Mastering Data Management

Bridge the chasm between technology and data governance

I’ve recently made a few assertions that haven’t stirred up much controversy. In fact, they seem to be fairly well accepted, and for that I’m thankful.  These assertions are:

1. Data governance is an iterative process whereby business leaders make declarations of policy etc. that, when enforced, will materially improve the quality of the data under their control.
2. These declarations cover the “who, what, when, where, why and how” of managing your most important data.
3. These declarations are codified as principles (why), policies (what, where & who), processes/procedures (how, when, where and who), business rules (what and how) and metrics (why)

If you don’t agree, let me know in the comments!

So, if we can generally agree on these declarations, let’s ask two questions:

1. How should these declarations be codified?
2. How do you tie these into technologies that are inevitably involved in CRUDing data?

Most declarations I’ve worked on have been recorded in some type of document, but some declarations are quite well-suited to technologies (seen the table below). The question is, How can we tie these together so that declarations are living assets and do not become “shelfware?”

It makes sense to me that we continue to codify these declarations in a form that most people can simply read (without having to know how to use a specialized design/development tool).

But these textual representations of intent, policy, constraint, etc. lose their value if they’re not tied to actual “assets” (services, messages, etc.) that Create, Read, Update and Delete (CRUD) data. This is the “great divide” in data governance.

So, I propose that we start defining declarations as text documents, like we do today, but that we transform them into some online, active form (like you’d find in a wiki). We can then begin tying key concepts to an underlying metadata repository to capture and share definitions of the people, organizations, roles, places, things, events, messages, processes, instrumentation, etc. that will implement the declarations.

The idea is that, instead of having shelfware or “brochureware” codified in a Word document or presentation that’s collecting dust and is completely passive, we should be able to have a document that’s online, with the ability to click on any word or phrase and find out the following:

• What does this term mean and where is it CRUDed?
• Who approved of this thing/event/process/action? How does it work?
• What are the conditions under which this is allowed or disallowed, under which it occurs or is triggered?
• What are peoples’ roles in this thing or event?
• Where is this implemented and is it running right now according to specification?
• What are the business implications of this thing happening or not happening according to plan?
• Where is this failing to meet standards and how often is this happening?
• Who/what is affected by problems associated with this thing or process?

Although space doesn’t permit me to elaborate on these (yet), does this make sense to anyone? What would you add? What would you challenge?