Government is not a Startup

Before going into the details, let me first couch myself in saying: I really like Steve VanRoekel. I think he was the glue that held together the FCC's New Media endeavors, I think, after meeting with him and talking with him on several occasions, that he's really smart, and I'm really excited that he's the federal CIO. In his first talk in that position, VanRoekel gave an outline of his agenda, the Future First Initiative and I think, all in all it's great stuff and good thinking.

But there's one eensy part of the talk that nudged me the wrong way. And it's something that we hear a lot coming out of the government, and out of silicon valley -- this concept that government should behave more like a silicon valley startup. It's not our CIO saying that, either. It's a consistent meme running across the techno-elites at the intersection of government and technology.

This meme -- like it's older sibling "government should be run like a business" -- is a particularly attractive one because it helps to simplify how government works, and make it into something that we can identify with. Who in Silicon Valley wouldn't want to hear that government wants to behave more like them? It's an honor! It's also attractive because it's filled with the great emotional stories of innovation and success and of the take-no-prisoners, change the world stuff of the late Steve Jobs and the rest of the innovators. After all, the United States government put a man on the moon, unleashed the power of the atom, sponsored that nifty Siri on your iPhone, and yes, developed this platform you're reading these words on called "the Internet." So it's not like government's innovation, especially in the IT sector isn't sewn into the DNA of every startup coming out of Silicon Valley (whether they like it or not).

The metaphor of the startup can be appropriate sometimes, I think it's time for us to move beyond the vague "startup" concept into more specific changes in policy. While a startup's agility is a great, romantic notion, it's a distraction from what our current problem with IT is inside of the federal government: a lack of ability to critically think about technology. Specifically, I think this breaks down into three things that need to get fixed.

First, there's culture. You have contractors who want to deploy "enterprise grade" software that costs tens, if not hundreds of millions of dollars for the government. Because of the way the human brain works, bureaucrats think they're saving the taxpayer money when they take a 50 Million Dollar proposal and knocking it down to 20 Million Dollars, but more often than not -- especially when it comes to the web, that same project could cost less than a million dollars for anyone in the public to develop. For instance, that Recovery.gov's website budget alone is more than three times larger than the annual budget of the Sunlight Foundation -- the leading and one of the larger organizations focused on transparency at the Federal level.

Part of the reason why that budget got to be so large is because the Recovery and Transparency Board hired contractors to write the RFP, who are part of this same culture of enterprise as the people building the software, and thus they asked for a bunch of things they didn't need. Read the original RFP, and you'll see. Ask your favorite web developer if they've ever set up an XML Firewall, or have any experience with "Enterprise Data Cubing."

Regulation is another big issue. Because procurement is such a headache, and often adds between 8 and 18 months just to get a contract in place on top of development time, in the world of technology, this means that government is forced to live a cycle behind on Moore's Law, something I spoke about at Code for America a few weeks ago. We need for procurement rules to change so that government can get better access to people and talent, and so that the gap between what kind of technology people expect and what government can provide (what VanRoekel calls 'the Productivity gap' but I think is more of an expectations gap) starts closing. We have a system designed to prevent fraud, and mitigate failure, and what that does is give government yesterday's technology at tomorrow's prices.

A third, and I think often overlooked reason why government cannot think critically about technology is that it does not participate, socially, online in the same way that everybody else does. As I mentioned in my post about tools for brainstorming and dialog if government wants access to expertise, the best way to do that is not to use or develop new systems to get that expertise, but rather to go to where the experts are and ask -- it must stop carving out special places for citizens to participate with government in an official capacity, and start soliciting input and sharing its own expertise where experts are to be found.

All three of these things can be solved without government "behaving more like a startup." Nobody really wants government to behave like a startup -- nobody wants their government to "pivot" or to bet the farm. We certainly don't want it obesessed with it's own margins, and until Marlboro Miles become legal tender, there's a bright line between what government can do and what business can do. We certainly don't want the government to worship sunk costs, and we do want it to be "more agile, like a startup", the fact is: most startups fail.

And more often than not, government cannot afford that luxury. Take the story of USAJobs.gov. Government did the thing that I've been hoping they'd do for a long time -- they stopped relying on contractors, got some developers and developed their own software in-house. It ought to be that USAJobs.gov can launcha minimum viable product and iterate, but instead what happened was, the website launched with some bugs and Congress flipped out. Congress isn't requesting iteration, they're requesting heads on platters. The final step and the most difficult one is to change the expectations of the public and, frankly, of Congress. Government can't innovate unless there is less political outrage in failure.

Government's best moments of innovation were when government leveraged its strengths of being government, and used those resources and its patience in order to innovate. So I think it's time we moved beyond this simple "government should act more like a startup" idea. We need to admit to ourselves that that notion isn't pragmatic: the economic incentives aren't right, and the cost of failure is too high. Government need not be a startup in order to innovate. Instead, let's focus on what the actual barriers to innovation and change them.

Comments

[Carl Malamud]

You mentioned the USAJobs.gov hooptedoodle where "the website launched with some bugs and Congress flipped out." The issue here is leadership: the administration should have stood up more forcefully for their development teams if they were convinced they were on the right track.

Too often, government executives treat their tech staff as contractors, be they in-house civil servants or beltway bandits, caring only about paperwork and meetings with superiors, and not about building real systems that serve real needs. As you say, that takes iteration and elbow grease.

Steve VanRoekel showed impressive real involvement in the nuts and bolts of the FCC when he was managing director. I'm hopeful he'll bring that same day-to-day involvement to his new post.

[Tilak_joshi]

Excellent points!

Having seen this happen from the inside, I completely agree. The inefficiency and unawareness is frustrating at best. The standard process (deliberation, approval, implementation, testing, review) for technology solutions, enterprise and, less commonly, in-house, ranges from six months to two years at times. This can be lifetimes for a startup. This level of inefficiency is the norm, and precedents need to be set to disrupt this pattern.

Furthermore, lack of social media participation is preventing transparency and, many times, makes it very difficult to identify a problem until it's too late. There needs to be a move towards more crowd engagement like Bloomberg's NYCBigApps competition applying OpenData and other crowdsourcing initiatives. Social media is definitely an asset in breaking open the flaws in government technology.

With more transparency and crowd engagement tactics using social media, and governing initiatives at a central level, efficiency in technology will be encouraged not only within federal agencies, but also outside of federal agencies.

-tj

Tilak Joshi
website: http://www.tilakjoshi.com
twitter: @tilak_joshi

[Gregorylent]

government is inertia, its only job is to perpetuate itself.

anything gov2.0 that really changes things will be resisted 100%.

accept that, and act accordingly.