The art and science of data

There is both and art and a science to improving agency management through the use of performance data, said John Kamensky, senior fellow with the IBM Center for The Business of Government.

That means federal managers have to be more creative about the data, said Kamensky and other management experts.

The data is not the key thing. Rather, the improvements that the data can enable are what matters. "A patient cares about a doctor fixing things, not about how the doctor got his degree," said Jon Desenberg, policy director at The Performance Institute, a private, non-partisan think tank that looks to improve public- and private-sector performance. "Data is a scoreboard. In a game, you look at the scoreboard, but watch the players."

Kamensky, Desenberg, management process expert Robert Behn and others spoke in panel discussions and keynote presentations at the Performance Institute's Government Performance Summit on June 11 in Arlington, Va., covering a wide range of meta-management subjects. They touched on PerformanceStat assessment programs in federal agencies and the wave of big data emanating from those agencies.

Behn, who has worked with the governor of Massachusetts on government performance issues, and is faculty chair of the Kennedy School of Government's executive program at Harvard, noted that adhering to PerformanceStat assessments that include annual goals and other milestones can be tricky in tight economic times. "The targets can be negotiable," he said. "Resources change, targets change. All targets are negotiable."

Understanding that you can't do everything with limited resources is helpful in goal-setting, he said. He advised summit attendees to "accept we can't do all we wanted," but excel in what they can accomplish. Setting goals that are lower to ensure more areas are covered isn't the best path.

Kamensky advised looking deeper into internal data to glean novel solutions that may not be immediately apparent. The City of Baltimore Education Department's effort to increase graduation rates by searching through truancy data, ultimately chasing down chronic student illnesses in particular neighborhoods and working to improve health conditions, is one good example, he said.

Big data produced by agencies for external consumption by the public, press or other agencies, said Michael Nelson, senior technology and telecommunications analyst at Bloomberg Government, "needs a good bumper sticker" to indicate exactly what it is. Without a stated identity, masses of data generated by federal agencies for their customers will not hit home, he said, rendering it underused or largely useless.

Nelson, who worked at the White House as a science advisor, as a technology analyst at the Federal Communications Commission and later as part of President Obama's campaign team, pointed to the National Oceanographic and Atmospheric Administration, the Department of Transportation, and the Environmental Protection Agency as stars in presenting big data on the federal stage.

NOAA, he said, has eased the use of complex, huge volumes of weather data for commercial weather forecast companies, while DOT has a vast user community that relies on its safety information. Both agencies have created data with users in mind. The EPA, meanwhile, has set understandable standards for data generated by its employees for internal as well as external consumption. "They tell employees to make spreadsheets like everyone in the world is going to see them."

Nelson advised federal agencies to "know their audiences" for data. There are separate ways to package and provide it, depending on whether it's destined for procurement, budget, legislative, regulation or the media.

Knowing the nine steps in handling big data also helps, he said. (See box at right.) Applying the data and protecting it, he said, can be the most challenging steps.

Nelson pointed to the recent release of information on NSA's PRISM program as an example of how even the best protections sometimes can rest on unpredictable ground. An NSA contractor who apparently had wide access to that agency's classified intelligence gathering apparatus and disclosed its existence to the media.