Strong data governance frameworks are fuel for analytics

As organizations become more data-driven, implementing a data governance framework is integral to their success with analytics.

Data governance is critical to keeping compliant with government regulations and avoiding the penalties that come with violating such rules. But beyond preventing potential problems, a good data governance framework can serve as an enabler of analytics success and empower more employees within an organization to confidently work with data.

Without a data governance framework that enables users, data can be a source of fear and doubt. But with a data governance framework that inspires confidence, data can be fuel for the kind of curiosity that leads to insight and action.

That was the message of a panel discussion on data governance during Domopalooza 2021 -- the virtual user conference hosted by analytics vendor Domo on March 24.

"Data governance is hard," said Wayne Eckerson, founder and principal consultant of Eckerson Group and moderator of the panel. "It's very hard to do, and that's why a lot of organizations don't do it very well."

One of the main reasons it's difficult is that it's needed in order to do two things that are seemingly in conflict with one another, Eckerson continued.

"We've got some polar opposite forces working," he said. "One is the need to control data, and the other is the need to empower people and open data up, so how we balance the control and empowerment is really the essence of data governance."

One of the main hurdles many organizations face with analytics is simply convincing decision-makers that a data governance framework is necessary before they suffer the consequences of not having one, whether regulatory noncompliance or disorganized data that inspires a lack of confidence in decision-making.

Another is simply defining data governance and establishing the goals of a data governance framework.

Perhaps the biggest goal is ensuring the quality of the data, according to Dave Luna, senior technical specialist for data analytics at CAE USA, a technology training vendor for U.S. defense and government agencies.

Data governance is hard. It's very hard to do, and that's why a lot of organizations don't do it very well.

"The first iteration for everybody on their transition to a data path forward is whether their data quality is good," Luna said. "Everybody loves things like a free text field that gives them power, but they're so dangerous if you have no controls, no definitions, no monitoring of them. My focus on data governance started with quality, cleaning up the data. Data has to be right before it can tell you the right thing."

Another goal is to make data easy to find, and to that end consistency is a key part of any data governance framework, according to Kelsey McMahon, manager of marketing analytics at CME Group, a global markets company that is the world's largest derivatives exchange.

"Where we started our journey around data governance is around definitions and naming conventions," she said. "That makes it easy to find that data, you're not duplicating the data, you're not duplicating [extract, transform and loads] and wasting time. That's where we focus our journey."

Finally, access control is vital to a successful data governance framework, according to Ben Schein, vice president of data curiosity at Domo. And that access control is not solely to prevent people from working with sensitive data but also to enable them to confidently work with the data that is appropriate for them in the their role.

"It should be done with a lens toward empowering people and not keeping people out, not being a gatekeeper but being a shopkeeper to let people in and securely do [their work]," he said. "If your access is limiting how people can engage with data in a way that is counter to your business strategy, it's dangerous to your business. Governance needs to be strategy."

Once organizations buy into data governance and establish a framework, the framework has the potential to expand what they can accomplish with their data.

Luna said that before CAE USA adopted a data governance framework, employees lacked confidence in data they didn't prepare themselves. They questioned the quality of each individual data point, slowing the decision-making process, and rather than take advantage of modern BI tools that speed up the analytics process, reverted to their own Excel spreadsheets.

Once confidence in the data was established, however, employees felt better about data exploration.

"Before, no one wanted to talk about what the data told them," he said. "They were always questioning where it came from and saying, 'You need to prove that to me first.' Now that we've got a good [data governance] foundation, we're instead getting people asking what else the data can tell them. It really matters that we've implemented the data governance."

Similarly, Schein said that before the implementation of a data governance framework employees often fear that the data they're working with is incorrect. In addition, someone might be afraid that they're looking at data they shouldn't be seeing.

"Data is a very emotional thing, and data governance is part data therapy in a lot of ways," he said. "A big piece of data governance is mitigating that fear and giving the confidence to explore."

While establishing a data governance framework is critical to successful data-driven decision-making, the panelists cautioned that data governance work is never done. Regulations are constantly evolving, so data governance needs to evolve with those changes. But empowering employees through data governance is also an evolution, and it can always be improved.

Following the framework should be kept easy and it should use an organization's existing infrastructure, according to Schein.

"The mission," Eckerson said, "is to create a culture of governance to create a culture of analytics."