Getty Images

Experts outline keys for strong data governance framework

Developing guidelines that simultaneously protect organizations from risk while enabling and encouraging employees to work with data can be difficult, but not impossible.

The aim of a data governance framework is to strike a balance between mitigating risk and enabling the agility that leads to innovation.

As the volume of data that organizations collect increases exponentially, and as data becomes more important than ever to deal with fast-changing economic conditions, data governance is critical. And a data governance framework, when developed well, provides an organization with the structure for data-driven decision-making.

It's needed to manage risk, which could be anything from the use of low-quality data that leads to a bad decision to potentially running afoul of regulatory restrictions. And it's also needed to foster informed decisions that lead to growth.

But setting limits on which employees can use what data, while further limiting how certain employees can use data depending on their roles, and simultaneously encouraging those same employees to explore and innovate with data are seemingly opposing principles.

So a good data governance framework finds an equilibrium between risk management and enablement, according to Sean Hewitt, president and CEO of Succeed Data Governance Services, who spoke during a virtual event on April 26 hosted by Eckerson Group on data governance.

A good data governance framework instills confidence in employees that whatever data exploration and decision-making they do in their roles, they're doing so with proper governance guardrails in place so they're exploring and making decisions safely and securely and won't hurt their organization.

"The aim is to create a balance between governance controls and then agility and innovation," Hewitt said. "You need to use enough governance to mitigate those risks and [at the same time] optimize the value you're getting from the data."

Striking that balance, however, is not simple.

But there is a process organizations can go through to develop a successful data governance framework, according to Hewitt.

Sean Hewitt (middle right) and Josh Reid (bottom) of Succeed Data Governance Services take part in a virtual conference
Sean Hewitt (middle right) and Josh Reid (bottom) of Succeed Data Governance Services take part in a virtual conference on data governance hosted by Eckerson Group. Also pictured are Wayne Eckerson of Eckerson Group, Gaurav Pathak of Informatica, Prukalpa Sankar of Atlan, Ashwin Nayak of Zaloni and Matthew Helmrath of OneTrust.

Implementing data governance

Before ever assigning data stewards and setting limits on who can access what data, organizations need to understand exactly what data they have.

Without knowing metadata -- data about data -- enterprises risk being disorganized.

"Metadata is the fuel for data governance, so organizations need to build that intelligence," Hewitt said. "If they don't have the metadata available, they're flying blind."

Metadata is the fuel for data governance, so organizations need to build that intelligence. If they don't have the metadata available, they're flying blind.
Sean HewittPresident and CEO, Succeed Data Governance Services

Metadata enables organizations to know such things as where their data originated -- such as a point-of-sale application -- whether it was manually or automatically input, how many people have viewed and used a data point or data set, whether data has been manipulated or altered and by whom, and the age of the data.

Metadata also enables organizations to define and catalog their data so it can easily be found when needed for decision-making.

"The starting point for data governance is metadata," said Josh Reid, partner and senior advisor at Succeed Data Governance Services. "Metadata really helps you build your data intelligence. All the things we document are metadata, and they are critical for a data governance program."

After gaining an understanding of the data at hand, organizations should assign responsibility for different data sets and data domains to a data steward or owner.

And what exactly those responsibilities are need to be explicitly laid out, according to Hewitt.

"Organizations need to be crystal clear because confusion is the enemy of data governance," he said.

Data stewards/owners are the ones likely empowered to set limits on who can use a particular set of data to mitigate risk, while simultaneously making sure employees can safely and securely access the data they need to fulfill their responsibilities.

They're also usually the ones entrusted with keeping data up to date and making sure the metadata is accurate so data can be easily found and properly used.

Once data stewards/owners are determined comes the development of a data governance plan -- the actual steps that lead to the balance between risk management and employee enablement. Developing and implementing processes and policies support the desired outcome of a strong data governance framework.

"Vision, mission and goals bring everyone together, creates common understanding and drives everyone in the same direction," Reid said. "Together, with a strategy and roadmap, the leadership team and the people doing the work have a clear understanding of what's going to be done, when it needs to be done and what success is going to look like."

Finally, organizations need to ensure everyone is set up in their roles to execute their data governance plan -- including education -- and they need to monitor and improve their plans over time.

That process of monitoring and improving includes setting up a set of metrics that can be looked at on a regular basis to monitor progress, confirm that goals are being achieved and make improvements to the data governance plan where needed.

"All those things need to be baked into a data governance framework in order for it to perform optimally," Hewitt said.

Overcoming barriers

Despite what seem like logical steps for implementing a strong data governance framework, many organizations struggle with data governance.

And at the heart is cultural mindset, according to Hewitt.

Too many organizations view developing a data governance framework as a project rather than a process. A project is an undertaking done once that has an end date, whereas a process is something that is ongoing.

"This is why a lot of organizations struggle," Hewitt said. "They don't accept that data governance is a cultural transformation."

The ones recognizing the need for data governance -- often those in IT departments -- therefore need to be adept at change management, he continued.

"They need to understand the mindset, history and traditions of their organization," Hewitt said. "They need to understand the power relationships so they can use those power relationships to help overcome some of the roadblocks."

And they need to show early success, Reid added.

That can cultivate executive support. Then, as momentum increases, those benefiting from the data governance framework -- the end users enabled to work with data with proper guardrails -- can gain confidence. Finally, after success with pilot programs in selected areas, data governance can spread across departments until it's part of the organization's culture.

It's an approach that begins from the top down, then goes from the bottom up, and finally from the middle out, according to Reid.

"It shows that everyone in the organization has a role in data governance," he said.

Dig Deeper on Business intelligence management

Data Management
Content Management