Skip to main content


Service Deskin kehittäminen

Common pitfalls of choosing an MDM platform

Content sections

Common pitfalls of choosing an MDM platform


If you today own Dell Boomi or OneData and think you have a Master Data Management-platform, you actually don’t. Read this blog to understand what you have and what gaps in expectations and capabilities you need to address.

If you already have a Master Data Management-platform, that is a good start! You might actually have one, but it is not a given.

If your Google skills are timed and accurate you just now found a blog that will save you a lot of headache. You are not doomed if you thought that you bought an MDM-platform but ended up with an integration platform, but you have dug yourself a hole that requires some effort to get out of. Read this blog to find out what to look for, and especially what pitfalls to avoid.

I have talked to several customers over the years about master data management, only to find out that they indeed are paying licenses for something labelled as a MDM  but actually is nothing more than an integration platform (as if that was a bad thing 😊).

In 99 % of the cases, integration MDM is only handling reference data, not master data. As this blog explains, “The Master Data Model vs the Information Model”, master data management has different meanings and thus different approaches to handle.

In short, the reason is that, an integration platform works with an Information Model that is tightly coupled with how information is stored and processed in IT-systems. A master data model expresses how the business views and defines its own concepts and entities for governance purposes.


The Domain Approach - Analytics MDM

The Domain approach is a top-down-perspective on Master Data.

It is also known as Analytics MDM, since it usually is related to a strategic analytics initiative. One or several Master Data Entities are identified by a selection based on one single question: What information is business critical?

Business critical information is usually customers, products, assets, employees etc. Each entity being its own domain. The Master Data entities are often core elements of business processes, affected by legislation, guarded by security requirements or poses a financial risk. In short, needs loving care and strict governance.

Together, the master data entities, its definitions, attributes and relations make up The Master Data Model. This model exists to ease and support business ownership and stewardship of business critical information.

An MDM-platform that supports the domain approach should be structured around the following capabilities:

  • Multi-domain modelling
  • Information ownership
  • Master Data Entity definitions and change management concerning the definition
  • Policy definition and enforcement, e.g. security, retention, GDPR, data quality etc
  • Stewardship interface
  • Monitoring of data quality and workflows
  • Mapping of Master Data Entity members to compliant source system data members.

This MDM-platform type aims at providing the business with a tool for the daily maintenance of their critical information assets. The outcome of this process is that any domain member has a Golden Record (clear definition, policy compliant, high quality) available for subscribing systems and processes.

Multi-Domain is important. There are actually MDM platforms that are single domain, resulting in that you have buy another platform for the remaining domains. Crazy, I know!

In the Domain approach, the Master Data Model is used as a looking glass against source system data, and the model acts as a bridge between related data in separate sources.

The Master Data model very much becomes a set of business rules that guides comparison and, both conceptual and technical integrations.




The Transaction Approach - Integration MDM

The Transaction Approach is a bottom-up perspective on Master Data.

One important note is that from an integration perspective all non-transactional data is Master Data, which is not correct. You must separate Master Data from Reference Data, otherwise you will end up buying the wrong tool for the job.

An Integration Architect instead matches Master Data on data point level trying to solve the integration one data point at a time. Also, the Transaction approach address the questions of “where is my data?” and “how do I connect to it and move it somewhere else?”

An MDM-platform that supports the Transaction approach should be structured around the following capabilities:

  • Data source analysis
  • Reference data identification
  • Golden Record-record creation (model)
  • Mapping of Reference Data Entity members to compliant source system data members
  • Integration rules for match, merge, deduplicate etc
  • Integration data flows from-and-to data sources

This MDM-platform type aims at providing IT with a tool for the daily maintenance of reference information assets.


The Domain Approach - Analytics MDM (operational)                                                       

Analytics MDM actually has two flavours

  1. Analytics MDM for analysis purposes; the Golden Record is only used in the analytics platform for accurate decision making (uni-directional)
  2. Analytics MDM for operational purposes; the Golden Record is used both in the analytics platform and feed-back to the operational systems to improve (bi-directional)

The Analytics MDM (operational) approach is the Holy grail. Platforms that support Analytics MDM for operational purposes, fulfil the combined business driven master data governance requirements with integration capability requirements, including bi-directional data flows.

To implement an end-to-end Master Data Management platform for operational purposes, you must both establish the business ownership, -definitions, -policies and governance model as well as chose an integration platform to build solid, resilient, low-payload and trusted integrations that actually brings it all together on the lowest technical detail.

This certainly is not as easy as pulling a rabbit out of a hat. Especially since it requires a very structured and thorough process that does not leave room for skipping any level in the above pyramid. Every layer and details need to be handled properly. That said, you can still do it in an agile fashion.

When high-quality, governed and trusted master data is available in the operational systems, the number of suboptimal manual workarounds, duplicated in each silo is reduced to a fraction. The volume of errors and faulty decisions due to dirty data is pretty much eradicated.


Platform guidance

So, what MDM platform do you have? Did you require an Analytics MDM (analysis) platform but actually got an Integration MDM platform? Then you know what gaps to fill.

A quick rundown of the major players and what their platforms actually deliver is shown in the picture below:


MDM platforms


Do you want to know more about this? Please don't hesitate to contact me.



Magnus Hagdahl is a principal consultant at Enfo