By Joseph Marks
August 1, 2012
In mid-2010, the already frenetic Todd Park was in overdrive. President Obama had just signed the Affordable Care Act into law, the most significant reform of the American health care system since Medicare. It was Park’s job to figure out how government could use technology to make the law’s implementation as smooth and fruitful as possible.
As the Health and Human Services Department’s first chief technology officer, Park had a broad mandate but not a lot of staff. The idea behind the new position, he says, was to act like a “lean startup” smack in the middle of a mammoth federal agency, pulling together tools and resources on the fly to meet a particular need.
In this case, one of the tools Park decided he needed was Ed Mullen, a freelance Web designer in Jersey City, N.J.
Four months earlier, Mullen had been listening to talk radio and stewing over the controversial health care legislation making its tortuous way through Congress. The big debate was whether the law would include the so-called public option, a government-sponsored insurance plan. Mullen was preoccupied with another issue: the president’s idea of state-regulated insurance exchanges.
He was concerned Democratic lawmakers didn’t understand how transformative the exchanges could be and might jettison them in a bid to save the public option. “I thought the exchanges weren’t getting enough attention partly because people just couldn’t envision this abstract thing,” he says. “As a designer, I can visualize these things a bit better than the average person.”
Mullen had no training in health care. His only experience with government was as a political junkie watching from the sidelines. He knew how to design websites, though, and this problem in part seemed to him to be about design.
“Professionally, these are the kinds of problems I work on,” he says. “I do a lot of transactional sites and a lot of complex content. If you strip away the politics and the health care aspect, this is the kind of design concept I work on.”
Within a day and a half, Mullen built a mock-up of what an insurance exchange might look like. “I just needed to get it out of my system more than anything,” he says. He published the mock-up on his blog and sent out a tweet to White House New Media Director Macon Phillips. No response. He figured that was it.
Four months later, though, after the Affordable Care Act was on the books, Mullen got a call. Park had pulled together a small team inside HHS to build a website called Healthcare.gov. It wouldn’t be an insurance exchange, which the law didn’t require until 2014, but it would include something similar: a tool to calculate a person’s best insurance options based on his or her location, age and other factors. Phillips had shown Mullen’s blog post to the team and they wanted his help.
“We called him out of the blue and said, ‘Are you willing to come join the Healthcare.gov team?’ ” Park says. “We said, ‘You’d need to be here in basically a week and we’d pay you much less than you’re currently being paid and you’d have to work like a barking rabid dog, but are you willing to do this?’ ”
Mullen said yes. He signed on as the team’s design lead in May 2010 and two months later Healthcare.gov was online.
Bringing people like Mullen into government for short-term stints has become a hallmark of the Obama administration’s push to narrow the efficiency gap between government and the private sector, and it’s come to define Park’s work in government, first at HHS and, since March, as the nation’s second federal chief technology officer.
Taking Up Residence
Private sector participants are a vital ingredient in what Parks calls the “lean startup mode in government.”
They bring experience in fast-paced, competitive environments that people inside government sometimes lack, he says. Depending on the project, they can be generalists to offset federal experts or experts to complement government generalists. The produce-or-perish world of private sector enterprise also can prime the pump of federal workers’ own creativity, he says. These outsiders are less used to the strictures of government work and are more likely to think of creative ways around them.
“There’s this idea of a beginner’s mind,” Mullen says. “If you use an acronym, I won’t know what you’re talking about. I felt like, more than anyone else, I was the target audience [for Healthcare.gov]. As a self-employed person, I’m one of the people who will be buying health insurance from the exchanges, and I have trouble evaluating what’s a good health plan and what’s not. I know that pain.”
Park made millions of dollars as a health care software entrepreneur before joining government in 2009, and he speaks with the mile-a-minute enthusiasm and splashy metaphors characteristic of the startup set. He sometimes refers to these short-term government tours as “mashing up amazing talent inside government with amazing talent outside government.” More often, though, he uses the gentler term “entrepreneurs in residence.”
The government has fielded three major pilots of the entrepreneur in residence, or EIR, system so far. The first paired a team of Food and Drug Administration medical device examiners with manufacturers, doctors and academics, with the goal of finding a way to speed the approval process for life-saving stents, artificial joints and similar products. The second program paired experts from the venture capital and tech startup worlds with U.S. Citizenship and Immigration Services employees to devise a way the agency could allow more foreign-born entrepreneurs into the country.
