2014-05-27

Co.Labs

Why Exactly Does The Government Suck So Badly At Software?

Everyone agrees that the governments should become platforms for the services they provide to citizens--so why is that so hard? Here’s how one company is trying to fix Washington's digital disasters--and what still needs to be done.



Earlier this month the Senate Subcommittee on Financial Services and General Government held a hearing in which they called on Steven VanRoekel, the nation’s top chief information officer, to explain why the federal government is so bad at developing and managing IT projects. This wasn't just about the rollout of Healthcare.gov last year, but also the other 200 major government IT programs that are delayed, over budget, or at risk of catastrophic failures.

These 200 programs currently have a whopping $12 billion total budget that most startups--even established companies--would kill for. And that $12 billion is just a fraction of the $82 billion the government will spend on IT projects this year alone. If the government clearly has the financial resources why are citizens more likely to see federal digital services launches go the way of Healthcare.gov instead of blooming like Facebook?

The answer is twofold, according to the people I spoke to. First is the government’s inability to think and act in an agile way. Second is failing to make "coding for the Government" sound like a sexy gig for young developers. Here's what Uncle Sam needs to do to turn things around.

The Government Doesn’t Think Like A Startup

Matthew McCall is a health technologist who last November created a popular petition to get the government to open-source Healthcare.gov when it was going through its darkest days. But besides being an open source activist, McCall works for the nonprofit , where he’s helping to build and grow a community around VistA, an open-source government-created Electronic Health Record currently run by the Department of Veterans Affairs. This means that McCall is one of the only people among the large group of critics of government IT projects that is actually working from the inside to try to fix them.

“In short? The government doesn’t take the same approach to software development that startups do,” McCall says when I ask him why the U.S. government is so bad at rolling out big digital services when it has tens of billions of dollars at its disposal while startups with virtually no money and a skeleton crew often achieve wildly successful product launches.

McCall says a startup is focused on building a “Minimally Viable Product” where it brings the solution to market as quickly as possible in a basic form, and then iterates on the product using agile methods. This allows the startup to gather user feedback and rapidly iterate and improve the product and is almost the exact opposite of how the government works, which is more of a waterfall method.

“Government development focuses more on gathering comprehensive requirements up front, issuing a contract for the work, and managing the contractor during the buildout,” says McCall. “This ‘big bang’ approach typically means longer development time with little to no customer validation. If requirements do change over time, it is usually driven by competing organizational interests rather than customer needs, which makes it very easy to get away from understanding if what is being built is actually useful or relevant any more.”

McCall says there are unfortunately a number of regulatory policies in place that make it very difficult for the government to function more like a startup. He points to The Paperwork Reduction Act as just one example. “[It] makes it nearly impossible for developers in the government to ask questions of the public in a timely manner, so user-validated development is very challenging. There are also policies that dictate which technologies you are allowed to develop in, and how you implement them. These are generally a good thing, to make sure software is maintainable and secure, but it adds a great deal of overhead in terms of process and time, and makes using emerging technologies which a startup would use a much tougher sell.”

But it’s not just bureaucratic red tape that stymies billion-dollar IT projects. It is also a cultural difference. McCall says the government is risk-averse by nature--something anathema to the ethos embraced in Silicon Valley. But of course, if a startup gets hacked, it’s not nearly as big of a problem if the government does. It’s a point McCall concedes, but he notes it can often be used as a red herring to keep the organization from trying new and interesting things.

“While I was in government working on Blue Button, I was yelled at by a developer claiming that I was going to wind up killing patients because I was giving them access to their medical records,” he says. “[It was] a completely untrue notion, but it was enough to give me pause, and I’m sure there are similar stories all over government.”

OpenGov To The Rescue

At the Senate Subcommittee earlier this month Sen. Barbara Mikulski, D-MD., said, “We spend billions... on technology projects that are often ineffective or lacking in utility. And often, we end up doing it again and then again and then again. And I think it’s time we get our arms around this.”

It is those billions that probably worry most senators more than any other problems with government IT projects. After all, we live in an age of austerity and “government waste” is sure to anger voters going to the polls later this year. But the thing is the government is rife with waste--and not just at the federal level. State and local governments waste billions every year due to inefficiencies with sorting through financial data to see where they are actually spending their money. It’s a problem in search of a technical solution that you would think most state governments would be actively working to solve. But in fact virtually every state and local government still relies on the 30-year-old spreadsheet to sort through their big data.

That’s where Zac Bookman and Mike Rosengarten come in. Bookman is the CEO and Rosengarten is the CTO of OpenGov, a startup that empowers state and local governments to ditch their spreadsheets and visualize big data all through the humble web browser.

On the outside OpenGov may seem like a relatively “boring” startup compared to Oculus Rift or other hip headline grabbers, but it very well could have a greater impact on your day-to-day life than anything Facebook offers. That’s because it solves the problem of the government’s big data overload by organizing government’s financial data into dead simple visualizations officials--and residents--can act on.

“There is an epidemic in governments of all shapes and sizes across the country in that they use enterprise accounting systems that run on software written 30 years ago,” says Bookman. “If you are the mayor of a city and I ask you a basic question about your data like, ‘How much have you spent on police hours over the last five years?’ you probably don't know of the answer. You often will have to call IT to run a report from the accounting system, or you'd have to call finance to weigh into a 20,000-row Excel spreadsheet to pull out a few disparate lines of data. Some people would laugh at that but we find it to be a really serious issue and so do those who are on the inside who are struggling with the technology because they need that data to make real-world decisions.”

