Master Data Management and Information Modeling: On End or Means

Good news! MDM is alive! I published 2 blogs on Master Data Management (MDM): Benefits and implementation. As it turned out this subject is popular, because a hose of reactions has come my way. A large number of them deal with data and information modeling. However, the question is whether data modeling is a means or an end.

Quite a lot of the reactions on my MDM blogs originate from enterprise information architects. They claim that MDM can only contribute to the business if it is preceded by creating a perfect information model of the same business. I am not sure if I agree on this.

Somewhere in the eighties of last century I have given advanced classes in information modeling using the data modeling methodology NIAM (Nijsens Informatie Analyse Methode) and ERD (Entity Relationship Diagramming). NIAM is based on the language theory as developed at least 90 years ago, and is a very strong method for information modeling approach. A few years later, dr. James Martin enriched ERD, part of the Information Engineering (IE) methodology toolbox, with some of the strongest concepts of NIAM, never approaching its elegance and sophistication.

Especially in NIAM it was considered an art to specify business rules as much as you could in static data model rules. This makes sense, as there are less business rules left to program. Top of the art was when you created an information model that represented the universal model of everything. Such an approach had a few drawbacks though:

• The information model became extremely abstract, taking a lot of time to understand, and even much more to change it.
• The group of persons who could work with it was extremely small, and therefore the risks high. If someone left the project, it could take months to reach the old level of fluency.
• Programming against such a universal model was extremely difficult. Testing it and implementing it were real implementation challenges.

If brought to live, such a model is timeless. I happen to know that some of those models are still in production, after more than 25 years unchanged, and they still easily outperform today's customer relationship management systems, and handle changes of business requirements without any problems with a extremely short implementation time! The data model is stable, no programming required, and only configuration is needed.

With Information Engineering (IE) and Information Engineering Facility (IEF) the situation of implementing universal models improved drastically, because programming was done with Action Diagrams working against the information model, and generation the full source code automatically. There was no longer any need to transform information models manually, and manually program against this transformed implementation model. This approach removed the previous double inhibitor.

Readers of the Results2Match blogs may have noticed that we prefer the simple approach. A tool should serve the work (Peter Drucker). Keep it simple, and let MDM make your life easier, not more miserable! Avoid overly complex implementations. MDM is supposed to serve the business strategy and business interests, and enable those. Focus on that, make MDM the 'vehicle' is should be, such as a carrier for real business improvements like customer self-care solutions.

Therefore I stick with my earlier conclusion: Also with MDM, business benefits should lead you to business requirements, and not the universal model of everything.

What are your experiences with MDM? Please share your experiences with the Results2Match Community! We welcome your feedback!

This blog is part of a series on Information and Data Governance:

1. Data Governance for Dummies: Part 1 (Business Requirements)
2. Data Governance for Dummies: Part 2 (The Replica Data Security Solution)
3. Master Data: Business Asset or Cash-flow Burner?
4. How to beat the Master Data implementation challenge?
5. Master Data Management and Information Modeling: End or Means (this blog)
6. Information Governance Solves the Challenges of IT Governance and Data Governance! Or?

Results2Match has a strong vision on successful business management solutions and result driven implementations.

This blog is written by Hans Lodder. Hans is a very experienced management consultant and interim manager. You can contact Hans by email.