Data Gurus Take Charge

Twenty years ago, Scott Shoup was a recent college grad working for the Federal Emergency Management Agency, helping earthquake-stricken Angelenos sign up for assistance after the devastating 1994 Northridge quake. Now, he’s helping lead the agency’s efforts to leverage data across the department to be more “survivor-centric.”

Niall Brennan, a self-professed “data geek,” crunched numbers for years at Beltway think tanks and the Congressional Budget Office. Now, he’s massaging massive amounts of data at the Centers for Medicare and Medicaid Services to help agency leaders make decisions.  

Shoup and Brennan are two federal officials with one of the newest—and still largely unfamiliar—titles on their business cards: chief data officer.

Over the past few years, a growing number of federal agencies have appointed a chief data officer, or CDO, to help them manage their data as an asset—releasing it to citizens and entrepreneurs alike and scrutinizing it internally to derive new insights and drive efficiencies.

In February, the White House jumped on the CDO bandwagon, naming DJ Patil—the Silicon Valley alumnus from LinkedIn, who’s widely credited with popularizing the term “data scientist”—the first governmentwide data chief.

Government Executive sat down with experts, researchers and five recently appointed CDOs to discuss their mission and whether there’s room on the federal organizational chart for yet another “chief.”

Agencies Get Hip to Data

In terms of setting policy governing access to federal data, the Obama administration has hardly been a slouch. Since the administration created the online repository Data.gov in 2009, agencies have released more than 120,000 data sets through the site, according to the White House.

In 2012, the Obama administration unveiled a digital government strategy, calling on agencies to make application programming interfaces the new default in sharing data with the public. And in 2013, President Obama issued an executive order requiring federal agencies to collect and publish information in open, machine-readable formats and for the first time inventory all of their data assets.

“I think agencies simply recognize what a vast treasure trove of data we have as a government,” says Brandon Pustejovsky, CDO for the U.S. Agency for International Development, which last year began requiring its aid missions around the world to start an extensive data collection effort.

A Role Largely TBD

But can agencies keep up with the acceleration as new data sources keep feeding the ever-bulkier big-data beast and as steps toward greater transparency heighten scrutiny of government spending and performance?

Enter the chief data officer.

A number of agencies have already carved out space on their leadership teams for a new data chief position. That trend has accelerated in recent months; no fewer than six newly minted CDOs have been appointed since July.

The working definition devised by federal IT officials describes a hybrid role. The CDO is “part data strategist and adviser, part steward for improving data quality, part evangelist for data sharing, part technologist, and part developer of new data products,” according to a position description posted on CIO.gov. 

Dan Morgan, who became the Transportation Department’s CDO last summer, has focused on strategies to figure out how to help employees better leverage data in-house. “Those are the people who are really asking different kinds of questions of their data and really trying to extract as much value as possible from it,” Morgan says. “So, being in front of the computer is so much more important than being behind the computer.”

At USAID, Pustejovsky helped finalize the agency’s open data policy that requires organizations doing development work—including scores of contractors and grant-based organizations—to systematically collect data generated by their work and submit it to headquarters. Teams in Pustejovsky’s office comb through the data, which is beamed in from some 80 missions around the globe, scrub it to remove any sensitive material and then publish it online.

The data runs the gamut from livestock demographics in Senegal to HIV prevention efforts in Zambia. “While some data sets might seem arcane to one person, to the next person it’s the desperately sought-after answer to a researcher’s key question,” he says.

Too Many Chiefs?

There is quiet grumbling across government, however, that the C-suite has already grown overcrowded with trendy new, fill-in-the-blank “chief” titles. The skeptics ask: Is the answer to the government’s data challenges really the elevation of a new role that could splinter off even more of the CIO’s authorities?

“The first issue organizations should ask themselves is if they’re really happy with the way they’re leveraging their data at the moment—then they should not do anything at all,” says Peter Aiken, the founder of the Data Blueprint consultancy, and a longtime advocate of a beefed-up data role. His book The Case for the Chief Data Officer (Morgan Kaufmann, 2013) argues that most organizations do not fall into that camp—and the solution is not to simply pile on yet another portfolio to the CIO’s to-do list.

“CIOs are not paying enough attention to data,” Aiken says. “But it’s also very appropriate to recognize that they are being asked to do a tremendous number of things.”

Another reason agency CIOs may not be cut out to lead their agency’s data agenda: their focus on program management. Yes, that can be a bad thing. For years, the IT shop was the wild, wild West, with slipping deadlines, ballooning costs and little accountability. Critics will, of course, point out that is still often the case in the federal government. A program management mentality seeks to impose order on that chaos. But the best attributes of that discipline can actually be hindrances when it comes to data management.

“Data cannot be developed in the same mind-set and with a project-driven mentality,” Aiken says. “The only possible outcome that produces is actually more small piles of data. And if that’s your goal, we’re doing a terrific job. But most people realize that more small piles of data is really a bad idea.”

CDOs recognize the key to decoding their role isn’t necessarily a new piece of hardware or software. When asked to describe their work—and the challenges associated with it—most cited some variation of the decidedly low-tech themes of people, process and culture.

Even as analytic capabilities get quicker, faster, smaller and ever more massive quantities of data easier to manipulate—or “munge,” in the parlance of data geeks—smart humans remain the “key secret sauce,” Brennan says.

A ‘Golden Age’

The most ambitious visions of what the government can accomplish are easily tempered by the “legal, bureaucratic and practical hurdles” to data sharing across agencies, and fears that some activities “could run afoul of privacy advocates worried about how the government tracks its citizens,” said a recent article on FiveThirtyEight. The Bureau of Labor Statistics, for example, reports monthly jobless figures using imperfect survey techniques, in part because more direct data sources—such as the government’s national database of new hires—are off limits to BLS statisticians.

Still, chief data officers have begun to notice a change. “I went from walking the halls trying to get people to care about data 10 years ago, to being just overwhelmed with opportunities now,” says Shoup, adding that government is finally poised to enter a “golden age” of data.

There’s been fierce competition for talent in recent years. A 2012 Harvard Business Review article co-authored by Patil trumpeted the role of data scientist as the “sexiest job of the 21st century” and kicked off a hiring blitz for data geeks.

Experts say this isn’t yet one of those stories in which the private sector has already lapped the feds when it comes to swooping up talent. “So, government can either play catch-up,” Aiken adds, “or government can lead.”