In May, the White House launched the largest EIR program yet—Presidential Innovation Fellows. The program is modeled partly on Code for America, a nonprofit that pairs eager, young Web developers with cash-strapped cities to use technology to improve city services. Park, who seems always to search for the most explosive phrase he can get away with, describes the fellows as “the baddest of the badass innovators.”
The 20 fellows have six months to complete five projects each sponsored by a different agency. Projects include building a system to issue electronic payments rather than cash to people working as subcontractors on U.S.-led development projects in poor nations, as well as a tool to give U.S. citizens secure online access to their medical records.
The projects were chosen in what Park called a “cage match” with representatives from the interested agencies trying to sell the others on their ideas until only the five best remained. If all goes well, Park says he’d like to continue the program with new agency-generated projects each year.
Reforming the way Citizenship and Immigration Services handles entrepreneurs’ visas seemed like a natural fit
for the EIR approach, USCIS Director Alejandro Mayorkas says.
More than three-fourths of patents awarded to the top 10 patent-producing universities listed at least one foreign-born inventor in 2011, the Partnership for a New American Economy, a nonprofit co-founded by New York Mayor Michael Bloomberg, reported in June. Most of those patents were in science, technology and engineering. Yet the visa adjudication system is geared toward traditional jobs at established companies, not bootstrap entrepreneurs with one great idea and some financial backing.
Mayorkas gave the 10 agency team members and five entrepreneurs in residence from the tech startup and venture capital worlds three months to sort through training materials and policy guidance to come up with a solution.
“When we started off, it was almost like two different groups speaking completely different languages and only half understanding each other,” says Paul Ford, one of the EIRs and vice president for community development at Dallas-based SoftLayer Technologies, a cloud computing provider.
“We’d talk incubators and capitalization tables and the internal team would be using tons of government acronyms,” he says. “We’d look at them like, ‘What the hell are you talking about?’ It took a couple of weeks before we all started to understand each other. But once that clicked and we started working through all the different visa classes, those aha moments just started happening over and over again.”
One of the first problems the team encountered was the issue of control, according to Ford. Most work visas require sponsorship by an employer who will “control the . . . worker’s daily duties,” the statute says. That’s a tricky proposition for an entrepreneur who aims to run her own company.
The resident entrepreneurs proposed the idea of treating venture capital boards as the equivalent of bosses, which fits the language of the statute if not its original intent. The control a VC board wields over an entrepreneur isn’t much different from the power an employer exercises over a particularly independent employee, Ford says, especially if it has majority voting rights. There’s also plenty of paperwork to document that control—a critical part of the adjudication process.
The EIRs also lobbied for a more liberal reading of the requirements for the O-1 visa, reserved for applicants who possess “extraordinary ability” and are “one of the small percentage who has risen to the very top of [a] field of endeavor,” the statute notes. Entrepreneurs are often at an early stage in their careers, but may show extraordinary promise, Ford says.
“Have they flipped a company in the past? Created a startup? Been written up in Wired or The Wall Street Journal? That’s pretty darn extraordinary,” he says. “A lot of times it’s about reading between the lines.”
The EIR team was scheduled to wrap up its work at the end of July with recommendations about how to revise USCIS training materials and policy documents. Besides adding private sector perspectives to the visa adjudication process, the program also brought together field adjudicators and policy-makers, which sparked new ideas and forged professional connections, Mayorkas says. He hopes to launch another program focused on reforming visas for the entertainment industry, he says, but there’s no time frame yet for when that will happen.
The entrepreneurs in residence system, for Park, is part of a larger enterprise: making government function like a “lean startup,” with projects being managed by ad hoc and energized tactical teams from throughout the agency—and occasional outsiders—rather than by the existing bureaucracy.
These teams should include representatives from all phases of a project’s development to be sure nothing is missed along the way, he says, and the team should start building something right away —even the smallest component— so that members can spot bad presumptions before it’s too late. In the software world this is known as agile development. Engineers contrast it with waterfall development, in which each project phase follows in succession.
The lean startup model has been endorsed by a number of government watchers as a way to speed up notoriously
slow agency operations and to benefit from private sector advances.
“I think the starting proposition is we don’t have the same resources we used to and the pace of challenges has increased substantially,” says Max Stier, president and chief executive officer of the nonprofit Partnership for Public Service. “So we have to be innovative in government and adopt the best practices of the private sector.”
