Rackspace CTO John Engates offers a prescription for HealthCare.gov's complete recovery and continued health.
When HealthCare.gov launched in October, it was plagued with problems. It was slow. It crashed. Most importantly, the user experience was just plain bad.
But many of the site's main problems weren't necessarily technical. HealthCare.gov suffered from a severe lack of transparency, accountability, and oversight. There was no single owner -- no centralized management.
Late last month, I was invited to the White House and to the HealthCare.gov operations center to get an inside look at the work being done to fix the site. It seems a number of the initial problems are being fixed. One major first step toward righting the ship was implementing a tech surge. The government brought in some outsiders to work on the site and put all the contractors under a single roof. There is finally a single point of accountability.
During my visit, I went behind the scenes and saw how new life is being injected into the ailing healthcare exchange site. I met the vendors and contractors. I saw the hardware and software upgrades. I heard about the bug fixes. I saw where the bottlenecks were dislodged so traffic could flow. I saw firsthand the monitoring upgrades, the optimization overhauls, and the database updates.
Since then, one question I've been frequently asked is: Does HealthCare.gov need to be torn down and rebuilt from scratch? I don't think so. Like most modern websites, it is built in a modular fashion. It's not one big, monolithic code base. Modules can be updated, modified, and upgraded as needed. And through the use of APIs and web services, these modules can be tied together to create a seamless experience. It's not all that different from assembling an airplane. The wings, the fuselage, the electronics, and the engine are built in different locations, but they all come together and fit. But you can't build the wrong engine, try to bolt it on, and expect it to work. It's the same principle with HealthCare.gov. Unfortunately, that's what happened in this case. The different contractors built modules that couldn't be pulled together. In the end, it didn't fly. The tech surge is fixing that.
It is, however, still necessary for HealthCare.gov to bring aboard specialists to run the site in the long term. It will require the expertise of a team experienced at managing websites at massive scale. A number of disparate contractors are doing the work, but there needs to be a full-time, clearly defined team in charge of the site's day-to-day operation to replace the ad hoc team that came as part of the tech surge. A single group must oversee software and code updates and hardware upgrades, plan for changes, and maintain the architecture. Updates must be made frequently and proactively, not as just-in-time fixes in reaction to traffic swells. Think of it like a major e-commerce site that starts planning for Black Friday on Jan. 1. HealthCare.gov must start prepping for the next major deadline months in advance to avoid another major fumble.
One thing's for certain: For HealthCare.gov to succeed, it must be laser focused on the customer experience. When it launched, it was a bad experience. Aside from the crashes and latency, the workflow was all wrong. The first thing users encountered was an Apply Now button. Now they're prompted to compare plans before applying. HealthCare.gov should take a page from online tax preparation software. Greet users with a checklist of what they'll need. Offer them the ability to browse and then input their data, sign up, and check out. It's a lengthy, time-intensive process, and users must input a lot of data. Laying things out from the start and offering a clear step-by-step path to completion will help customers through the experience.
The site must also be elastic and nimble. It has to scale to handle peak loads. Some users have reported trying to apply only to be told to come back later. That's a terrible experience. The site should be able to accommodate people when they have the time. It's estimated that the retooled HealthCare.gov has been able to handle 1 million users in a day. That's great progress, but it also raises questions. Did all those people make it through the process? If not, where did they drop off? Where did they stall out? When did they abandon their session? All this information will be helpful in understanding and improving the customer experience.
This can be achieved with tools such as monitoring, analytics, and auto-scaling, so the site managers can know what's happening when, and the site can adjust itself on the fly to add capacity to accommodate a crunch of users. It's really no different from a sophisticated e-commerce site that has seasonal spikes and lulls but must remain available when those spikes hit.
Overall, I feel HealthCare.gov is on the right track. There is nothing fundamentally wrong with the site or the technology. The team in charge of improving the site now has a clear mission and purpose, and there is an increased level of intensity and accountability -- none of which was there before. The bar has been raised. A few common-sense updates, an experienced team of experts leading the charge, and a heavy focus on the customer experience can push it to success.
Interop Las Vegas, March 31-April 4, 2014brings together thousands of technology professionals to discover the most current and cutting-edge technology innovations and strategies to drive their organizations' success, including BYOD security, the latest cloud and virtualization technologies, SDN, the Internet of Things, Apple in the enterprise, and more. Attend educational sessions in eight tracks, hear inspirational and industry-centric keynotes, and visit an Expo Floor that brings more than 350 top vendors together. Register for Interop Las Vegas with Discount Code MPIWK for $200 off Total Access and Conference Passes.
InformationWeek Tech Digest August 03, 2015The networking industry agrees that software-defined networking is the way of the future. So where are all the deployments? We take a look at where SDN is being deployed and what's getting in the way of deployments.