The importance of data governance is fast ballooning, as information piles up amid burgeoning compliance obligations, marketing strategies and budget pressures. Data governance remains a somewhat ambiguous discipline, however, in part because there is no consensus on its definition or how to turn it into a system of best practices, writes Jill Dyché, co-founder of Baseline Consulting.

The term "data governance" is widely used by business executives and IT executives alike, but it isn't always clear who should be in charge of it and what kind of work it entails, Dyché writes in a post at Baseline magazine. On one hand, it is generally understood that the business owns the company data and is the driver behind its policies and oversight. On the other hand, there is a data management component to governance, and it falls to IT to execute the policies. 

Business units are catching on to the idea of data governance as they seek easier access to information they need. "I've developed my own database to protect myself from other people's databases," said a clinician at a health care operation. "I can't trust other people's interpretation of data about my patients. If I have to stay late and maintain my own Access database, that's what I have to do."

It is a mammoth undertaking to locate, collect, annotate, consolidate and deliver relevant data to the many units and systems within a large organization, and the effort sometimes ends up involving costly replication and unnecessary risk. To mitigate these problems, IT's role in enabling data governance has to be made clear. Establishing this function is always a great start, but making data governance a sustainable set of practices has been a more challenging task. 

"Often a visionary manager would convene like-minded people on both the business and IT sides to agree that data was an asset, that data quality was poor and that someone needed to clean it up. The natural next step was to convene a data governance council, the de facto decision-making body for data governance," Dyché writes. "Then things got really quiet."

The main challenge was in turning a big mess of siloed data--often duplicated, unusable or irrelevant--into a strategy that would serve the whole organization, and often there was no tactical plan in place. Companies have had a hard time getting past debates over data ownership and priority. Dyché offers several practices that companies should consider before delving into a data governance initiative.

You should start by identifying the business problem behind the need for data governance. Next, scope out the project road map and define ways to measure the project's value. Get senior level support. Avoid being overly ambitious at the outset, and be sure that at least part of the business problem is addressed before you try to turn data governance into a organization-wide initiative.

For more:
- see Jill Dyché's post at Baseline

Related Articles:
How NOAA handles 80 TB of data a day
Data management saves trucking firm $6 million a year
Five steps to tackling the data challenge
Why "data scientists" are in demand