Innovative State

by Aneesh Chopra

  That experience lit a flame for Tejeda, Reavis-Bey, and Pandel that they weren’t prepared to blow out when their fellowships ended. They sought to make the progress sustainable. “It shouldn’t happen that a project goes well, and people like it, and then it just dies because there’s no one there to support it,” Tejeda said. “This felt too important. It was New Orleans only, but hopefully we can expand it out to different cities that have similar needs. This is not an area where there are a lot of people who want to build tools for communities, and this is an important space, where you can have a real impact in fixing cities and helping people be involved in their government. If we can continue to work in this space, and people see the potential in this product—I wanted to make sure that we could.”

  So they passed over opportunities with established technology companies, staying in New Orleans and signing up for Pahlka’s incubator program, which allowed some teams to continue their work while benefiting from the access to advisers and structure already in place. They came into this phase with contacts, confidence, and ambition, perfecting their pitch that, as Pandel put it, “there’s space for innovative software with a user-centered approach, using the lean startup method, that can actually better meet the needs of cities and cost them less,” and that it made more sense for a city to hire them to improve its communication with citizens than pay a major corporation to attempt to replicate their work.

  During an interview in late 2012, Pandel related that “when we’ve been expressing the (relatively low) cost of this, people have been kind of surprised; they don’t know what to make of us. ‘Wait a second, this is a software system that provides value to me, and I don’t have to go through procurement for it. I don’t understand.’ So we really think there’s room in the civic space for projects like this.”

  Tejeda spoke of the potential to be “trailblazers.”

  “We get to be the first to try and disrupt this space at a fraction of the cost. We can possibly build a relatively big, successful company, and what we’re doing is displacing these big, gigantic dinosaurs that are just not providing the kind of value that they should to cities at the cost that they charge,” he said. “So I think we are all relatively optimistic as far as the space. There are legal challenges, as far as procurement, but this is what we are going to be working on now. But that’s the gamble of this opportunity, not whether the market is big enough.”

  The gamble would continue paying off in 2013, when the team began incorporating permitting and licensing data to its application, moved back to Northern California, changed its name from Blight-Status to Civic Insight, began working closely with the city of Palo Alto, entered the $3.2 million Knight News Challenge on Open Gov along with 800 others, and was named one of eight winners.38 That came with a $220,000 Knight Foundation investment grant.

  “So we have legs for a little while longer,” Tejeda said during the summer of 2013. “We can talk to other cities.”

  Code for America, in concert with Civic Insight, is smoothing the path for adoption by other cities, with the former creating a Code Enforcement Housing Data standard and the latter working to create an API that would connect to it.

  “There’s a lot of momentum,” Pandel added.

  “They get to go from cogs in the wheel to creating something long term that they are a founder of,” Pahlka said. “It’s not just doing good, it’s a narrative of success.”

  It’s an open innovation bonanza, featuring elements of open data, standards, and challenge competitions, all put to great use by entrepreneurial talent that government encouraged to assist.

  Meanwhile, Pahlka continued writing her own narrative of success. Upon taking the one-year leave from Code for America to work as Todd Park’s deputy, she oversaw the expansion of the Presidential Innovation Fellows program in its second year, from 18 to 43 fellows, and from five to 10 projects, some of them second phases of first-year projects (such as Open Data; Blue Button; RFP-EZ; and MyUSA, the next iteration of MyGov) and some (in the areas of disaster response, VA modernization, and federal agency accounting) representing new technological spins on long-standing issues.39 She also brought in the U.S. General Services Administration (GSA) to assist in scaling prototype solutions to multiple agencies. Her intention, after institutionalizing lean startup and collaborative principles inside the federal government? To bring the lessons back to Code for America as new fuel for the virtuous cycle.

  So the narrative of open innovation will continue. It is a hopeful story of people coming together—through handoffs and ­handshakes —to create the “wiser, frugal government” that Thomas Jefferson envisioned: one that is more transparent, participatory, and collaborative; one that is defined not by its size but by its smarts.40 One that, believe it or not, can work.

  Postscript

  You may not have heard much about any of this. Not the gatherings of influential voices from different sides working in good faith to seek common ground, such as the December 2010 CEO summit at the Blair House. Not the actions taken in the aftermath of those gatherings to meaningfully address some of the nation’s structural and regulatory issues constraining economic growth, such as the National Wireless Initiative and the JOBS Act. Not all the progress made in open government over the course of President Obama’s tenure, with federal agencies experimenting in ways big and small: from the Department of Health and Human Services giving life to a new industry of private sector health information applications to the Department of Defense reaching beyond the usual Beltway bandits to a former Sizzler waiter for the design of a combat support vehicle.

  You may not have heard much about the Washington to which I bore witness, one that was so much more collaborative than is customarily portrayed, especially when the camera lights were off, and ideological groups weren’t raising a ruckus. Thus, you may be surprised that Eric Cantor, the Republican Majority Leader whose beliefs don’t typically align with the Obama administration, sent this tweet on January 27, 2012: “Just heard the news that CTO @AneeshChopra is stepping down; his work, both in Virginia and Federal, have helped advance open government.” It wasn’t just about me either. On November 18, 2013, long after I was gone, he tweeted: “Providing open, machine-readable access to exactly how your tax dollars are spent is a major win for #opengov.”

