What’s In Store for the Future for Master Data Management (MDM)?
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 important two 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.
The 4 Steps to Become a Master at Master Data Management now.
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.