Master Data is a Process not an Output

The term master data gets tossed around on social media as this opaque thing that is critical for business intelligence, analytics, and AI. However, most of the content poorly defines the term, and conflates master data across transactional and reference data. So in the spirit of clarity, master data are defined as the set of tables that are the lookups for your business, they store the attributes about your vendors, products, or customers. Essentially, they help describe your transactions.

An example of master data is Customer A is located in Pennsylvania and buys through a distributor, or Product X is in the perishable category and has a shelf life of 14 days. A closely related concept is reference data, which are the set of choices that can be assigned as master data. For example, the shelf life master data field could have multiple choices (i.e., reference data) such as 7,14,30, or 90 days. So master data and reference data are symbiotic. Finally, transactional data are self explanatory -- transactions are the activities (i.e., sales, purchase, transfer, cancellation) that are described by master and reference data.

In my career, the master data has been by far the biggest problem. There are many reasons for this but most of it comes down to a lack of ownership, dedicated time, and process workflows. For example, vendors, customer, or products can have dozens of attributes that describe them and those attributes span the whole business. Therefore, you need a collaboration between functions (e.g., sales, manufacturing, finance), supported by a workflow, to ensure all of the product's attributes are populated and accurate. To complicate things further, the attributes to transact the business are different that the attributes needed to analyze and plan the business. And in my experience the transactional attributes are usually given more organizational attention and are tend to be more accurate. So as it stands, master data is a complex, multi-functional business process akin to S&OP or new product rollouts. Few people think about it this way, and see data as an output rather than a transformational business process.

The key takeaway is that master data are complex and requires an enterprise approach. I can definitely help you assess your master data and recommend strategies to ensure your master data is accurate -- contact me today and we can discuss!

JimSwallow@AccuDataNow.com