  You may not have heard about any of this progress even—or especially—if you are a voracious consumer of cable news media. Those outlets are conditioned to focus on now rather than next, strife rather than substance, conflict rather than cooperation, and most of all, caricatures rather than context. This is especially true when they confront voices and topics that don’t conform to the typical ideological constraints. Either they simply don’t know what to do, or they don’t care to try.

  This is what Eric Ries has experienced. Outside of the public eye, he has seen the excitement spread for the application of his lean startup principles in government, from the local (including Mayor’s Innovation Fellows in San Francisco and the Office of New Urban Mechanics in Boston) to federal levels, as evidenced by the Presidential Innovation Fellows. “They all have their unique spin on it, but the core principles are very much alike,” Ries said. He has also brought the message to his most avid followers, by inviting me and my successor, Todd Park, to address more than 10,000 people at his annual Lean Startup Conference in 2012, and explain how these methods can work in unexpected areas.1

  Yet Ries still struggles to find the right spin to make all of this meaningful and palatable for the national media machine. He has been shouted down and cut off on cable news shows, by hosts and guests who can only debate tired old dichotomies like taxes and spending, small and big, left and right. He believes the public would appreciate and embrace fresh approaches, if they only knew more about them. Instead, he characterizes the overall level of awareness at “zero.”

  “Zero,” Ries repeated. “It’s hard for the traditional political media to
report on, because they are so locked on to ‘he said, she said’ and polarized reporting, that this doesn’t fit the category. It’s pro-growth and pro-government. They’re like, ‘What!? How do you write that article?’ When I try to talk to the traditional political press, they don’t know what I’m talking about. Every story, they’ve already written a million times. They’re just changing the names.”

  At the time this book was going to press, the American public was becoming more aware of technology in government, but hardly in a flattering light, due to the extremely rocky launch of Healthcare.gov.2 The site was designed to allow citizens to shop for new health insurance plans, much like online marketplaces for airline tickets or shoes. The intent was for the American people to have the freedom to choose among any of the dozens of insurance and benefit options that are available in a given community—and to find more affordable prices for those plans, based on a family’s income.

  Although this was not widely reported, the Healthcare.gov site was not the first of its kind. While serving as the HHS CTO in the summer of 2010, Todd Park had led the construction of a less complicated predecessor, a project that went off without a hitch, launching a mere 90 days after the passage of the Affordable Care Act. In fact, it was the first website demonstrated by a sitting President, as President Obama took a few minutes to walk users through the features and benefits of the newly constructed comprehensive catalog of all available health insurance options in the public and private sectors in every community across the country.3

  The construction of that site—the first Healthcare.gov—­embodied some of the key elements outlined throughout this book. From day one, the project designers utilized not only open data but also Ries’ lean startup principles, in terms of their agile, modular, incremental approach to adding features, and the way that they assembled their staff.

  In January 2010, while Congress was actively debating the law, Ed Mullen tweeted out an unsolicited mock-up version of the marketplace, largely to help the public visualize what the law would mean for them in terms of a shopping experience. The tweet caught the attention of the administration and, within weeks of the law’s passage, HHS offered Mullen the role of lead designer.4 Guided by Park and White House New Media Director Macon Phillips, the entrepreneurial team launched a “minimum viable product” within ninety days to assemble the comprehensive list of all available insurance plans for users to sift through. The initial data set included a roster of plan benefits, as well as additional information—such as how frequently each plan denied coverage by applicants on account of preexisting conditions, and added surcharges based on a patient’s medical flags. Within a few months, the site increased users’ searching power by adding pricing data and, by March 2012, standardized the access to that data via a new “Plan Finder API” with the intent of allowing third-party developers to spread it across many different sites, beyond Healthcare.gov.5 To this day, that API powers sites including the U.S. News & World Report Health Insurance Finder.6

  That initial Healthcare.gov served its purpose for the time. But the passage of the Affordable Care Act meant that the site, by the start of open enrollment on October 1, 2013, needed to do more than merely present the existing public and private insurance options. It needed to become an online marketplace for a new generation of qualified health insurance plans that were compliant with rules effective on January 1, 2014. There would no longer be a need for information about surcharges or denials, since insurance companies could no longer legally discriminate against patients due to medical history. But there would be a need for a more robust e-commerce application that gave consumers the freedom to shop the qualified plans that were right for them, while simultaneously calculating new tax credits that would offset some of the cost burden, based on a person’s income and family circumstances.

  Above all, there would be a need for the underlying technology to work. And it didn’t. At least not in the early weeks after the official October 1 launch.

  So what happened?

