![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Thoughts on the healthcare.gov stuff...
Oct. 23rd, 2013 09:10 amJohn McCain said something understandable but really rather dumb the other day when he demanded they fly some clever people in from Silicon Valley to fix the ACA site. It wouldn't help, and not necessarily for the reasons people think.
First off, everything I've heard about the project sounds like classic large scale IT project crap. Huge goals, huge teams, lots of people with limited technology understanding wanting things included etc... it also sounds like the company they picked has limited experience in consumer facing web properties and they had a very poorly thought out test strategy and even more limited/non-existent beta program.
The core issue here, for me, is that the way anybody sane from Silicon Valley would have built this product is exactly how it can't be built, and that's really at the heart of the issue.
Complaint 1 I've heard is that unlike Amazon you can't just start shopping. No, you can't. It's closer to care insurance in that regard, and this is an area where the US really is way behind the UK, you need to put your details in before you can shop for car insurance, you need to put your personal details before you can get a quote for medical insurance. It's not really like Amazon at all. They also have the issue that insurance systems didn't just leap onto the stage as perfect web based tools. Behind every insurance web system, there's a very very ancient but effective system churning out quotes. The ACA has the problem of needing to interface a bunch of systems that didn't previously exist.
Complaint 2 - speed etc... again, if you asked your Silicon Valley team how to do this, they'd probably dump the hosting issue on Amazon and let them worry about server loads. I'm fairly sure that apart from the HIPPA related nightmare, people might raise an eyebrow about having millions of American's health details stored on commercial third party hardware.
Complaint 3 - wrong products being sold. See 1. The testing for something like this is a nightmare, I'm sure they didn't have robust enough a test regime, on the other hand, given the scale of the system, I'm not sure what could be done.
Complaint 4 - it's failed it must be scraped. Not a recommendation I'd make until you'd like to blow another $250m.
So, what would I have done differently. I'd have looked for a company with experience in large insurance systems before and combine them with other consulting companies who really understand the US insurance system, rather than pulling in a monolithic systems integrator as they seem to have done. I'd have made sure the testing and beta program was planned and I've have begged for a staged roll out. And this is really the problem, I'm not sure that's possible.
We're launching a new product at the moment. We spent 6 months building stuff and testing it on lead, tame customers. We then opened it up to a beta program with a small number of users and took their feedback and the bugs they found and fixed those. Now we're ready to go into an open launch. We're bound to hit more roadblocks on the way and find more bugs. But that's the way people release software, and unfortunately it's a terrible model for a national healthcare based IT project.
There's a reason the rest of us have single payer systems.
First off, everything I've heard about the project sounds like classic large scale IT project crap. Huge goals, huge teams, lots of people with limited technology understanding wanting things included etc... it also sounds like the company they picked has limited experience in consumer facing web properties and they had a very poorly thought out test strategy and even more limited/non-existent beta program.
The core issue here, for me, is that the way anybody sane from Silicon Valley would have built this product is exactly how it can't be built, and that's really at the heart of the issue.
Complaint 1 I've heard is that unlike Amazon you can't just start shopping. No, you can't. It's closer to care insurance in that regard, and this is an area where the US really is way behind the UK, you need to put your details in before you can shop for car insurance, you need to put your personal details before you can get a quote for medical insurance. It's not really like Amazon at all. They also have the issue that insurance systems didn't just leap onto the stage as perfect web based tools. Behind every insurance web system, there's a very very ancient but effective system churning out quotes. The ACA has the problem of needing to interface a bunch of systems that didn't previously exist.
Complaint 2 - speed etc... again, if you asked your Silicon Valley team how to do this, they'd probably dump the hosting issue on Amazon and let them worry about server loads. I'm fairly sure that apart from the HIPPA related nightmare, people might raise an eyebrow about having millions of American's health details stored on commercial third party hardware.
Complaint 3 - wrong products being sold. See 1. The testing for something like this is a nightmare, I'm sure they didn't have robust enough a test regime, on the other hand, given the scale of the system, I'm not sure what could be done.
Complaint 4 - it's failed it must be scraped. Not a recommendation I'd make until you'd like to blow another $250m.
So, what would I have done differently. I'd have looked for a company with experience in large insurance systems before and combine them with other consulting companies who really understand the US insurance system, rather than pulling in a monolithic systems integrator as they seem to have done. I'd have made sure the testing and beta program was planned and I've have begged for a staged roll out. And this is really the problem, I'm not sure that's possible.
We're launching a new product at the moment. We spent 6 months building stuff and testing it on lead, tame customers. We then opened it up to a beta program with a small number of users and took their feedback and the bugs they found and fixed those. Now we're ready to go into an open launch. We're bound to hit more roadblocks on the way and find more bugs. But that's the way people release software, and unfortunately it's a terrible model for a national healthcare based IT project.
There's a reason the rest of us have single payer systems.
