XChange Public Sector: Why Large Federal Programs Fail

Large federal programs tend to fail in predictable ways, at specific, identifiable points in the process. Similarly, successful programs have elements in common that can be replicated.

Those are the fundamental findings of William Eggers, global director of Deloitte Services Research Public Sector, borne out by research he conducted for a new book, "If We Can Get a Man on the Moon: Getting Big Things Done in Government."

At the opening general session of XChange Public Sector, Eggers recounted how he and his co-author, John O'Leary, explored why some big government initiatives succeed while others fail.

"We studied more than 100 projects since World War II -- the War on Poverty, some real wars, dozens of big IT initiatives, both at the federal and state level," Eggers said. "We think it was the biggest research project ever conducted on public initiatives."

What they found were systemic barriers to success, built into the processes that have developed for taking on challenging programs. Eggers said there is a map that can be followed to overcome those barriers.

"There can be a good idea, with an implementable design that works in the real world, that achieves democratic commitment," he said. "But it has to go through the wormhole, from one world to another. We call it that because one side, that's the political world, the policy people. On the other side is the bureaucracy, the executive branch, and there's this chasm in between. In the private sector that chasm doesn't exist."

On either side, there are traps and pitfalls that emerge consistently, he said, such as the "Tolstoy Syndrome" -- looking only at evidence that confirms one's view of the world, a failing that is particularly relevant in the government world.

One way to overcome this syndrome is through crowd-sourcing, an idea NASA has put into use, Eggers said.

"NASA has always said there are more smart people outside an organization than inside, so how do we access them?" he said. "So NASA is posting on Innocentive.com, which connects people with very difficult technological problems with smart people" interested in contributing ideas.

Another technique to overcome the Tolstoy traps is using analytics to provide hard evidence to justify decisions, Eggers said. "Analytics is going to be absolutely taking off in the next few years; spending will go through the roof," he said.

The design stage of any program is another very important potential pitfall. Eggers and O'Leary looked at a lot of initiatives that appeared to be implementation failures, but they actually turned out to be design failures. They asked members of the government's Senior Executive Service about the problem, and the executives were scathing in their comments about how government does designs projects so badly; only 16 percent of those running the biggest programs said the government designs policies that can be implemented. Eggers termed this the "design-free design trap."

"Policy design is dictated from the top down, for ideological or partisan reasons," Eggers said.

NEXT: Key Is To 'Design For Execution

'But it is not just the legislative part of the process that's broken, it also is in the actual design. They looked at 22 major acquisition systems since World War II. During the early Cold War period, it took about four years for the project to be completed; in the late Cold War years, it was up to six years. Now it's up to 13 years, and by the time it's completed, the technology is obsolete.

Eggers suggested the way to overcome the design-free design trap is "design for execution." That is, fail more, but fail better, by doing more small bets, modularizing them, and not swinging for the fences. He pointed to Google as one example; it was in beta testing for just one day. Twitter was in beta for just two weeks.

"Put them out there and get them into the real world quickly," he said.

He offered three recent examples that demonstrate this is an approach that will work for the government:

Unmanned Aerial Vehicles, which used to be considered high-tech toys, but this year the U.S. Air Force will train more joystick pilots than new bomber and fighter pilots combined. A $4 million Predator UAV is doing the same job as a $60 million manned fighter jet.

In 1978, RFID technology was used for the first time -- to monitor cows. Now it's being used in place of incarceration for over 100,000 people facing time in minimum security prisons, at a cost of $5 a day, rather than $77/day for prison.

Finally, the emergence of mobile phones as supercomputers are remaking a host of fields, Eggers said.

"We're doing a lot of work on how to transform health care through mobile applications," he said. "We can turn a smartphone into a stethoscope or a heart monitor, or even a hearing aid, with the right applications."