Master Data Management may be considered at best, mundane and boring. But the topic is grossly underrated, especially by those who aren’t directly impacted by data governance challenges in their respective organizations. This is not an unusual situation when the data in aggregate holds obscured problems that only manifest when you reach or make use of a particular piece of the whole data.
Despite the prevalence of complex computing systems and massive amounts of diverse data across all shapes and sizes of businesses, the reality is that master data, in its many forms, touches every industry and operation in varying degrees and often at an individual contributor level in ways that may not be clearly understood.
At the lowest level, we can relate to master data in retail when we attempt to buy something using a barcode, only to find that the barcode has no associated price lookup in the checkout system. That issue can be written off as a pricing or point-of-sale issue but in reality, it is still a master data issue.
Data Management Resiliency
Many may also underestimate the significance of a robust and resilient data management approach to data collection, data handling and data distribution (what Pretectum refers to as data syndication). This is particularly important when you consider the customer master.
Imagine a situation where, every time you engaged with a financial institution, you had to provide all your vital statistics to deposit or withdraw money from your accounts. Imagine too, a situation where you don’t have your account details readily at hand, and the only way that they can identify your accounts is by providing all this data. If they don’t have all that data attached to those account IDs then finding the right accounts to do withdrawals and deposits or inquiries, is a problem.
Technically, we solve for these issues with indexes, account codes, and classifiers. In the past, we might have coped with data inadequacy given that we had smaller volumes of data and smaller numbers of customers but today that’s simply not practically possible to do.
At its most basic level, the customer master facilitates multi-dimensional categorical identification and association of a customer entity with one or more other customer entities, transactions, and other data assets.
As a discipline, customer master data management requires several key aspects of the data governance and data management processes and tools to be in place
A central repository of defining attributes or metadata about how customer master records should look is one key to understanding how the customer dataset should look and feel as a whole. In the past, this might be handled by a data dictionary but it could be defined and documented in several different ways. The question is how would you enforce this definition for the examination of existing data and any new data that you capture.
This central repository could be a system of record, a specialized master data management system, or a data lake or data warehouse that is curated using a variety of tools. This will define things like honorifics, name data, contact data, and other key customer attributes all tied to a customer key or account number.
Your MDM central repository is the home to the metadata (data describing data) and the data itself.
Transparent data maintenance
It’s essential that the way that data is created, maintained, used, retired or archived, is well understood and consistent by all impacted teams within an organization.
People should know what they have to provide as minimum data and what values are considered acceptable. Some sense of understanding around interdependency is also useful. Most importantly, the critical data that is needed for the business should be defined, called out, and guaranteed to be provided and maintained to provide data quality assurance.
If decisions need to be made by teams or a hierarchy of decision-makers then that should be a part of the customer master data definition.
Consistent data syndication
Master data that resides in yet another silo is not that useful except as a point of inquiry or reference. What is needed is a way to disseminate and distribute the data held centrally to all people, systems, and services that need it, appropriately and consistently.
A data syndication facility needs to be provided by any data management system that allows the data to be passed to wherever it is needed as single records, groups of records, or whole datasets according to the specific needs of those downstream participants and processes.
Continuous control
The reason that you define the schema, the rules, the validity criteria, and then the processes around the collection and distribution of the data, is that you want to ensure that you have a consistent picture of the customer master at all times.
You need to decide if a system of record or your MDM is the master and the other systems are all subordinate or recipient systems. A model wherein the MDM is merely an assessment, reporting, and remediation platform is an ok approach as long as you continuously work on improving the system that you deem the master.
If you choose the MDM to be the controlling system then the effectiveness of control is entirely dependent on your MDM configuration.
Learn more about how Pretectum CMDM can meet your needs for customer master data management by contacting us today.