  That’s a reasonable question to ask in light of my advocacy for the use of technology in government throughout this book. Some may use this painful episode to repudiate the administration’s technical work and, on a larger scale, question government’s capacity to use technology as a means of delivering vital public services. On the contrary, it should serve as a teachable moment, speaking to the need to reform procurement processes for the purpose of closing the technology gap with the private sector. The objective should be to operate in the most open and collaborative way possible, especially when dealing with a project of such massive size and scope.

  After all, in addition to presenting a new list of qualified health plans, Healthcare.gov had to assess program eligibility and calculate tax credits based on income and citizenship, facilitate consumer enrollment, manage financial services, and provide all of this with the highest standards for privacy and security. While trying to tackle this monumental task, HHS ran into limitations, some of its own construction, and some beyond its control. And, while at press time in December 2013, we don’t have enough facts to offer a definitive judgment about what played the overriding role in the site’s troubles, two areas do warrant further exploration, in order to avoid similar stumbles in the future—outdated and restrictive IT procurement rules, and political interference with technical implementation.

  As noted in Chapter 7, President Obama understood early that, to close the technology gap between the public and private sectors, we needed to reform the way we acquired IT goods and services. The administration added new tools such as authority to use prizes and challenges, and opened up more competition through initiatives like “RFP-EZ.”

  Unfortunately, much of the $80 billion spent on government IT goods and services is purchased through existing, broken contracting vehicles. In August 2011, career staffers at the Centers for Medicare & Medicaid Services (CMS) requested services to build the new Healthcare.gov system through a 2007 “Indefinite Delivery Indefinite Quantity” (IDIQ) contract, which had established a pool of just 16 vendors for providing a broad set of enterprise technology development services.7 Many of the sixteen (including Northrop Grumman, Lockheed Martin, and ultimate winner CGI Federal) had a long history of earning government contracts to provide IT services.8 But the list stood out more for the companies absent, such as Amazon and Salesforce.com, household names with greater experience in building Internet-scale e-commerce sites in the public or private sectors, but that weren’t competing for that line of government work, or that specific contract, back in 2007.

  At press time, HHS had not issued any statement explaining the 2011 decision to use the 2007 contract vehicle, rather than reopen the competition, though I suspect it mostly had to do with time constraints. It would be virtually impossible in today’s federal acquisition culture to post an open call for bids, select a vendor, and get the work done within a couple of years. It is an escalating practice among competing vendors to immediately protest an award they did not win and, as the Department of Defense noted in a 2007 memorandum, the resulting litigation process leads to delays that are largely out of the control of the agency needing to get the work done.9

  Whatever the reason, the unsatisfactory result led President Obama to reiterate, in even stronger language, his frustration with these procurement processes; he used terms such as “inexcusable” to “mind-blowing,” and even went so far as to suggest that the country needs to “blow up how we procure for IT, especially on a system this complicated. We did not do that successfully. Now, we are getting it fixed, but it would have been better to do it on the front end rather than the back end.”10

  Without question, there were tactical mistakes not just during procurement, but in the actual design of the project. As of this writing, however, it is not entirely clear what mistakes were made, which were most damaging to the site’s development, and who made them. Here’s a partial inventory of public speculation: that the
system wasn’t tested end-to-end early enough to leave time for corrections; that the software development methodology failed to account for a more dynamic environment where requirements might change (such as the number of states the federal system would have to support); and that there was an overreliance on remotely hosted enterprise software, as opposed to more scalable cloud computing architecture commonly used by Internet companies.

  My contention, however, is that the corrosive political climate also did much to undermine the overall cause. Simply, many public officials were not especially motivated to help the project succeed. Rather, they saw ideological and electoral benefits in its failure, and took actions that made that more likely. This was especially evident in the states, which were supposed to play a major, constructive role in the implementation of the law. That law had empowered governors and key stakeholders to lead the development of exchanges tailored toward that state’s citizens. The federal government was largely to stay out of the way, focusing instead on stitching together the myriad required databases into a data hub that would easily connect with each of the state marketplaces. That hub, featuring linkage to the IRS systems when calculating a family’s income, as well as to immigration status via Homeland Security and date of birth via Social Security, would eventually launch on time and without trouble.

  But the effort to engage the states didn’t go as smoothly, in spite of the federal government’s efforts to lay some groundwork. In October 2010, HHS announced a competition among states to serve as pioneers in building their own sites. A team from Oklahoma responded with a compelling approach to not only deliver on the baseline requirements, but a vision to upgrade the delivery of many other social services via the Internet. In February 2011, the state received the highest dollar award ($54.6 million) among seven “Early Innovator” winners, each of which committed to “assuring that the technology they develop is reusable and transferable.”11 Unfortunately, the new governor, Mary Fallin, wasn’t on board with health care reform. In March, Governor Fallin rejected the grant, returned the money and killed the project.12 In August, Kansas Governor Sam Brownback similarly returned his state’s $31.5 million share of the grant funds.13 Ultimately, 36 states—as of November 2013—had put themselves in position to be entirely reliant on the federal government’s version of Healthcare.gov, which was an odd stance considering that many of those states’ decision makers typically tout states’ rights.