Setting up a data governance organization -- and getting the executive buy-in necessary to keep it running -- means that sooner or later the time will come to demonstrate the business value of data governance to company leaders.
But putting dollar signs on things like improved data quality
The panelists -- who convened for an hour to discuss data governance best practices, roles and responsibilities -- said any justification for investing in data governance should focus on advancing business goals and enhancing the bottom line. It’s important, the panelists said, to zero in on the business problems that data governance can fix and be sure to show what those problems are costing the company.
Sometimes the justification for a data governance program is clear-cut: Poor data quality is leading to a disastrous online experience for customers. But other times the process of identifying the business problems and quantifying their cost to the organization can get murky, said panelist Michele Koch, the director of enterprise data management and the data governance office at Sallie Mae.
Data governance best practices: Using the ‘five whys’ approach
Business users often don’t understand precisely why something is a problem or how that problem translates to increased costs or reduced revenues, according to Koch. A good way to identify business problems and ultimately arrive at their cost is to start by using the “five whys” approach, she said.
That means asking business workers to describe their biggest pain points when it comes to data quality. For example, the problem could be an application or system failing to deliver reliable results. Then ask the business users why it is a problem for them. Next, ask why the organization should address the problem, and so on. Eventually, Koch said, you will arrive at a deeper issue that directly affects the bottom line. For example: Poor data quality is leading to a disastrous online shopping experience, and X amount of revenue is being lost each month.
Once the business issues are clearly defined, Koch said, the process of assigning an approximate dollar value to the problem gets easier. For example, if a salesperson spends three hours performing manual processes because of poor data quality -- and that salesperson usually sells something every 1.5 hours -- then the company has just failed to close two deals.
“It all comes down to increasing revenue or decreasing your costs,” Koch said.
Data governance best practices: Defining roles and responsibilities
Setting up a data governance organization means defining and assigning the many roles and responsibilities associated with data governance success, according to panelist David Plotkin, data governance director for AAA of Northern California, Nevada and Utah.
At AAA, this was accomplished by dividing up the data governance roles and responsibilities into
three major categories, including the overseeing body, the business element and the IT support
team, according to Plotkin.
The overseeing body at AAA is based in the business side of the company and charged with handling the day-to-day aspects of running the data governance program. For example, it makes sure the approved business glossary is readily available and that data quality standards are continually being written and enforced.
Business users understand the data they create better than anyone and should therefore be
largely responsible for ensuring proper data governance. That's why the second category of data
governance roles and responsibilities at AAA is all about the business.
"I tend to think of [the business category] as the part you can't get by without," Plotkin said. "You [need to] have business buy-in [and] business providing those resources."
The business level of the data governance organization consists of the data owners, or the data governors, as they are sometimes called. These are the people who make the call when questions about data policies and procedures arise. And then there are the data stewards, Plotkin said.
"These are the folks who are out in the business; they have their day jobs, but they are the people that their peers tend to turn to with questions," he explained. "They know the data."
Plotkin said the data stewardship level is where "the real work gets done" because stewards are responsible for reporting any problems to the overseeing body -- a task that often leads to new data quality policies and procedures.
The third level consists of technical personnel who can explain data quality issues that crop up, for example, as the result of extract, transform and load operations, Plotkin said. They can also diagnose issues that exist within individual systems or applications.
"I will also toss in that it can very well be true that your best allies in IT are the enterprise architects," Plotkin said. "They get what we do, in large part."
Panelist Kira Chuchom, who manages the data governance organization at networking giant Cisco Systems Inc., agreed that steering committees are a key component of the data governance hierarchy. At Cisco, she said, the steering committee is made up primarily of company executives and other key decision makers.
"We are also big believers in data stewardship,” Chuchom said. “All of our data stewards [come] from the business.”