It is strange that, with all the brilliant developers around the place, big sites like this can end up so poorly coded and built. It seems to be the same everywhere. I needed to submit info to a new govt website here is Aus last week and, firstly, the site didn’t even work in Chrome ( ), and in other browsers it was slow, spewed all kinds of errors, had no design or UX grace at all … pretty amazing.
As for the cost, I can only assume that such a gigantic system, that has to tap into so many resources and involves so many people, is a pretty overwhelming thing to build. I’d assume it involved a lot more than just the coding we would envisage.
The problem with gigantic sites like these with no business/profit motive is that they are politically driven by bureaucracy and not handled anything like a site of this size would be in the business world. So, politics, bureaucracy, and the general rule of finding the lowest bidder for a project all combined tends to lead to all sorts of technical and project management issues. Simply picking the wrong contractor for the job or not properly vetting them can also lead to problems.
Here’s a smattering of recent links now that the company who build the website has been splashed across the headlines. They might give an inkling as to what might have been happening behind the scenes on this project, but it’s important to note that some of the links contain political commentary and analysis, which should be taken with a grain of salt because not all the facts are out (or have been confirmed) and supporters of each political point-of-view are trying to use the situation as a political talking point. That said, you may have to weed through some things in order to get at what you’re looking for:
The amount of money spent by the federal government to create their website is half a billion dollars. But many states are also creating their own websites. The state I live in spent $100 million to create their website, market it, hire “navigators” to help people through the process and they got a grand total of 3,800 enrollments for health insurance of which only 400 were private purchases without subsidies.
I cannot rationalize how billions of dollars are being spent at the federal and national level to create websites where people can comparison shop for health insurance when these websites have been around for years. eHeathInsurance.com is one such site that has been around for 15 years. That site does everything the government health insurance insurance websites do with the possible exception of checking government records to see if a person qualifies for government subsidies. The government could have bought an existing site like the one I mentioned and tweaked it to do what it needs. Or, better still, one developer (or team of developers) could have been hired and the resulting websites shared among the different states instead of all the states throwing money away to create duplicate sites.
Part of the problems they are having is the crush of people going to the websites at the same time. For half a billion dollars and the billions being spent by the states to create their own websites, a huge data center could have been built with hundreds or thousands of servers. A system should have been put in place to limit the number of people allowed access, such as an alphabetical ordering or something.
Sometimes private sector businesses have problems with their new sites. Target had big problems two years ago when they debuted their new online shopping site. When I read about the problems I visited the site and it was very slow. They fixed it. I would argue that a site like Target’s is more complicated than an insurance comparison shopping site. And I’ll bet Target did not spend billions of dollars creating their site.
While there have been some problems, it is not like there aren’t enrollments. For example, in the state of Oregon already they have signed up 56,000 reducing the number of uninsured people in the state by 10%. Certainly it becomes a big deal if they don’t get it faster soon, but it would be expected to have some problems on a roll out of this size (49.9 Million don’t have health insurance). There might even be some hackers who are trying to slow it down. I’m sure there are some competent people that will tweak it and make it run better in the coming weeks.
The US GDP is about $16 Trillion and healthcare represents about 16% if that or or $2.5 trillion. Spending $500 million is a small price to pay to start organizing a better system than we have now. Hopefully both parties will get together instead of encouraging Americans to rally against one another. ACA is a new beginning, and like anything this complex, it will take continued improvements over time.
The issue isn’t that the websites are having problems. That is to be expected. The real issue is the insane amount of taxpayer money spent on them to duplicate the offerings of existing privately owned comparison shopping sites. Billions of dollars are being spent on these websites nationwide and for that amount of money they should have had more than enough capacity to deal with the crush of people viewing the sites. Or at least had a system in place to mitigate problems like people whose last names start with “A” can register on one day, “B” the next day, and so forth.
As far as the Target website I mentioned earlier, I viewed that website at something like 3 AM and it was really slow. It might have been a problem of too many cooks in the kitchen or it was rushed to meet a deadline. Same thing might be behind the problems the government sites are having. I always underestimate the time something is going to take me to complete. I’m probably not the only one. But how many developers have a budget of $500 million?
The issue that the cost is ridiculously excessive for a website of this size and complexity. It has nothing to do with how much of the GDP the health insurance industry uses up. In short, the cost is so high because the project was bungled so spectacularly.
Hopefully both parties will get together instead of encouraging Americans to rally against one another. ACA is a new beginning, and like anything this complex, it will take continued improvements over time.
The problem is one party just wanted a feather in their cap for implementing ACA for everyone, the other says it’s broken, expensive, unsustainable, and hurting the economy.
Here’s just my 2 cent guess and I won’t be surprised if this is actually true.
The project had length of 5 years. I’m curious to know how many people have transitioned in/out of the project. Something tells me that this project had high turn over due to various reasons. Lack of management, extraordinary stress, customer unwilling to comprise deadline, customer who values deadline over code quality, lack of developer motivation due to stupid code that someone wrote many years ago, and etc… I’m pretty certain this is the reason why it’s in poor state.
According to an article in this week’s Bloomberg Businessweek magazine the problems are associated with a government contractor, CGI Federal which is integrated into the exchange websites to look up relevant data to determine options and costs. CGI Federal operates a proprietary back end that was/is not scaled properly to handle the load. The article then goes on to fault CGI Federal for not participating in and taking advantage of open sourced software.like Firefox or Google Chrome.
I believe we’re not considering the very difficult complexities of the sytem. Things like HIPPA constraints, state to state and company to company data transfer interfaces that need to be properly built, usually between disparate technology systems, and the immense complexity of medical-related data structures.
I doubt the front-end web access has been much of a problem, most everything I’ve read says the problems occur with data-related processing initiated from the web interface, things like account creation and getting available plans for a given geographical area.
Politics caused some of the problems too, what with so many states refusing to participate in medicaid expansion, which pushe a much larger-that-expected number of users to the federal healthcare website.
I do find it entertaining to listen to all the non-tech talking heads in media spouting off about technical problems that they have no understanding of, giving simplistic “solutions” that have no chance of working. So for the amusement value alone, the healthcare.gov website is a rousing success
My analysis is that there is no problem with the website itself, but it uses dependencies which are being overwhelmed. There was a last minute request from the government for contractors to require user registration. Since the registration form queries a large variety of databases for validation, those databases probably went down. As those databases were not associated specifically with the website, they most likely didn’t get any of the $$$, and the programmers assumed that their API call would always go through.