I am Saurabh Adhikari, and I think someone should emphasize that the talk of XO-1 laptops from OLPC costing only $189, $200, or $219 including deployment is clear fantasy. Having been privy to one XO laptop procurement in India, let me give you the breakdown.
How much does this really cost?
One Laptop Per Child charges $229 per laptop (and upwards) for anything size purchase and shipment below 100,000 units and the price keeps going up if the size goes down. OLPC blocks the money for 4 to 5 months or longer and interest on that will be more like $15 for those who want to buy an OLPC. Add it all up and you end up with a very different number.
Here is the average cost per unit of an XO laptop deployment, for purchases below 100,000, averaged after acquisition:
$229 XO laptop price for order below 100,000
$70 Duty etc per unit
$20 Door-door shipping charges per unit
$10 Cost of creating LC per unit
$7 Cost of hedging against Forex. To learn about forex trading visit www.top10forex.net
$30 Solar chargers per unit
$23 Maintenance cost per unit @10% annual
$10 Training per unit
$20 Servers and installation
$30 Cost of procuring an order
$12 Back office support per unit
$461 - Total cost at ZERO margins
That is pretty close to the estimates in Australia that came to $490+. So I would say that it will be safer to advise the buyers the cost as $229+ taxes+ shipping+ LC costs+++ and it will be $410 landed + solar chargers and installation etc. So AT NO PROFIT, OLPC laptops cost over $410 in another country even though the manufacturers produce it for $200 and OLPC charges the overheads that are reasonable compared to market.
It will be great if every country discussed it on this forum to create a realistic budget for OLPC deployment.
While I can't comment on the specifics, this assessment seems reasonable, but lacking context (how much would a competing device cost? what is the relative non-economic costs such as the impact of pollution at the end of the life of the device? what is the annual operating cost?) it's just one part of analyzing costs against benefits.
[A pencil per child is much less expensive, burdened or otherwise, but a pencil can't offer the same range of utility, so I presuming we're comparing the xo with other learning-focused laptops.]
Even if burdens end up stripping OLPC of much of its cost advantage, that may not lessen the value of an xo if it outlasts the alternatives under pretty rough use conditions. Also, are repairs cost-effective and doable in situ or do they require special tools and skills? If an alternative computing device fails sooner, there's a variable in the cost of disposal to consider as well.
In effect, I'd like to convert that "first" cost into a "per year" cost to really understand whether the juice is worth the squeeze...
Many of the 1:1 laptop programs that I've been involved in are financed over a four to five year period so the payments are not all up front and this includes the cost of the device, software, service/repair and other.
I can't speak for how OLPC deployments are funded or underwritten but the yearly expense is not necessarily the same as a full term buy-out. Additionally, I've seen more recent emphasis on leasing of systems to further minimize costs and address obsolete inventory.
I don't believe there is a blank check approach to purchasing technology either. Every country or region has terms and conditions. How they vary I don't know.
I am in charge of Plan Ceibal, the only full 1:1 deployment in the world with more than 420.000 laptops and covering 1st year school to 2nd year highschool
By no way we paid the cost suggested in this article. Our TCO for 4 years is less than us$ 100 a year, including laptop, servers (every school), connectivity (99% of the children have internet in school), support, spares , etc
Miguel, thanks a lot for your comment. If I remember correctly you had previously (late 2009 I think) mentioned a 4-year TCO of $276 for Plan Ceibal. Is still figure still valid or have there been any changes based on your experiences since then?
Also, I think it is extremely hard to make comparisons between different countries as the question of what costs are and aren't associated with a project differ significantly (let alone different costs for parts and labor, pre-existing infrastructure, etc.)
For example I believe that in Uruguay's case a lot of the costs related to setting up and providing Internet connectivity are handled by Antel, therefore not being directly attributed to Plan Ceibal, or did I misunderstand that?
Similarly in Paraguay Personal is sponsoring the first two years of Internet connectivity at the schools but last I checked it was unclear who would pay the bills after that initial period.
Anyway, thanks again for your input, I think this is an important issue which needs to be discussed more openly. :-)
It's good to see you posting here, Miguel. You would be doing a real service if you made available the TCO details in Uruguay. Then people could see exactly what is included and what not and how much specifically each of the costs are in a large national deployment. I realize you and your CEIBAL team are busy with lots of other things, but making that information available would be quite helpful.
Only $490?
We have seen massively larger TCOs for the laptops
$2700
http://www.olpcnews.com/sales_talk/price/5_year_tco_computers_in_schools.html
$490 but on wrong data
http://www.olpcnews.com/sales_talk/price/xo_laptop_total_cost_of_ownership.html
$972
http://www.olpcnews.com/sales_talk/price/the_real_cost_of_the.html
So $490 is at the very low end of the scale.
TCO is one of the more devious metrics the industry has come up with in the recent past. The metric itself is on target, if done correctly. Keyword is "total".
Context is always important. So is the scope. At what point do you draw the line and say the "total" is limited to certain items? Do we include the wall charger, but not the electricity? Do we include the salary of the man who runs the generator? The usual rule of thumb is to include the costs incurred by the deployment agency.
Then there is the issue of what was already available before the deployment. For instance, when Microsoft does TCO studies against Linux, they include training costs for Linux, but not for Windows, saying that people already know Windows. That's an important assumption.
While TCO figures are important, keep the context and breakdown of costs in mind as well. Otherwise its really sticker price we are talking about, and its no different that $249 I paid for a Asus EeePC for my mother on Woot.com compared to Saurabh's $229. I did not include my mom's lunch/dinner expenses while training her to use the netbook :-)
Having said that, I think it is important to take lead from Saurabh and put together such detailed cost breakdowns that other deployments may have (maybe on the OLPC wiki?). It may not help with reducing the $229 figure, but may assist in cutting some other costs by changing purchase/acquisition strategies.
Sameer -- Discussing satellite costs to help share best practices from different deployments is a great idea. Uruguay has led the way with a fairly detailed breakdown over four years. It wuold be good to compile the same for other deployments.
Saurabh, could you provide a few sentences of context for each of the costs you mention here?
I'm not surprised at your numbers. Solar powered deployments are much more expensive than grid-powered ones such as Uruguay. It would be great to have a public repository of the expenses incurred by previous deployments. Feel free to use the OLPC Wiki to store this info.
There already is a spreadsheet which can be used to estimate hardware and power costs at: http://wiki.laptop.org/go/Deployment_Guide/Workbook (spreadsheet at http://wiki.laptop.org/images/b/be/Deployment_Workbook_v14.xls ). It doesn't try to estimate teacher training and program management overhead costs.
I heard that the OLPC Association was working on a simpler version of this. As another commenter already pointed out, these cost estimates are highly country specific.