Slashdot Mirror
Cost of Healthcare.gov: $634 Million — So Far
First time accepted submitter Saethan writes "Healthcare.gov, the site to be used by people in 36 states to get insurance as part of the Affordable Care Act, has apparently cost the U.S. Government $634 million. Not only is this more than Facebook spent during its first 6 years in operation, it is also over $500 million above what the original estimate was: $93.7 million. Why, in a country with some of the best web development companies in the world, has this website, which is poor quality at best, cost so much?" That $634 million figure comes from this U.S. government budget-tracking system. Given that this system is national rather than for a single city, maybe everyone should just be grateful the contract didn't go to TechnoDyne.
Or to put it another way, the procurement process selects contractors who thrive in the presence of bureaucracy, not those who actually deliver quality products on time and on budget. This is a well-known and long-standing problem.
[Sir Garlon] is the marvellest knight that is now living, for he destroyeth many good knights, for he goeth invisible.
The site had how many people try to sign up in the first day? If you want to compare it to facebook (a popular metric here no doubt) the number of people who attempted to access and sign up on healthcare.gov in the first day dwarfs the first several years of enrollment at facebook. If they had attempted to build a website to handle the load they faced (which will of course taper off quickly once the first wave of enrollees are signed up and done shopping) we would be bitching that they overbuilt the site because they would have tons of servers sitting mostly idle after the initial surge is done.
We need to wait until it has been up for a while before we go around calling it a failure.
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
This figure is not just for building a website.
It is for all spending with CGI Federal over the time that they have been doing business with the Federal government, including payments from fiscal years before Obamacare was even passed.
The figure is now being regurgitated by various right wing websites without anything that even passes for thinking.
And also now slashdot, which is disappointing.
That figure covers 114 separate contracts (see http://usaspending.gov/explore?tab=By+Prime+Awardee&fiscal_year=all&idvpiid=HHSM500200700015I&typeofview=transactions ) Not to suggest that it still wasn't overly expensive, but consider the fact that the system is a national transaction application that has to dip into numerous other federal data sources - and has a mission criticality above and beyond facebook. Still, many of us could have done it better and cheaper, but then again very few of us would actually enjoy working for the federal government and conducting our business the way any federal contractor is required to.
-- Religion is not an exact science
When a site loads 50+ .js files after you click an 'Apply' button, something is wrong with the design.
Handling what is potentially HIPAA-covered data? Much harder to do than just working with credit card information.
That's a lot of discrimination.
One website, that's expected to have incredibly heavy loads[...]
Well here's the rub. In regular operation, the loads aren't going to be incredibly high. They'll be "very" high, but not ridiculously. You could argue that their single largest mistake was trying to do a massive roll out to everyone in the country all at once. They should have rolled out to a small number of people, worked the kinks out and come back in a month with a slightly larger roll out. Rinse and repeat until it's available for everyone and you have some idea what your actual day to day usage numbers are going to be.
So all government contracts go to female African-American small business owners who are veterans? No wonder the government doesn't get anything done, the poor women must be totally overworked.
Plenty of talented developers and teams get crushed by government red tape, bureaucracy, and the simple inability of most government agencies to manage their contracts. I can't figure out why but there is an enormous attraction for government program managers to micro-manage. Having worked on a handful of very expensive, very large government programs I can tell you that either side can make a project a disaster. But I've been on teams that can roll out a successful commercial project in 3 months that takes 3 years for nearly identical functionality in the public sector (DoD in my case). It's not incompetence at the individual level, either, in my experience; it's something institutional. Too many regulations that cause inflexibility and twisted risk/reward feedback for both costs and personal performance, and the antithesis of an evolution-as-improvement driven culture to match changing development standards.
Way to over simplify website design, there's a completely lack of understanding of the system here. There's a huge amount of infrastructure that has to be put on the back end to make this work (Servers, Database licensing, Maintenance agreements, Security, Data Centers etc...)
For a system of this size, It's expensive. I agree with GP, $600 million is pretty cheap for a system intended to serviced over 100,000,000 people. Less than $6 a user is a pretty good deal.
Ars has a great article up going into more depth of why this happens so often here: http://arstechnica.com/information-technology/2013/10/why-us-government-it-fails-so-hard-so-often/
the government has lots of conditions you have to meet if you want a contract and you have to prove that you meet these conditions preference is given to women, minorities, veterans, small businesses, etc. its not a lowest bidder deal
Notice how everyone points out their favorite political cause as the reason for the failure, while the actual one dwarfs them all by comparison yet goes unnoticed? Anyone who has worked with the government before knows that the main reason everything is so expensive is bureaucratic red tape and auditing.
This is why an LED that costs less than a penny winds up costing the government $50 over its total ownership. I've looked at military contracts; Every LED in the system is individually serialized and tracked. You can't just order a bin of them, and put them on a shelf like you would in a normal factory. Even a ten cent screw has to be vetted through approved vendors, assigned its own serial number, etc. And that's just the screws for the toilet paper holder in the Pentagon. You don't wanna know the kind of process screws destined for fighter jets are subjected to.
So don't say "oh noes, it's because minorities are given preference!" ... which is a patently stupid thing to say anyway since they're paid the same as the non-minorities. That adds very little to the cost -- maybe a .1% bump due to the extra recruiting needed -- unlike the stuff I mentioned, which balloons it to many multiples of what you'd see in the same project in the private sector.
#fuckbeta #iamslashdot #dicemustdie
In early 2000's, I was involved in a loan processing project that cost the bank more than $500 million (if inflation adjusted, it would be more than $600 million) that never came on-line because it never really worked. It had similar requirements as the health care site (gather a lot of info from the customer, figure out the best loan option for them), and had much less user load requirements. It also employed one of the three top consulting house to develop it. The government is not the only ones who can screw up a tech project.
I've put in many RFQs on government dollars at universities, national labs, and private businesses (I've never been a direct employee of the government). All the law requires is that I get a quote (which usually turns out to be a no-bid) from a minority or woman owned business, and if that quote comes in over, the money still goes to the lowest bidder. The only extra cost is my time in getting another quote. Fair enough.
Pretty much every extra cost that I see comes about because someone abused the system in that specific way that the rule addresses. You can simply look at the process and see, ah, that rule or requirement was instituted because someone was either dumb or dishonest. No matter how rare or unlikely to occur again, however, the bureaucracy will institute a rule or procedure. Because that's what bureacracies do, private or public.
Toss in empire-building and that explains most of it. Though honestly the national labs have been far better places to work than the businesses or universities. Businesses are just as subject to these tides of human behavior as governments. They're just not as transparent, and you get fired for making them public.
I'm not saying this was that Healthcare.gov was the most efficient use of resources ever. On the other hand, the Facebook comparison is ludicrous. FB didn't have to serve 40 million users on day one; they got to scale up slowly. HC.gov is in the unenviable position of having to have a system which will handle millions of users (and almost certainly an overload) the moment it opens, then never having to handle that great a load again. In addition to having to do it in a way that absolutely protects the users HIPAA PII (so don't say cloud), unlike FB, which is in the business of making PII public and faces no penalties if it gets hacked.