By 
There has been a lot of talk about data management and the different types of data. Master data management, reference data management, digital rights management — those are a lot of DMs!
Naturally, as we leap into the age of AI and machine learning, it would behoove us all to get our data straight. After all, there are so many types of data out there: transaction data, research data, Big Data, metadata, enterprise data, customer data, golden data, data virtualization, data analytics, etc. To sort out all types of data is beyond the scope of this article. We can, however, discuss two very relevant types of data: master data and reference data. It is through exploring nuances of meaning that we can settle the master data vs reference data controversy.
Master Data
Master data involves data about companies and business entities. This data provides context for business transactions. Examples of master data include but are not limited to financial structures, parties, products, customer relations management (CRM), and locational concepts.
The science of maintaining and managing master data is known as master data management (MDM). Master data management is essential for verifying data sources and ensuring data quality.
The legendary, ancient Chinese general, Sun Tzu, was famous for his seminal text, “The Art of War.” His insights regarding strategy have been employed by armies throughout history, and today, these insights are employed not only on the battlefield but in the boardroom.
One of his most renowned quotes applies perfectly to the science of master data management: “Know thyself. Know thy enemy. “
Master data management is the science that allows a business entity to know itself. To be a true master of master data, you need accurate data on all key business information within an entire organization.
Knowing one’s enemy, which in the business world is knowing one’s competition, can be more difficult because that could involve some type of corporate espionage and might be illegal. There is no excuse, however, not to know one’s self.
Reference Data
Reference data is a subset of master data. Reference data relates to data that is used to classify other data. The idea is that as a database gets bigger and bigger, it can be more and more difficult to organize and aggregate raw data and Big Data into data sets.
However, if we have some constant metrics that would not change too often, like reference data values, we could sort out this raw data and Big Data into reference data sets. This is a reference data model. Examples of reference data include country codes, zip codes, kilograms to pounds conversions, and many other different things.
If, however, there is no external authority for this reference data, the data must be changed if the business changes (i.e., business transactions, business events, business activity). In that case, such reference data becomes internal reference data.
The science related to the study of reference data and its management is called reference data management (RDM). Without some type of reference data management to provide context, large hierarchies of data cannot be aggregated because there will be no units of measure that cannot be changed.
The Difference
Master data relates to personnel, locations, and things connected to an organization. Reference data relates to data that is used to classify other data.
Although both master data and reference data are used to provide context for business transactions, there is a distinct difference in function. Master data is related to the business entity itself. Reference data is related to classification and is just one type of master data. The truth is that both master data and reference data are essential for preventing data inconsistencies. Without maintaining data quality, other data sciences, like data analytics and business analytics, would not be possible.
Comments