Those real-world decisions can sometimes affect lives. For instance, cities using the OpenGov platform can now see in a matter of a few clicks how much they have spent on fire and safety and if they can pull money from any other areas to give more to fire precincts ahead of deadly wildfire seasons.

Because of OpenGov’s ability to quickly visualize complex financial, census, and other big data it has often been called the Mint.com of city governments. And it’s a service major cities like Los Angeles and Palo Alto are using.

“What's unique about OpenGov is it was developed in a creative partnership with a sophisticated local government--the City of Palo Alto--so it represents the best of both worlds,” according to Rick Cole, Deputy Mayor for Budget & Innovation, for Los Angeles. “It originated from the desire of a local government to create a tool that didn't then exist in the marketplace, but it now has the industrial strength of a constantly evolving product with a broad customer base that supports ongoing development.”

Cole notes that OpenGov isn’t only useful to city officials either, as it has a front-facing portal any resident can view in their browser. “It's to our benefit as public servants to demystify budget data to rebuild trust through transparency and accountability. As a user-friendly and dynamic platform, OpenGov allows us to show residents how we are investing in specific priorities.”

But again, if OpenGov is the tool local governments have needed for years to make sense of decades worth of big data, why hasn’t the government itself created the solution before now?

The Talent Problem

The answer to that question, according to OpenGov CTO Mike Rosengarten, is that while many local, state, and even the federal government might have had the drive and even the money to produce such digital tools, they often lacked the talent.

“This particular topic resonates deeply with me because as a software engineer and now an entrepreneurial software engineer, I see it from a few different angles, but one of the most important and prevailing issues for me is that the best engineers don't want to go work for government. They want to go to Google. They want to go to Facebook,” Rosengarten says. “So when Healthcare.gov or any of these other local municipalities need internally built tools they build their own solutions and they're just often subpar and really expensive. The best engineers are just not going in to local government saying, ‘Where can I get my hands dirty? How can I help? How can I innovate?’

The lack of interest from young developers who want to work for Uncle Sam is primarily down to the poor recruitment tactics from state and local governments compared to the big tech companies.

“I've done a fair amount of recruiting at this point, especially out of colleges, and what I've come to find is that if you're in front of the student's eye, you're going to get their attention,” Rosengarten says. “I've yet to see the City of Pasadena come to my school and say, ‘Come work for me.’ So I think part of it is just the presence, the recruiting effort that these large companies are taking, and maybe it's that they have a financial backing to do it.”

“I think if more students understood the problems or that the potential opportunity to solve real hard challenges with the local governments, they would get more excited and they'd probably join. But the problem is they're just not getting that information, and so they're diving into whoever is buying them dinner, taking them out, and whatever else recruiting techniques these big companies are using.”

Bookman points out that he and Rosengarten took a group of top students from the University of Maryland to dinner once as the students were visiting companies in Silicon Valley. “They visited Twitter, Square, Dropbox, and Google, and all these students are talking about how cool it is that Twitter has an arcade in their office. At that point they don't even seem to be thinking about the business of Twitter itself. They're focused on perks. How can local government compete on perks against Google?”

The perks and sex appeal of the major companies are indeed one of the obstacles the government needs to overcome in recruiting top talent, agrees McCall. But he stresses the most important is thing the government can do is explain to coders why their talents are needed and how those talents can be used to code a better world.

“The government’s culture is very different from that in the startup world or at companies like Google,” says McCall. “I’ve worked at some companies with great perks and I get why people want to work for those companies. I don’t think working for the government will ever be ‘sexy,’ but that isn’t why people should want to work for the government.”

“I think what the government could do is a better job of explaining why you would want to be part of the government. Working for the government is a form of public service, and you have an opportunity in doing so to help be a part of the solution to our country’s challenges, rather than an armchair critic. The guys that fixed Healthcare.gov are heroes to me. I got a great sense of pride in working for the government, and I know many people in government who love doing what they do because of their sense of duty. They might not be as big of heroes as our troops, but what they do every day matters. If government can attract and retain people who want to make a difference and are given that opportunity, I think it will go a long way.”

The Future

While the government has many obstacles to overcome McCall is hopeful about the future of a government’s ability to handle massive technology projects. As evidence he points to how the U.K. government has recently completely revamped their web services with www.gov.uk.

“They are doing some really interesting stuff in the U.K. right now, trying to simplify how their citizens interact with government via a unified experience,” McCall says. “Rather than having the thousands of websites we have here in the States, they are trying to have a single point of contact, and affect cultural change at the same time.”

But McCall cites examples closer to home. He points to 18F, a digital services agency within the United States federal government that is based on the lean startup model. It was created after the disastrous rollout of Healthcare.gov and aims to reform the front-end of government systems its citizens interact with on the web.

“I’ll also throw in a plug for the Presidential Innovation Fellows Program, which is now starting up its third class,” he says. “It’s a great way to have C-suite support for trying innovative things in government, and tackling pointed challenges.”

Are these initiatives enough to ensure a debacle like Healthcare.gov never happens again? Of course not. But each shows the government shares at least one crucial element with the startups that are successful: Learn from your mistakes and keep trying.

[Image: Flickr user Oliver Kliewe]




Add New Comment

2 Comments

  • The government doesn't suck at software; they suck at hiring the right contractors and then managing them.

    As astronauts have commented since Mercury, it was built by the lowest bidder.