“What is the definition of ‘customer’?” How many meetings (and wasted hours) have been spent attempting to answer this question? After being involved in literally hundreds of conversations regarding the never-ending process of resolving critical metadata definitions, I never tire of the opportunity to level-set about data governance, stewardship, and the metadata/harmonization albatross.

Perspective 1: Even asking the question “What is the definition of ‘customer’?” betrays a gap in assessing information use horizontally across the organization. The fact is that the precise semantics of terms like “customer,” “product,” “part,” or “supplier” erodes as each term is used in more contexts. And since everybody already has an innate understanding of what the term means, there is no reason to be explicit about its definition. The only problem happens when stakeholder A wants to use Stakeholder B’s “customer” data for A’s own purposes, and it turns out that they are actually talking about two completely different things, with different definitions, attributes, and quality characteristics.

Perspective 2: Business applications are typically architected to meet functional requirements supporting transaction processing or operational activities. Analysis is an afterthought, and relies on grabbing data from the operational systems.  Yet the advocates for metadata are the data warehouse/BI folks, grasping for some meaning onto which to hang their hats (ok, a bit of a mixed metaphor, but cut me some slack…).

Perspective 3: What is the level of effort necessary for effective metadata management? OK, do a thought experiment – In your organization,

• How many applications are there?
• How many databases does each application touch?
• How many tables are there in each database?
• How many attributes in each table?
• Multiply these numbers together – that is the number of data elements to be cataloged in the metadata repository.
• How much time does it take to rationalize a data element’s metadata? Let’s speculate optimistically with 2 hours each.
• Multiple the number of data elements by 2 and that is the number of hours it will take to get your metadata cataloged.

Perspective 4: When is metadata used? Despite all the expert advice about enterprise metadata and its value to the organization, we are still seeing some questions regarding its use, except for specific purposes such as developing a reporting and business intelligence program that is rationalized with the business applications.

From these perspectives, it becomes a good idea to limit the scope of your metadata activity and integrate it much earlier in the system development life cycle. Define your data concepts before you build the databases, and that will simplify the associated activities (such as source-to-target mapping and data migrations).

But getting back to the first question about what is meant by “customer” – we’d love the ability to put on a pair of 3-D glasses and have all the different versions merge together into one beautiful, rationalized definition that everyone can share. But that image is just an illusion: there is no one definition for most of the commonly-used terms. More likely, there are qualified concepts that are distinct and business processes can benefit from a differentiation and a precise definition for each.

Instead of one customer, there are “prospective customers,” “good customers,” “business customers,” “residential customers,” “wholesale customers,” (and the list can go on). Each has attributes that are slightly different (and relevant) to the consuming business processes. These qualifications are based on the business use, and the differentiation process also will reveal the common characteristics that benefit the downstream data consumers. So use those 3-D glasses in reverse: look at what you think is one concept and try to break it out into its component (red and blue blurry) visages, and then figure out why they don’t all mean the same thing. That may do two things: reduce the amount of time arguing about that single definition, and creating more clarity on what things really do mean.