Evan Burfield is the co-founder of 1776, an incubator platform for start-ups in education, health, energy and smart cities. He founded and, from 2006 to 2011, was the chief executive of Synteractive, one of the contractors that built federal Web sites such as Recovery.gov and Treasury.gov.
HealthCare.gov was launched Oct. 1 with bugs that made it impossible for many Americans to purchase insurance. Although these problems are the focus of many tirades and jokes, it’s not a disaster — in a few months the site will work as expected.
Moreover, in the range of federal IT debacles, HealthCare.gov doesn’t come anywhere close to the worst. Over the past five years, agencies including the Justice Department have had to scrap software projects that cost far more than the $70 million to $125 million (at least) reportedly spent on HealthCare.gov. In 2010, Justice’s Litigation Case Management System was canceled after four years of development because it was going to cost an additional $193 million to make it usable. But HealthCare.gov’s issues remain troubling, and the government can still learn from the site’s failures.
Government Web sites don’t have to be costly to handle extreme amounts of data and work well. In June 2009, the Recovery Accountability and Transparency Boardsearched for a team to build a Web site that would enable Americans to understand how the $900 billion in Recovery Act money was being spent. Numerous data sets would have to be collected from more than 100,000 organizations nationwide and then transformed into something meaningful.
The digital agency I led at the time, Synteractive, was geared toward building Web software for start-ups and businesses. We found a partner, Smartronix, to help us navigate the federal IT world and spent two weeks cranking out a proposal. Figuring that we didn’t have much chance of winning, we included all sorts of ideas that we thought would be powerful but could be difficult to bring to fruition. We proposed making the primary user experience an interactive map, making all of the data accessible via open standards for developers, and hosting the entire site in the “public cloud.” Such things had never been done before in government work at that scale or level of visibility.
After the Recovery Accountability and Transparency Board gave us the green light, we had less than four months to create the Web site. The back-end data were collected by a separate team, but we had to figure out everything else. With our start-up experience, our approach was to keep the team small and work side-by-side with government officials to shorten the decision-making loops. Simplicity was the mantra for everyone involved.
After Recovery.gov was unveiled, critics pointed out errors in the first set of data reported by funding recipients — but that was the entire point. By giving citizens direct access to the raw data on where their money was going, they could hold both the government and the recipients accountable. The data became better and better each cycle. Newsweek eventually referred to the site as “perhaps the clearest, richest interactive database ever produced by the American bureaucracy.”
Because Recovery.gov was built on a public cloud, it was able to respond to demand as needed. Even when a major outage on the cloud crashed other sites, Recovery.gov didn’t go down.
The cost to taxpayers? About $5 million. Given what we were able to produce on that budget, it’s hard to take seriously charges that HealthCare.gov was underfunded.
What start-ups understand but the government often fails to appreciate is that having more money and time can often impede building great technology because it allows people to make things more complex than they need to be. Most federal IT projects continue to bet on methodologies and approaches that have long been discredited for commercial enterprises.
A start-up’s cost to build great software is one-tenth of what it was about 15 years ago. In that time, the federal IT budget has almost tripled, to at least $80 billion. Why are Americans seeing only incremental improvements? Because federal procurement rules make it difficult for agencies and contractors to change operations.
A number of practices contribute to persistent dysfunction. The government continues to prefer massive, long-term contract vehicles that prevent open competition among our most innovative technology providers. How can the government benefit from the latest practices when it has to choose — as in the case of HealthCare.gov — from among a limited subset of contractors preselected years in advance?
Furthermore, the government continues to select providers based on their ability to produce byzantine proposals rather than letting them compete by rapidly building early prototypes that showcase their ability to build innovative technology.
Recovery.gov succeeded because it acted like a start-up in the face of almost every structural impediment in the federal IT contracting system. That would have been impossible without government officials who were willing to find creative solutions.
The issue is not one of talent. The United States has some of the finest technologists in the world — many of whom work hard to achieve important missions for government. The issue is that this talent is thwarted by rules written for a previous generation.
This status quo cannot hold much longer. The best result that could come from the attention on HealthCare.gov is if Congress seeks to reform how the government procures IT contracts. Americans of all parties want an effective government that costs taxpayers less. Procurement reform would be a win for all of us.