US census ditches handhelds for clipboards
Back to pen and paper
Regcast training : Hyper-V 3.0, VM high availability and disaster recovery
The US is giving up plans to issue researchers with handheld computers for their 2010 census, as changing specifications have driven the cost of each one beyond $8,500.
The next census is scheduled for 2010, and the US Census Bureau hired Harris Corp to create handsets suitable for the half-million researchers who have to knock on doors collecting the data.
The project was supposed to cost $600m, or about $1,200 per handset, but the Washington Post reports that changing requirements and badly estimated costs have driven the price up to $1.3bn for only 151,000 devices.
The handsets do have GPS and wireless connectivity to allow dynamic collection of the collected data, but that's hardly sufficient to account for a per-unit cost of $8,609.
The Census Bureau blames Harris Corp for misunderstanding its requirements and underestimating the complexity of the project. Harris blames the Census Bureau for failing to provide proper specifications and making 417 separate changes to the project during its lifetime.
Various groups, including the Government Accountability Office and House Appropriations Committee, have been picking over the problem and concluded that it was lack of communication on both sides that led to the ballooning costs. Harris had failed to take into account the cost of supporting users, as well as the complexity of the task the devices are expected to perform.
The last US census, in 2000, is estimated to have cost about $4.5bn (about $16 per person), but 2010 is going to break all records with an overall cost of $14.5bn.
For comparison, the last UK census, in 2001, came in at £254m – about $10 per person. It remains to be seen how much Deutsche Telecom can spend on the UK 2011 census. ®
COMMENTS
$9,000 a pop, a bargin
Let's see 151,000 PDAs with a custom application to record census details, a web page won't do because the network signal can't be guaranteed, so you need to store the completed survey forms locally until a connection has been made, and then download them. This could be days in the US.
So the design will set you back a few million, plus a few more for a bunch of coders in Bangalore to write the application, the a few million to test it, a few million for a bunch of coders in the US to rewrite it so that it works, another few million to test it again. Then even more millions for the change requests, then you have to set up your unique remedy help desk, had to be government security rules won't allow you to share an existing one, to support the users and a software support team, so that's some more millions. Then you need to train 151,000 users to user the application, so that's a nationwide training program, so that's many more millions, don't assume that because you know how to use windows mobile the Joe or Jane next door does. Then there's rollout costs, profit margin, and contingency so there you go $500,000,000, allow some scope creep, and the the usual double the number you first thought of and add 50% for any government contract, and Bob's your uncle, $1.5 Billion. You aren't buying a simple excel spreadsheet you know. So how much was it then.
By the way it would be really interesting to know how the break down that $8,609. It can't all be for just the device, can it?
ok 417 changes
WHat the hell did you think was going to happen. THey were going to say OK we will do it for the same price. Would love to know what the original specs were and then see those 417 changes that were requested.
erm
I got a PDA phone, with several flavours of wireless, cellular tech, GPS, etc for free on my contract. You can buy them off the shelf for less than US$500 and the discount for 151,000 would be quite good.
IT infrastructure monitoring strategies
Requirements Checklist for Choosing a Cloud Backup and Recovery Service Provider
Cloud based data management
Enabling efficient data center monitoring
Agentless Backup is Not a Myth
