This all started for me in October 2013.
Healthcare.gov had just launched, and if you remember, it wasn’t going too great. I was minding my own business in California working for Google. Although I had volunteered for the Obama campaigns, I had no other connection to the Administration and no government experience at all.
I was visiting some friends from the campaign in Chicago when through some complicated series of connections that I still don’t understand, I heard that the White House was looking to organize some kind of assessment of the website, and was asked to call in to a phone conference where I would get more information. The phone call was at 8:30 Eastern, so 5:30AM for me, so I was not too happy but I did it anyway.
On the phone was a guy named Todd Park. I had never met Todd; I looked him up on Wikipedia while on the phone. With a couple more phone calls, I was talked into flying to Washington, D.C. for a few days to evaluate the situation.
Here was the situation when I got there:
The government shutdown was over, and the failure of Healthcare.gov was the lead story on cable news all day every day. The original estimate from the Congressional Budget Office had been that 7 million people would enroll in this first enrollment period, but Todd said that given the circumstances if we could manage 4 million, that would be a home run.
He said they needed an independent outside team to figure out how bad the problems were and if they could be fixed to save the site this year. Also that if it couldn’t be fixed, then it would be a catastrophe for the policy, with the 2014 election coming and the unpopular law having failed to launch. Democrat support would probably fracture, the election would be a disaster, the President would be a lame duck two years early, and nobody would try again for a generation. But you know, we want your honest assessment.
“The whole system had worked as normal and produced the expected result,
which was a web site that was overpriced by hundreds of millions of dollars
and did not work, at all.”
We looked around and found some really surprising things. One was that there was no monitoring of the production system. For those of you that run large distributed systems, you will understand that this is as if you are driving a bus with the windshield covered. Second was that there were hundreds of people and dozens of companies involved, but nobody in charge. Third was that there was no particular urgency about the situation. As I would come to understand, nobody was acting like there was anything out of the ordinary because there was nothing out of the ordinary.
The whole system had worked as normal and produced the expected result, which was a web site that was overpriced by hundreds of millions of dollars and did not work, at all.
Since there were so many really basic problems, we ended up recommending that the site could probably be fixed to work well enough to enroll the 4 million people. Of course, we were then asked to stay to follow through and implement those recommendations. (This is called commitment escalation, and we will do it to you. This is the only time I will warn you.)
My 3 day trip had turned into nearly 3 month stint.
I submitted an invoice for the nine week period ending December 31, 2013, and the mean hours per day worked was 17.5. I was hallucinating and having other problems from not having slept enough for three months. This was the hardest thing I have ever done and I hope nothing ever comes close to it again.
“Things did not go back to normal for me; they only continued to get weirder.”
When I went home, other people had taken over the day to day operations and helped finish the enrollment period, which went to March 31. But things did not go back to normal for me; they only continued to get weirder. There had already been media interest in the people from the “tech surge,” and in March the final enrollment number of over 8 million was announced, and three of us from the “tech surge” team were on the cover of Time Magazine. In May I went back one more time for an event where we met the President.
I tell you all this not to prove how amazing we were, but exactly the opposite.
Nothing we did was technically difficult by any standard we are used to. There were simple problems with simple solutions. The solution to the lack of monitoring was to install monitoring. The solution to nobody being in charge was to make everybody come to the same room where we could coordinate. And so on. It was not a hard engineering problem, and any of you could have done it, and it was also more important and meaningful than anything I could have accomplished in a lifetime working at my old job.
Beyond the 8 million people that got access to health care for the first time, which for many of them was a matter of life or death, I don’t think we are ever going back to the time when people who actually need health insurance are not allowed to buy it.
So I went back to my old job and tried to care about it. I was not successful. On one hand the company does not need me; there are thousands of other engineers that are as good or better. On the other hand, if I succeeded beyond anybody’s wildest dreams the net effect is that some extra billions of dollars would go to one billionaire instead of a different billionaire. It was hard to see why I should bother, and still is.
But meanwhile, there were wheels turning in Washington, D.C.
People that had been working on government technology reform for years had received a huge boost from the Healthcare.gov meltdown. So pretty soon Todd Park was calling me again, asking me to come work for a group that has now become known as the U.S. Digital Service. It took a few months to convince me, but in August I finally left Google forever and moved to Washington.
“I expected that the difficult problems would be recruiting 10 more engineers to make this questionable life decision, and getting agencies to accept outside help.”
The basic theory of the U.S. Digital Service is to replicate the ingredients that were successful on turning around Healthcare.gov, for a few other high priority projects. With the money we basically had left over in an IT oversight fund, I expected to hire about 10 to 12 people and take on three projects, which I thought would be the Affordable Care Act, the VA, and immigration. I expected that the difficult problems would be recruiting 10 more engineers to make this questionable life decision, and getting agencies to accept outside help.
Since then several unexpected things have happened.
Where I was worried about recruiting 10 people, I actually got over a thousand applications. The demand from the agencies is also more than we could ever satisfy. We have met with 22 of them and identified around 60 projects that need attention.
And possibly the most surprising thing was that Congress passed an omnibus spending bill in December, and it included the full $20M funding request to operate the U.S. Digital Service. This will allow us to grow to about 40 people and take on 10 projects in 2015.
And this is the part where you come in.
Right now, at least, in this moment, we have overwhelming demand for engineers like yourself in the federal government. And we have overwhelming supply of people that want to come help.
For these reasons, we have proposed in the President’s 2016 budget a total of $105 million to create copies of the U.S. Digital Service team inside all of the 24 major agencies, for a total of around 500 people. The artificial barrier that has kept the technology industry and the public sector separated on different evolutionary paths is porous right now, and if enough people cross over, it can be destroyed.
Some of you, not all of you, are working right now on another app for people to share pictures of food or a social network for dogs. I am here to tell you that your country has a better use for your talents.
“All of these are design and information processing problems and all of these
are matters of life or death to millions of citizens and all of them are things
you can fix if you choose to.”
The Affordable Care Act, I just told you about. The Social Security Administration mails checks from a mainframe running COBOL, which might kind of be OK except that more than half of the workforce that maintains it is at or near retirement age. What happens then? The Department of Veterans Affairs has a serious backlog of disability claims. The U.S. Citizenship and Immigration Service processes permanent resident applications and everything else on paper. If you lose your green card, it will take you 6–8 months to get a new one and in the meantime you will not be able to prove you have the right to be in the country or get a job.
All of these are design and information processing problems and all of these are matters of life or death to millions of citizens and all of them are things you can fix if you choose to.
So that’s what it all boils down to.
This is a market economy and you — and your talents — are the crucial capital. Whether you know it or not, you are making decisions about how we allocate resources. We can solve any of these problems but we have to choose to do it.
The most sobering thing about my time in government is to really understand on an emotional level that this country belongs to you and me and it is exactly as good as we make it. Grownups are not going to fix it for us and billionaires are not going to fix it for us. We either do it ourselves, or nobody does. I will take questions.
Ready to get involved? Apply to join the U.S. Digital Service here: www.whitehouse.gov/us-digitalservice