It's clear that there are a lot of challenges in providing good, consistent data to the business. People misuse and misinterpret data. They don't measure data quality. They don't align data access with business policies. They don't make data available in a sustainable way. Plus, the lack of accountability and workflow tools renders data governance a paper exercise, where artifacts are generated and then forgotten on a shelf somewhere.
To use a somewhat well-worn but apt example, there's a Maslow's hierarchy of need here. First we need to understand what terminology and definitions are for enterprise data. Only then can we introduce data management, including the demarcation of the rules associated with the data and who's allowed to see it.
Data governance is really a framework for supporting new rules, definitions and policies and to support changes in those rules, definitions and policies. It's a series of SLAs for data. Taken that way, there are different components. You need to deconstruct, identify, and put processes around those components. This not only helps you know what to do, but what not to do.
This was first published in September 2009