Data-Driven Government? Not Yet. Feds Still Manage by Hunch
No, agencies don't really a "big data" button. Yet. Roobcio/Shutterstock.com
In a true data-driven culture, leaders need to be willing to rethink their plans and perhaps even change course based on what the data says, according to EPA CTO Greg Godbout.
If you peruse the federal budget, you’ll find more than dozen references to “data-driven” reviews and programs. No fewer than six agencies have named chief data officers since last summer, and the White House appointed the first governmentwide chief data scientist earlier this year.
Experts say parsing through ever-larger quantities of data can help agencies derive new insights -- even lead to creative business breakthroughs -- and operate more efficiently.
But data-driven decision-making is still just an idea in many quarters of the federal government.
“Government sort of manages on hunches,” said Greg Godbout, chief technology officer at the Environmental Protection Agency, at AFCEA Bethesda’s Data Symposium June 30.
Godbout is the co-founder and former executive director of the General Services Administration’s in-house digital startup, 18F.
"It's great to have a hunch,” Godbout said. “Hunches are important. But we need to then have evidence to back up those hunches . . . We need to start having real metrics to do this."
In a true data-driven culture, leaders need to be willing to rethink their plans and perhaps even change course based on what the data says, which requires a “massive culture change,” Godbout said.
"I wonder, is the federal government ready for leadership that's willing to say, upon seeing early signs that things are going wrong, 'OK, I suck. This was a bad initiative. Let's go in a different direction'? Really, that's what we're talking about."
Agencies under President Barack Obama have made overtures to becoming more data driven. The IT Dashboard, for example, tracks the spending and performance of agencies’ major IT projects on a public website.
But even in data-driven programs, “we follow metrics that are silly,” or that don’t tell the whole story, Godbout said.
Case in point: The day HealthCare.gov launched in 2013 -- and promptly crashed -- it still maintained a “green light” on the dashboard, meaning the agency’s tech officials considered the project low risk.
"That dashboard, to me, loses credibility,” Godbout said. “It's not working. We need to dig in and figure this out."
Godbout said the government needs its own pipeline of data scientists and other data-savvy talent.
Patil, the Obama administration’s new governmentwide data guru, co-authored a 2012 article in the Harvard Business Review, which hyped the role of data scientist as the “sexiest job of the 21st century” and kicked off a hiring blitz for data geeks across the private sector.
"We need to get an army of them coming to government,” Godbout said. “And I just happen to know there is an army coming. They're hard to hire; they're very hard to find. They're out there and they're very interested in working for government."
The 2014 Digital Accountability and Transparency Act, approved by Congress last spring, could also nudge agency leaders toward data-driven decision-making.
The law requires agencies to report all financial information using common standards so it can be easily tracked.
In the past, though, simply collecting data has rarely led to management breakthroughs.
Since 1990, agencies have been required to collect reams of financial information for auditing purpose. But despite spending countless hours putting together reports on agency spending, that data is rarely used, said Karen Lee, chief of the federal financial systems branch at the Office of Management and Budget.
"We certainly have not been able to really crack the nut on how to use this federal spending data to improve our management,” Lee said.
Even now, “baby steps in using data are major steps," Lee said, noting that some federal offices are still transitioning from hand-written spreadsheets to Excel.
"We have too much data, to be frank,” she added. “And our challenge is how do we package data in ways in which federal managers, recipients of federal awards, those who are interested in helping the government manage better and target resources to those who need them can actually make better decisions.”
(Image via Roobcio/Shutterstock.com)