IBM users reveal five data governance best practices to remember

IBM customers Nationwide Insurance and Cardinal Health reveal five rules of thumb they learned while implementing highly ambitious data governance programs.

Organizations launching an enterprise-wide data governance initiative must prepare for setbacks and be willing to employ a mashup of different techniques to succeed, according to IBM customers Nationwide Mutual Insurance Company and Cardinal Health.

Representatives from Nationwide Insurance and Cardinal Health -- two companies that have launched data governance programs on an enterprise scale -- spoke to a crowded room of attendees an IBM conference late last fall to share their experiences and explain five key data governance rules of thumb they learned along the way.

Five data governance rules to remember

1. Keep it realistic -- especially in the beginning
2. Use top-down and bottom-up approaches
3. Enforce data governance rules while incentivizing compliance
4. Build a strong partnership between business and IT
5. Begin business glossary work early on

Keep it realistic -- especially in the beginning
Getting various departments within an organization on board with a data governance initiative takes time and patience, according to Frank Sheridan, associate vice president of marketing and information management at the Columbus, Ohio-based Nationwide.

Organizations should begin the process by building a “coalition of the willing” -- representatives from each department who recognize the value of data governance and want to help, Sheridan said.

But as those representatives begin meeting to discuss and agree upon definitions of business terms, there are bound to be setbacks. Different departments have their own way of doing things and change rarely comes easily.

Sheridan said the best way to deal with setbacks and disagreements -- particularly in the beginning -- is to take them in stride and continue moving gradually forward until agreements can be reached.

“You need to be a little bit more flexible in terms of being perfect every time we have [those] interactions and share that information,” Sheridan said. 

Use top-down and bottom-up approaches
A question often asked at information management conferences centers on whether data governance initiatives should begin from the bottom rungs of an organization as a grass roots effort, or whether they should stem from a “top-down” or executive-level mandate.  

The answer is that both approaches should be used where they make the most sense, according to Sheridan and Ted Friedman, a data management analyst with Stamford, Conn.-based IT research firm Gartner Inc. 

If the data governance rule or metric that an organization is trying to manage is important across multiple departments, then it should be mandated from the top down. In that case, it’s important to gain executive support early on. But when the metric applies primarily to one department, then a bottom-up approach to data governance may suffice.

“A lot of our clients try to do this comprehensive enterprise-wide data governance from the top down and they get stuck,” Friedman said. “Individual business units have the need to localize a little bit and have unique requirements, so that’s why finding a center of gravity and working a bit bottom up at times is also effective.”

For more on data governance best practices

Find out how Sallie Mae adheres to data governance best practices

Gain insight on Data governance best practices from governance pros

Find out how shipping firm A.P. Moller Maersk made enterprise data governance a reality

Enforce data governance rules while incentivizing compliance
Organizations should enforce data governance rules consistently across the organization. But as new data governance rules, policies and procedures take shape, organizations shouldn’t expect that mandate alone will get people to comply.

Organizations need to strike a delicate balance between enforcing the rules consistently and providing incentives for those who willingly comply.

“You can’t just force everybody to comply, nor can you just rely on good will and incentive and just hope that people get the job done,” Friedman explained. “You have to have a bit of both because one or the other alone isn’t going to work.”

Begin business glossary work early on
A senior project manager for information governance at Dublin, Ohio-based Cardinal Health, Brooks Zaremski is currently charged with building and maintaining the company’s master business glossary. He recommends that anyone considering a data governance program begin work on the business glossary as early as possible.

A business glossary contains all the mutually agreed-upon definitions of business terms and uses metadata to expose those definitions to the entire enterprise. Zaremski said business glossaries are typically built out gradually.

Friedman said organizations increasingly see the task of building business glossaries as an essential early-stage data governance activity.

“It’s all about getting people on that same page with the terminology,” Friedman said. “If you don’t do that early on, the whole effort is that much harder.”

Use common tools whenever possible
Organizations use many different technologies to ensure data quality and catalog and enforce data governance policies and procedures -- so many that it may be difficult to standardize on one or just a few.

But organizations should try to use common information governance and data quality tool sets wherever possible, because this will encourage the consistent use of business terminology across systems and departments, according to Zaremski.

“You have to have a common business language,” he said. “If you’re not speaking the same language, [data governance is] pretty challenging, if not impossible.”

Mark Brunelli is the Senior News Editor for Follow him on Twitter @Brunola88.

Read more on Business intelligence and analytics

Data Center
Data Management