It may sound counter-intuitive but the “what” of metadata is probably the least important thing.
Why do I say that?
We, let’s start by looking at what metadata is.
Metadata is information about other data that gives it context.
Two common examples are the data dictionary and the business glossary.
The “what” of the data dictionary is a list of tables and columns, the “what” of the business glossary is a list of terms and definitions
From a pure technology perspective it is tempting to begin building these lists – after all, any documentation is better than no documentation
However, in their own right these lists may add very little value.
In fact, in a large organisation these kinds of lists can be overwhelmingly large – to the point where they in fact get in the way of value.
Start with the “Why?”
Rather than rushing into building lists start by understanding which metadata is really relevant. What are the business goals and objectives that depend on the data we are looking to understand.
If, for example, I am looking to improve customer experience, I may want to enhance my understanding and decision making around customers and the channels they prefer. This kind of critical thinking allows us to filter our extraneous data and build our metadata around critical data elements
Of course, each organisation will have its own priorities.
A good starting point would be to identify the four or five key reports that go to the board each month. Is there a clear understanding of what each report is measuring, where the data comes from, and the trust worthiness of the data? By building this understanding for each key report you help the board to have confidence in the reports that feed their critical decision making, whatever the context.
Or you could focus on a key project – for example a new billing engine, or a regulatory requirement such as PoPI.
The key is to understand that metadata is only relevant if it is useful.
The “Why” gives us this.
Move to the “Who?”
In he above example, I suggested involving the board in your metadata program by starting with their priorities.
This may not be appropriate for you, or your organisation.
However, what is critical is an understanding who will use the metadata captured.
Engaging with multiple stakeholders helps us to understand the impact of changes to a data element or report, whether it can be re-purposed, why and may also expand our understanding of “Why” it may be important.
Most importantly, this stakeholder engagement sets a standard for accountability for data.
As we move toward the digital enterprise data is an increasingly shared asset, used for multiple purposes. Understanding who uses each element, and for what purpose, starts to break down data siloes and build a data-driven culture.
Worry about the “Where?”
Finally, we do need to worry about context.
Understanding the where of metadata is to build up a map of critical data elements and how the affect our business.
Where is this report generated? Which business process uses this element? Which is the master source for this metric?
Every metadata element is linked to others in a myriad ways. These relationship, again, will feed into the “why?’ and “who?” questions.
As more context is created so the value of our metadata increases – but so too does the complexity.
Of course, one can ask many questions as one builds our metadata capability. But as one moves beyond technical documentation to deliver business value it is clear that the “what” cannot be enough.