If an organization has data flowing across multiple systems and processes it's time for a formal data governance...
By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
program, according to data quality experts like Gwen Thomas, president of the Data Governance Institute.
"Do these organizations want rules to specify how the data should be structured, shared, accessed and used? Of course," Thomas said. "Do they want controls to enforce those rules? Of course. Do they want clear rules of engagement that show how stakeholders make decisions about these rules and controls? Of course. Now, in 2010, it's a given that organizations of a certain size want and need some form of formal governance."
Joseph Bugajski, research vice president at Stamford, Conn.-based Gartner Inc., goes even further. "In some cases, it is past time," said Bugajski, who was previously chief data officer at Visa Inc. "Almost every business has reporting requirements that depend upon reliable data at its source. To assert that the data is reliable requires governance. And in the simplest form, it means somebody owns responsibility for saying, 'Yea, verily, this data is accurate.'"
As defined by the Data Governance Institute, data governance is "a system of decision rights and accountabilities for information-related processes, executed according to agreed-upon models which describe who can take what actions with what information, and when, under what circumstances, and using what methods."
Launching such a system, however, can be a hard sell. First, many organizations of a certain size believe they do have data governance, Thomas said. These data practices typically don't have breadth and depth or alignment of a formal governance program, instead allowing individuals or business units to make their own rules and standards.
"If you take the approach of the sheriff who rides into town and says, 'Now we're going to have governance,' everyone sitting around the kickoff table will be looking at each other because they feel they have governance in place," Thomas said. "You've already alienated them."
Data governance: A political hot potato
Data governance programs are politically fraught, Thomas said. Managerial people tend to see governance programs as another bureaucratic layer that will slow them down, rather than make their work easier. People who work with corporate governance all the time understand the kind of ill will and havoc a poorly defined or unfocused governance program can inflict. And the C-suite will simply be horrified that the company didn't have a formal data governance program in place already.
Then there is the nature of data itself in the modern workplace, added Bugajski.
"Data is everywhere, like air. If the air around you is dirty, unless it is causing you sickness, you don't do anything about it," Bugajski said. "Everybody has tons and tons of electronic data, and very few of us know its quality and reliability."
Data is everywhere, like air. If the air around you is dirty, unless it is causing you sickness, you don't do anything about it.
Joseph Bugajski, research vice president, Gartner Inc.
In fact, CIOs and business leaders looking to start a formal data governance program should probably steer clear of using the term data governance until they have a clear idea of what their data governance aims to accomplish, Thomas advised last week at the MIT Information Quality Industry Symposium.
"Every organization has to find the words that will resonate with the organization. How you communicate your program is almost as important as what you do," she said.
It's important to "focus and frame" the problem that your data governance program will address, Thomas said. Successful data governance programs start with a driving need. They address problems that keep an organization from doing its business: a broken IT development life cycle, a data warehouse that needs updating, a business process with data flows that aren't right. Thomas recently worked with a troubled bank that was told it needed to institute a data governance program before it could get capital from its investors.
The instigating problem, however, should not define the data governance program. A data governance program is not a project, which implies a finish date, but an ongoing effort.
For John Bottega, a keynote speaker at the MIT symposium, making the case for a formal data governance program wasn't an issue. Bottega is chief data officer for the markets division at the Federal Reserve Bank of New York. Appointed in February 2009, his official resumé states that he is responsible for driving and implementing the division's data management strategy. That office "encompasses business, governance and technology in order to establish a sustainable business data discipline and technology infrastructure."
In addition, Bottega has the support of board members in high places. In February, Daniel K. Tarullo, a member of the Federal Reserve System's board of governors, testified about systemic risk before various Senate committees. "The recent financial crisis revealed important gaps in data collection," Tarullo said, adding that "greater standardization of data is required to decrease the chances" of a crisis of this magnitude from happening again.
Indeed, if the financial crisis at the heart of the global recession showed anything, Bottega said, it was that data quality in the financial services industry was poor.
"We had data, but it was not comparable. It did not satisfy requirements. It was not collected and captured at the source in a methodology and format that would enable analysts to effectively utilize this critical financial data to perform their analysis," Bottega explained.
Getting started with a data governance program
According to Thomas, five preconditions must be in place before data governance can be implemented.
- A driving reason for data governance, supported by high-level sponsors.
- Political power to overturn dissenters.
- Firm time commitment from the participants.
- Trusted project managers who can document and communicate with ease.
- Data analysts who really know the ins and outs of your data system.
Once these conditions are in place, data governance programs often include the following:
A support team, or working group, to do the actual scoping of the data governance program, sometimes led by a data architect and concentrated in a data governance office.
It's imperative that the data management person in charge can lead people. That means being able to express an idea in three or four different ways, if necessary, to get agreement. The data manager has to figure out how the data governance program will align with the organization's governance efforts. That likely will require assuring people, maybe even in writing, that the data governance program does not aim to take power from other governance groups.
A council of decision makers. Data governance raises hackles. A council of decision makers is necessary for brokering the inevitable compromises that are part and parcel of data governance programs. The council should include the people whose systems and processes will be affected by the data governance program and are empowered to adjudicate the required changes. Successful programs "are very careful about scoping what the data governance program will address, so that the council's time will not be wasted," Thomas said. As the scope of the data governance program increases, more people may have to be brought into the council.
Embedded workers from the business and IT to enforce the rules. It is critical that these experts understand what the data governance program aims to accomplish, as they will "be making judgment calls about the data every day," Thomas said. This group can also function as the eyes and ears of the working group, to help identify problems and provide "a feedback loop."
Let us know what you think about the story; email Linda Tucci, Senior News Writer.