The EIR programs’ chief critics, in fact, tend to fault Park for transforming the government building process too little rather than too much.
Kathleen Allen teaches courses in entrepreneurship at the University of Southern California and directs the school’s Marshall Center for Technology Commercialization. She spent three years as an EIR at a major government aerospace contractor leading its first foray into commercial technology.
Making a careful, risk-averse organization that was used to operating like government run like a small startup required much more than simply overseeing a few product launches, Allen says. It meant restructuring how the entire commercial division was organized, how employees worked together and how they were evaluated. It also meant changing the office’s overall philosophy.
Assigning EIRs to narrow projects may yield some results, she says, but if those outsiders can’t look beyond their projects to reform agency culture more broadly, then their work is unlikely to create a paradigm shift.
Allen’s model of an entrepreneur in residence as a sort of super-consultant with broad authority to shift office structures and operations is common in the private sector. Park often references this definition by describing his own work in government as an EIR role.
He’s defined EIR positions more narrowly at the agency or innovation fellow level, he says, to ensure there’s buy-in from top executives in advance and to make sure projects don’t lose focus. Consistent with his lean startup philosophy, Park says, he also wants to build innovation in government from the ground up, making sure kinks are worked out along the way, and solutions aren’t imposed from the top down.
Edward Eitches is president of the National Council of Housing and Urban Development Department locals of the National Association of Government Employees. He supports entrepreneurs in residence programs but calls them a Band-Aid on the larger problem of bureaucratic inefficiency in government and siloed spheres of authority.
Eitches helped design a program called the Personnel Clearinghouse at HUD, which would allow employees to move laterally into open jobs for which they have special expertise or where there is more work than their current position offers. The idea is to facilitate something similar to Park’s lean startup teams, making it easier for the right person to find the right work that would both energize the employee and serve the institution.
The program was delayed by months of bureaucratic wrangling, he says, and ultimately was limited to HUD headquarters in Washington and only with a manager’s OK. As of early July, only three people had switched positions through the seven-month-old program.
Eitches made headlines in the government innovation community in February when he lashed out at Park’s predecessor, Aneesh Chopra, during a farewell event at the Center for American Progress. He argued Chopra’s intense focus on the developing EIR programs did little to promote federal employees’ own creativity.
“All these things work to inhibit creativity,” Eitches says. “You have to go through layers of people to get things done, and after all that, performance evaluations generally aren’t written with rewarding innovation in mind. I haven’t seen an [evaluation] element that says you’re an innovator. Maybe there should be one.”
The greatest barrier to the lean startup model in government—and where the private sector’s aid and insight might do the most good—experts say, is government’s low tolerance for failure.
Allen experienced this firsthand during her work with the aerospace contractor, which she asked not to name. In the aerospace field in particular, she says, any late-stage failure can mean losing billions of dollars if not losing lives. As a result, avoiding failure is valued much more highly than risk-taking and creativity.
Even when there’s less on the line, federal employees have developed a strong fear of failure, Stier says, because most program breakdowns are followed by harsh criticism from Congress, the media and the public. The lesson that agencies often take from such experiences is to spend more time ensuring they’re above reproach and less on figuring out new and better ways to do their jobs, he says.
Stier cites the General Services Administration’s Western Regions Conference in April, a management debacle with a price tag of more than $800,000, which included $100,000 for pre-conference planning, $5,000 for private parties and a $3,200 mind reader. The agency’s response to the scandal and ensuing scrutiny was to clamp down on all conferences and training, a move Stier worries will cost the agency much more in good ideas that never find the right audience and possible project teams that never meet.
In his public speeches, Park hammers on raising the government’s tolerance for failure and calls “failing fast and failing smart” an important lesson EIRs can bring to government. By using the lean startup and agile development models for government projects, he says, teams can fail early in the process when the stakes are lower rather than falling down on delivery. He concedes, though, that it will take a major culture change for the idea to take hold.
“If you fail after six months of strategy and a year or two of work and $100 million, that’s a really big problem,” he says. “If you fail after three days by five people, that’s much less of a problem. It might help you on the way to a solution. The ironic thing is people who do ‘waterfall’ call it risk mitigation. To me, that’s risky. Lean startup to me is the most conservative solution.”
By Joseph Marks
August 1, 2012