"Clearly, increasing trust and accountability is an important goal of transparency, but it's also important to think about how dashboards have been successful in the private sector and apply that to the public sector," he said. "The goal in the private sector is to take that information and then use predictive analytics to think about how to make decisions on the future. I think it's still an open question how the administration and state and local governments will end up utilizing this in that way."
There's a debate underway about just how much government data to release and how to do it. On one side are people like Clay Johnson, director of open government group Sunlight Labs, who expresses concern that dashboards tend to over-editorialize data. While it's critical to release "big, ambitious metrics," he would rather be able to get raw data and let citizens make their own dashboards than look at colorful pie charts.
Government data is most useful when it maintains its "fidelity," said Oracle's Doolan. "What's really important is the degree of separation of the data from the operational systems," Doolan said. "If you have too many degrees, the data is stale, out of date or doesn't have enough relation to what's actually happening in the enterprise. The closer it is, the less reinterpretation of data you have."
At the same time, there's trepidation about publicly releasing government performance data. Government managers worry they'll be bombarded with calls from citizens concerned about small expenditures.
Data quality, or lack of if, is another issue. Releasing inaccurate data opens agencies to public criticism, as the feds found when the first round of stimulus funding data began pouring in to Recovery.gov. Such scrutiny ultimately leads to improvements in data quality better, Kundra has pointed out. Indeed, problems with data quality helped spur the creation of a federal spending data quality working group, made up largely of agency CFOs, which will work to ensure the government has adequate quality controls in place.
Treworgy said data quality is more important than quantity. "Just putting out reams of data isn't really helpful," he said, pointing to the fact that even IBM's CEO has a dashboard of only 10 metrics or so.
Because they serve much larger constituencies, however, government agencies will be expected to provide high-quality data and more of it. As a result, their new Web dashboards will have to be good at both.
InformationWeek's Government IT Innovators program spotlights innovative government IT organizations--at the federal, state, and local levels. Nominate your agency by submitting an essay on your most innovative IT initiative completed within the past 12 months. The deadline is April 30. Click here to take part.