While agencies are generating a lot more data than they used to, the promise of "big data" remains elusive for many organizations because the information is either insufficient, unreliable or managers don't know how to use it -- or all three.
"Our No. 1 objective is to make sure data is valid," said Angela Bailey, chief operating officer at the Office of Personnel Management, speaking at Government Executive's Excellence in Government conference in Washington on Monday. But accuracy isn't enough if it doesn't tell you what you need to know. For example, while OPM can say how many federal employees will be eligible to retire at a given time, it's much more difficult to predict how many are actually likely to retire at that point.
To understand if someone is likely to retire you'd need to know much more than their age or years of federal service, Bailey said. Are they married? Do they have alimony obligations or young children or children in college? Do they have a large mortgage? Are they in good health? Are their parents alive and financially secure or in good health? That's the kind of information people factor into retirement planning. Without it, you can't make informed projections about individuals' intentions.
Steve Goodrich, vice chairman of the Government Transformation Initiative, said agencies are much more adept at collecting and using data than in the past, but they still have a long way to go.
In particular, agencies need to focus on developing and promoting data-savvy managers and executives. "Education and training programs need to adapt" to new expectations and requirements of leaders, Goodrich said.
The Digital Accountability and Transparency Act, which President Obama signed into law on Friday, could increase exponentially the amount of data agencies generate and will have access to. It requires them to standardize the way they track and report procurement information with the goal of improving insight into federal spending.
The DATA Act is expected to have a huge impact governmentwide, but especially at the Defense Department, which completes about 9,000 procurement transactions every month, Goodrich said.
As data becomes standardized it will become much more useful, he said.