4 Master Data Management implementation styles

Do you have your customer data stored in multiple systems? Then it is a good time to think about Master Data Management. Since data is stored in several places, it is difficult to find all your customers' information. Moreover, this makes maintaining these data very complex, for example, when moving house. And checking your customer database against sanction lists is no easy task either.

Those who start thinking about Master Data Management are often shocked by stories about complex, expensive and long-term projects. Unfortunately, this is often true, but usually due to too high ambitions, a too complex product or an approach that is too large-scale. In order to achieve success, it is important to grow in MDM step by step.

Registry and consolidation style

Step one for a successful MDM implementation starts with the choice of MDM style. Several styles are distinguished and for a first step, it is advisable to choose the registry or consolidation style. In these styles, the data from multiple sources are brought together, matched with each other and Golden Records are created. A reference index is also created, which provides insight into which sources a customer is included in.

In the registry or consolidation style, data is often supplied in batches, for example once a week or once a month. With this style, timeliness is often less essential, but you do have quick insight into the overlap of data from the various sources (duplicates) and the quality of the data. Mutual differences become clear, so that mutual optimisations are possible.

If you are dealing with many different source systems, start with, for example, two or three sources and gradually expand. Start small, deliver results quickly and then build on the customer view. In this way, you will create visibility in the organisation, others will see the added value and you will be able to enjoy the benefits more quickly.

Coexistence style

In the first two styles of MDM, the focus was mainly on the consolidation of data from multiple sources. A useful insight, and based on this, it is of course possible to update and synchronise data in the various source systems.

Are you ready for the next style? Style three goes beyond consolidation and provides for continuous synchronisation of the source systems. This style is called the coexistence style.

In the coexistence style, the different systems continue to maintain their own data, but offer their new and changed data in real-time to the MDM system. Naturally, matching is also done now to trace corresponding customers, creating new Golden Records and updating existing ones. The MDM system therefore always provides an up-to-date picture. By setting the right rules, it is possible to determine exactly which source system is leading for each piece of data. Of course, the priority can also be determined on the basis of a last change date, so that the most up-to-date address is always recorded in the Golden Record. A combination of merging rules for the compilation of the data of the Golden Record gives the best results.

The core of the co-existence style is that every change to the Golden Record is immediately fed back to the source systems that are linked to that Golden Record. This makes it very easy to process, for example, a change of address or to enrich data known from another source. Of course, there are still points of concern. A system for pensions will not automatically adopt a change of date of birth or gender, unless that information is supplied from the basic register of persons (formerly GBA), for example. The rules of an MDM system must take this into account and must, of course, be configurable.

A MDM system based on the coexistence style provides you with an excellent basis for setting up a My-environment. Your customers can use this environment to manage part of their data independently, and by updating the Golden Records with this information, your source systems are also immediately updated. At least, if you choose to.

Centralised data

Earlier, we talked about the first three styles of Master Data Management. The first two styles are the registration and consolidation of data into Golden Records and the third is the coexistence style whereby changes can be made in both the source systems and the Master Records. Of course with a continuous synchronisation of these data.

The ultimate in Master Data Management, according to some, goes one step further. In this phase, there is Master Data that is stored and managed centrally. This data must be used by all connected systems. As a result, changes are only made to the Master Data. This creates a true Master - Slave relationship.

In the ideal situation, the data is only stored in the MDM system. Unfortunately, it is not always possible to formalize the Master - Slave relationship so explicitly. Many organisations work with standard packages and these usually cannot function without the relationship data being recorded in that system.

By only permitting mutation of Master Records in the MDM system and distributing those mutations from there, the uniformity of data remains guaranteed. The connected applications are obliged to adopt that data. By using a Master Record, these applications will subscribe to that record and thus remain continuously informed of changes.

We have talked about the flexibility of the data model. Now that MDM is no longer a collection of data from supplying source systems, another, more normalised, data structure can be chosen.

Another aspect that deserves attention is data quality. When centrally managing the Master Data in the MDM system, it is essential to carry out the validation and standardisation of the data there. Control at the source, i.e. at the point of input, is important. A First Time Right strategy, provided with adequate controls, provides for this.This functionality must be integrated with all input and change functions so that the necessary checks can be performed at the source.

Human Inference's DataHub offers a solution that also supports this phase of Master Data Management. We also offer adequate solutions for data quality control at the source, First Time Right.