Master Data Management: Techniques & Organization

Un texte plutôt court, un peu technique en langue anglaise, inspiré par la littérature que j’ai absorbée et avant tout mon expérience de 8 mois maintenant dans le domaine. Le MDM n’est pas seulement un domaine technique, c’est aussi du fonctionnel… et de l’humain !

As I introduced it some time ago, the management of Master Data is required to improve the quality of information (accuracy, availability) used within a company, and finally to get the business taking advantage from this consistency.

Now, it is interesting for You, dear readers, to understand that Master Data Management (MDM), is not only a technical area, but is also interested in very functional things, as well as involved in the human responsibility & knowledge! In fact, experts mention this as a triplet of ”People, Processes and Technologies”.
Indeed, the definition and consolidation of standard information will necessarily lead to discussions about who, where and how the master and referential data are managed in a company.


So, the first step of a MDM approach consists in knowing the current major processes, as “Order to Cash”, “Plan to Inventory” or “Procure to Pay”; unfortunately this appears sometimes to be unclear within companies. So more focused processes like “Order creation process”, “Part update process” or “Customer deletion process” can be interesting starting points to understand how business people handle data with systems, and with highlights to inputs and outputs of their work. Generally, given functions own those processes (as Purchasing would be responsible for managing Vendors), but actually even these smaller parts of process are transversal (Accounting is also involved for Vendor management). And this is why the MDM team is supposed to be very integrated in the company.

Once a process is identified, the flow of data known, the interest should slide to duplication and accuracy of information, its implications and the possible solutions. This enters now the realm of MDM, which will then consist in chasing duplicate records, duplicate locations and improve process and tools to guarantee a better accuracy and completion of information.

One way to reach a correct quality is to centralize the so-called “Master Data” (Parts, Suppliers, Customers, Chart of Accounts, etc) by consolidating the existing databases, cleaning information (e.g. remove duplicates), and then distribute correct information all over remaining front-end systems and govern data through validation & workflows. So maintenance of data quality is ensured.
This can lead to reviewing the tools landscape, as setting up a MDM database with interfaces to connect it to the legacy systems. Some of the challenges in there can be the part numbering, the location where to manage information and obviously the technical connections between tools.

All of this is MDM, and even further more that will be detailed here later.

Ciao
MfS