Over a million developers have joined DZone.

What's in Store for the Future of Master Data Management (MDM)?

DZone's Guide to

What's in Store for the Future of Master Data Management (MDM)?

Learn the important changes that MDM will witness in the coming years. The use of cloud, machine learning, and multi-domain models will change the way MDM systems work.

· Big Data Zone ·
Free Resource

The open source HPCC Systems platform is a proven, easy to use solution for managing data at scale. Visit our Easy Guide to learn more about this completely free platform, test drive some code in the online Playground, and get started today.

Master Data Management (MDM) has been around for a long time, and many people, like myself, have been involved in MDM for many years. But, like all technologies, it must evolve to be successful. So what are those changes likely to be, and how will they happen?

In my view, MDM is and will change in two important ways in the coming years. First, there will be technological changes, such as moving MDM into the cloud or moving into more 'software as a service,' or SaaS, offerings, which will change the way MDM systems are built and operated. Secondly, there are and will be more fundamental changes within MDM itself, such as moving MDM from single domain models into truly multi-domain models. Let's look at these in more details.

Waves of MDM Change: Technical and Operational

New and disruptive technologies make fundamental changes to the way we do most things. In the area of MDM, I expect that to change in two main areas. First comes the cloud. In all areas that matter in data we are seeing moves into the cloud and I expect MDM to be no different. The reasons are simple and obvious, the move towards MDM being offered as a SaaS offering brings cost savings in build, support, operation, automation, and maintenance and is therefore hugely attractive to all businesses. I expect that going forward we will see MDM more and more being offered as a SaaS.

The second area I see changes happening are more fundamental. Currently, many MDM systems concentrate on single-domain models. This is the way it has been for many years and currently manifests itself in the form of a 'customer model' or a 'product model.' Over time, I believe this will change. More and more businesses are looking towards multi-domain models that will, for example, allow models to be built that have the links between customer and partners, products, suppliers, etc. This is the future for MDM models, and already at Talend, our multi-domain MDM tool allows you to build models of any domain you choose. Going forward, its clear that linking those multi-domain models together will be the key.

MDM and Data Matching

Another area of change that is on the way is in regards to how MDM systems do matching. Currently, most systems do some type of probabilistic matching on properties within objects. I believe the future will see more of these MDM systems doing 'referential matching.' By this, I mean making more use of the reference database, which may contain datasets like demographic data, in order to do better data matching. Today, many businesses use data that is not updated often enough and so becomes of less and less value. Using external databases to say, get the updated address of your customer or supplier, should dramatically change the value of your matching.

Machine Learning to the Rescue

The final big area of change coming in the future for MDM is the introduction of intelligence or machine learning. In particular, I forecast we will see intelligence in the form of machine learning survivorship. This will like take the form of algorithms which 'learn' how records survive and will, therefore, use these results to make predictions about which records survive, and which don't. This will free up a lot of time for the data steward.


Additional changes will likely also come around the matching of non-western names and details (such as addresses). At the moment they can be notoriously tricky as, for example, algorithms such as Soundex simply can't be applied to many languages. This will change and we should see support for more and more languages.

One thing I am certain of though, many of these areas I mentioned are being worked on, all vendors will likely make changes in these areas and Talend will always be at the forefront of development in the future of Master Data Management. Do you have any predictions for the future of MDM? Let me know in the comments below.

Learn more about MDM with Talend's Introduction to Talend's Master Data Management tutorial series, and start putting its power to use today!

Managing data at scale doesn’t have to be hard. Find out how the completely free, open source HPCC Systems platform makes it easier to update, easier to program, easier to integrate data, and easier to manage clusters. Download and get started today.

master data management ,big data ,machine learning ,data matching

Published at DZone with permission of

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}