The healthcare.gov launch in the US in 2013 has been well documented as one of the most disastrous product launches ever. In the aftermath of that launch the United States Digital Service was formed with a focus on improving the way IT services are delivered in government. Dana Chisnell, who has a long and storied career helping governments with user experience and is currently part of the USDS team, joined us at Mind the Product to share the legacy of that failure – a new approach to software development in the federal government, and a new way to think about the citizen experience.
The first healthcare.gov system cost $200m to build and $70m/year to maintain. You’d think for that kind of money you’d have an amazing user experience and robust site uptime. But it wasn’t. On the first day 2.8 million people tried and failed to sign up on the site, and 70% of these errors were caused by something as simple and seemingly innocuous as the sign-up form.
Authentication is a repeated problem across all of the services and interactions users have with the government, practically every service the federal government provides has this problem. And this is because they have made it hard for themselves to deliver a service by forgetting that they are actually in the service business. And that service is a designed experience, whether you are intentional or not.
The redesigned healthcare.gov took 8 engineers to rebuild in less than a year, cost $40m to build and $4m/year to maintain. It’s error rate is less than a tenth of a percent – 10 errors in the first month, not 10%. The response time went from 20 seconds to 20 milli-seconds. And it has allowed millions who have never had health insurance to sign up for it for the first time.
Thus redesigning the citizen experience isn’t just about the user interface, but it happens at every layer of the service, from the back end to authentication systems to the many touchpoints for the citizen.
Watch this illuminating talk for more examples and insights into how the USDS is doing this on a grand scale by continuous delivering improvements.
8 engineers / 1 year and still 40M to build? I get that there’s overhead and people need to be paid well, but where did all the money go?