Here’s where I attempt to discuss governance as it relates to my core coverage domain: data management.
Data governance is a new buzzphrase with legs. I notice that three other industry analysts are posting to a blog sponsored by a data management vendor, and that they’re all tap-dancing around the topic of data governance…none of them has produced (in that blog) a clear definition of the term, though they’ve gone on at some length regarding the value of data governance, the “what’s it’s not” of data governance, the “what it sorta overlaps with” of data governance, and so forth.
I’ll back into my own definition of data governance. First, I’ll revisit my definition of federation, as one broad category of governance structures:
- “Federation is a governance structure in which autonomous domains choose to honor each other’s decisions and accept each other’s assertions in some realm of human endeavor—such as identity management, data management, or SOA management--subject to business contracts, trust relationships, interoperability agreements, and local policies.”
I implicitly describe “data management” as a “realm of human endeavor” to be governed (i.e., controlled). And I define governance in another post as:
- “Control structures on human and automated interactions, some of which emerge from the blur of decentralized, autonomous decision agents, and some of which are imposed by centralized authorities.”
Leveraging, converging, and extending these definitions, I define data governance as:
- A control structure on human and automated interactions within and among data management domains, addressing the full life cycle of functions necessary for comprehensive management of data as a business asset.
I have spun my own alliterative string of verbs to describe the various life cycle functions managed by a data governance environment:
- Mapping, modeling, and marking up data
- Moving and migrating data
- Massaging and manipulating data
- Massing and mastering data
- Monitoring and measuring data
- Mobilizing and extracting meaning from data
And so forth. Mmmmmmmmmmmmmmmmmmnemonics. Governing anything involves getting your head around a single conceptual model of the entire domain. I parse every data management vendor, architecture, approach, product, etc with this ontology in mind. ETL? (that’s primarily moving and migrating data). Data warehousing? (that’s primarily moving and migrating, massaging and manipulating, massing and mastering data). Business intelligence? (that’s primarily mobilizing and extracting meaning from data). DBMSs? (a bit of everything, actually). And so on and so forth.
Data governance is being used in the same breath as master data management (MDM) to describe this entire life cycle of data management functions. Now, repeat the mantra: mdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdmdm.