tool.

If you've ended up here, it may be because the same customer exists three times in the CRM and twice in the ERP, because two departments bring different figures for the same meeting, or because invoices keep going to an old address. Here you get a down-to-earth explanation of what MDM is, when it's worth tackling, and how little machinery is often needed at the start.

What is MDM, and what does it do?

MDM is a discipline in which business and IT work together to make sure the company's official, shared master data is consistent, correct and has clearly assigned responsibility. Master data is the products, accounts and parties that your business transactions are about. An order is a transaction. The customer and the product on the order are master data.

The central concept is the "golden record": the one correct version of a customer or product that the other systems follow. The system where the correct version lives is called the system of record. MDM is supported by technology, but it's just as much about people and processes. There are typically two roles: a data owner, who sets the requirements for definition, quality and security, and a data steward, who looks after the data day to day.

MDM typically covers:

  • Duplicates: finding and removing multiple versions of the same customer or product.
  • Standardisation: consistent formats for names, addresses and units, including in bulk corrections.
  • Rules at the point of entry: incorrect data is stopped before it gets into the systems.
  • Consolidation: data from several sources is merged into one version.
  • Distribution: the right data is sent out to the systems that need it.
  • Hierarchies and classification: for example which companies belong to the same group, and which product group a product belongs to.

There are several ways of doing it. The simplest is to name one source system per type of data. Other models gather data in a central hub and copy it out (consolidation), show one combined virtual version across the sources (federation) or copy data from system to system (propagation).

When do you need MDM?

Signs that the time has come:

  • The same customer or product exists several times with different spellings.
  • Reports from two systems give different figures for the same thing.
  • Goods or invoices go to outdated addresses.
  • You've bought or merged with another business and now have two customer registers.
  • You want to automate or use AI, and that requires the systems to agree on who the customer is.

Acquisitions are a classic problem. Every merger can bring yet another customer database, and over time they become hard to reconcile. An MDM project may be premature if you have one main system that everyone works in, and where master data is already in order. Then it's mostly about keeping it that way.

Danish businesses have an advantage: CVR, the Danish business register. The Danish Business Authority (Erhvervsstyrelsen) describes the Central Business Register (Det Centrale Virksomhedsregister) as the authoritative register for basic data on all registered businesses since 1999, and data from CVR is publicly available. The CVR number is therefore an obvious shared key for customers and suppliers, and name and address can be checked against the register. But be aware of advertising protection. Under section 19 of the CVR Act, private businesses must not use basic data on advertising-protected businesses for direct marketing, and if you pass on contact details, the protection must be clearly marked. The marking should therefore be a field in your master data.

Types of master data and their shared keys

Master data Example fields Typical source system in an SME Shared external key
Customers (businesses) Name, address, contact persons, payment terms, advertising protection CRM or ERP CVR number
Suppliers Name, address, payment details, terms ERP CVR number
Products Item number, description, unit, weight ERP, enriched in PIM GTIN (the number under the barcode)
Locations Warehouses, shops, delivery addresses ERP GLN (location number)

If you're unsure what belongs in the CRM and what in the ERP, read CRM or ERP: what do I need?

What does it cost?

In MDM the tool is rarely the biggest expense. Ask about these cost drivers:

  • Clean-up: finding and merging duplicates and correcting errors in what already exists.
  • Ownership and rules: meetings and decisions about who owns which data, and which fields are mandatory.
  • Integrations: every system that needs to read or write master data requires a connection.
  • Licence, if you choose a separate MDM tool: is it charged per user, per data domain or per number of records?
  • Ongoing care: time for the data steward, so quality doesn't slip again.
  • External data sources: lookups in, for example, CVR or other registers to enrich and check data.

Buy an off-the-shelf system or build your own?

For most smaller businesses, the first step isn't a new system. It's deciding which system is the source for customers, suppliers and products, and enforcing rules there: a mandatory CVR number, fixed formats and a check for duplicates on creation. An integration layer can then pass changes on to the other systems, so they don't lead lives of their own.

A separate MDM tool makes sense when you have many systems, several companies or countries, or when acquisitions have given you several registers that need to be merged. For product data that needs to be enriched with texts and images for sales, a PIM system is often the right supplement. Read more in PIM: Product Information Management.

Where does an AI layer fit in?

Master data requires many small judgements, and this is where an AI layer can help:

  • Finding duplicates: spotting that "Jensen VVS ApS" and "Jensen V.V.S." are probably the same customer, and suggesting a merge that a person approves.
  • Standardising: correcting formats of names, addresses and units.
  • Checking: comparing the customer's name and address with CVR based on the CVR number and suggesting corrections.
  • Sorting: suggesting category, product group or group affiliation for new records.
  • Warning: flagging records without a CVR number, products without a unit, or customers created twice on the same day.

What AI can't do: decide who owns data, guess values that don't exist, or merge records without checks. A wrong merge can be hard to roll back, so the final approval should rest with a person.

How to go about it

  1. Choose one type of master data to start with, typically customers or products.
  2. Name the source system and owner for that type.
  3. Decide the mandatory fields and formats, for example a CVR number for business customers.
  4. Clean up: find duplicates and correct errors, preferably with help from registers such as CVR.
  5. Connect the other systems, so changes are made in one place only.
  6. Measure quality on an ongoing basis, for example the number of duplicates and records with missing fields.

Agent360 can help you choose a solution or build the integration layer that keeps master data consistent across your systems, so contact us if you'd like to talk it over.

Sources

Frequently asked questions

What is master data?

Master data is the products, accounts and parties that your business transactions are about, for example customers, suppliers, products and locations. An order is a transaction, while the customer and the product on the order are master data. MDM makes sure this data is consistent, correct and has an owner across the systems.

Does a small business need to buy an MDM system?

Often not as a first step. Many smaller businesses get a long way by naming one system as the source for each type of master data and enforcing fixed rules there, for example a mandatory CVR number and checks for duplicates. A separate MDM tool makes sense with many systems, several companies or countries, or after acquisitions.

Can you use CVR data for marketing?

Not towards advertising-protected businesses. Under the CVR Act, private businesses must not use basic data on advertising-protected businesses for direct marketing, and if you pass on contact details, the protection must be clearly marked. That's why advertising protection should be a field in your master data.

#mdm #master data #data quality #cvr #integration

Which of your manual steps could disappear?

We map where data is moved by hand between your systems, and what can realistically be automated.