Intelligent Enterprise

"Governance," that wonderfully over-used term, is something that data management professionals love to preach. But to what extent do we practice it in our own professional activities, and how do we go about achieving this laudable objective?

Listen to a presentation from any data product or service vendor or to a speech from any data management professional – especially, one that is up to speed on the latest jargon – and you will hear the term "governance" mentioned sooner rather than later. Pay closer attention and you will find it's typically mentioned in the context of Master Data Management (MDM) or Customer Data Integration (CDI), and seldom otherwise. And therein lies the problem.

Wikipedia – that modern Bhagavad Gita, Bible, Quran, Torah, Tripitaka and more rolled into one – defines governance in the context of business as the development and management of "consistent, cohesive policies, processes and decision-rights for a given area of responsibility." But there is no reason that governance should be limited to packaged solutions like MDM or CDI.

Let's take the area of customized application development. There clearly needs to be a governance framework around project data management, which consists of roles, responsibilities, activities, policies, procedures and standards as applied to all data-related activities on a project (including data modeling, database development, data loading, data integration, database change control and database promotion across development/test/production environments). Too often business users, project managers and others take data and related activities for granted: developers create or modify tables as they go along, policies for test generation are prominently absent, with developers wasting valuable time on account of data conflicts (the infamous "junk data" in the development/test environment). Data quality assessments and awareness are an after-thought, and adequate data/ETL-related tools are lacking. Data issues are a widely known (and easily provable) contributor to project delays and even failure. Yet, project after project, we fail to learn from our own and others' mistakes, and to take pro-active steps to address issues that we know are bound to arise. There are many reasons for this phenomenon, but certainly an important reason, often times, is the lack of that pesky governance framework.

The Data Management Association (better known as DAMA) is putting together a Data Management Body of Knowledge (DMBOK) that should help data management professionals practice what they like to preach. DAMA recently released the second version of the DMBOK Framework, which spells out how the body of knowledge will be organized – the DMBOK itself is in the making, with the help of a large contingent of contributors. Loosely modeled on the PMBOK from the Project Management Institute, the DMBOK is intended to provide guidance on data management functions, best practices and vocabulary without detailing specific methods and techniques. In other words, it should help lay down some method to the madness, without requiring us to sacrifice our creativity. Sounds good!

Rajan Chandras is a consultant with a global IT consulting, systems integration and outsourcing firm, and can be reached at rchandras@gmail.com.