If you follow politics at all, you’ve likely got an opinion about ObamaCare, or the Affordable Care Act, passed by Congress and signed by President Obama in 2010. As the recent government shutdown began, a major provision of the law was activated. Healthcare.gov was opened to the public on October 1, and it became immediately obvious that the site was not ready for prime time.
Whatever your stripes politically, the failure of this large web application demonstrates a very large problem in Washington, and it is not about healthcare or socialism or the President. The problem is found in government procurement procedures.
As a professional startup guy, with success and failure under my belt, I can tell you with certainty that this is insane. $292MM is a massively successful exit for a large team. $292MM is an E- or F-round for a company about to go public. It is not the price tag for building technology.
In the world of startups, we, smart, driven people, take risks. The convince folks with money that our ideas are good, our teams better, and our plans flexible and perfect. When the idea, team, and money come together, we press forward, but the battle in which we engage is almost universally one of market, not one of technology. Technology is a P- (easy) problem. Market fit is NP (hard).
Healthcare.gov was a large challenge, but it was a problem of technology, not one of market fit. The parameters for the problem were specified, and manageable, and no one was going to have to pivot to meet market demand. Data-interchange is tricky, given. Synchronous access to data across multiple systems is also a challenge. Predicting traffic is a challenge, especially when being done across multiple systems.
This was a difficult technological challenge, but going in, the goals were set. The site simply needed user stories, tests, development, and quality assurance. Then it needed to be tested against a barrage of traffic. Even then, it might have gone down. But the catastrophic failure we’ve seen indicates that the problems with the site are fundamental and systemic.
Given the error messages that have been circulating on the Internet, it seems that engineers didn’t bother to envision such rudimentary failures as an offline database, or a missing secondary server. This suggests that the contractors building the system did not use even a modicum of industry best practices: Test-driven design, acceptance testing, load testing, etc. It is absolutely inexcusable that, given the tools we have at our disposal today, any technology fails unexpectedly, in the manner healthcare.gov has.
How is it possible that this colossal failure cost taxpayers almost Three. Hundred. Million. Dollars?
How is it possible that someone in government thought they needed to spend even half of that money to do something so linear? Is this what it means to be a government contractor?
That’s a lot of money for 1980’s technological paradigms, and test-free development of fault-intolerant system. It’s a high-stakes failure that ought to offend us, as startup devotees, as technologists, and as entrepreneurs.
I, for one, could have failed at this project for a lot less money.