|Google reveals details of Mountain View Wi-Fi service|
| 3:50 pm on Jun 22, 2006 (gmt 0)|
|Google does not plan to use ads to pay for the free wireless Internet service it's offering in its hometown of Mountain View, Calif., and there's no secret plan to monetize the service, a Google Wi-Fi product manager said Wednesday. |
"The reason it is free is because...we want to get a lot of people on it," Larry Alder of Google said during a panel discussion on wireless projects in cities at the Supernova 2006 conference here.
The service, which is fully deployed but not yet available to all Mountain View residents, is a test bed that will help Google understand the technology, Alder said.
Google reveals details Mountain View Wi-Fi service [news.com.com]
| 3:57 pm on Jun 22, 2006 (gmt 0)|
A trial run for their future plans, maybe?
| 4:49 pm on Jun 22, 2006 (gmt 0)|
It's obvious what they are doing here. Installing a mind control program onto the computers of all who use this free service. There's just no other explanation.
| 5:06 pm on Jun 22, 2006 (gmt 0)|
There are plenty of explainations. With such a large market share on search, getting people online will benefit them immensely and ultimately lead people to use their other services which do have ads. Over all its in their interest to put as many heads online as possible.
Not to mention this adds tons of branding and will show the world their capability to provide reliable free wi-fi. If this test-run works, it'll open up contracts in hundreds of cities and many other countries--where they may begin to serve ads eventually. This is just a small beta and the big execs don't care about the measly 10 or 20 million that comes from monetizing ads from one city, they care about making sure it works and establishing a strong image for the politicians and others who need convincing before getting on board.
| 6:18 pm on Jun 22, 2006 (gmt 0)|
|Google has been working with ham radio operators on sharing the spectrum the Wi-Fi network uses |
This is interesting... Wouldn't Google need to work with the FCC to accomplish this?
P.S. The article itself didn't waste anytime with Godwin's Law.
| 1:00 am on Jun 23, 2006 (gmt 0)|
I'm sorry, but this is just another one of those - let's throw it at the wall and see what happens tests.
|Google has hung about 350 nodes on city light poles, and they'll serve about 70,000 people in 12 square miles, he said. |
Let's see, they probably spent around $2,000 per node (on the CHEAP side) - that's $700,000 in capital.
How about ongoing O&M?
350 - T1 lines at $150 / month = $52,500 / month
NOC / Service Costs = $10,000 / month
Maint. (5% failure rate) = $3,000 / month
Depreciation (5 year equipment life) = $12,000 / month
Pole attachment fees ($10 / point / month) = $3,500 / month.
So we are talking roughly $80,000 per month and this assumes Earthlink kicks in a lot. So how many homes is this? Let's say they are passing 70,000 peeps at 4 peep per home or 17,500 homes. Let's say their saturation rate is 30%, that's roughly 5,000 users.
That works out to roughly $16 / user per month to recover their cost. Bump that up by 25% to pay Earthlink and we're talking an even $20 / month / user in revenues (at least).
Do you click on that many ads a month? If you think my numbers are way off, think about this. Cable refuses to lower their prices to beat DSL. Why? Because at $30 / month Verizon is practially giving the stuff away.
And this assumes Earthlink will set up the NOC for free, be the ISP, it doesn't include training costs...
| 3:32 pm on Jun 23, 2006 (gmt 0)|
Good workup BillyS, but you also didn't include R&D, salaries, insurance, legal, PR and advertising costs.
| 4:36 pm on Jun 23, 2006 (gmt 0)|
There are several other things to consider that counter the idea that costs of Google's Wi-Fi venture would be offset but increased revenue from the new users that get online.
1) If you assumed that all of Google's $2B of profit (or whatever it is) came from the 300 million x 72% = 216 million U.S. residents with broadband, that would amount to only $9 per person even with those ridiculous assumptions.
2) Broadband penetration in the US is already 72% and probably higher than that in Mountain View, so even if this got everyone in town using broadband it would still be only be a 28% increase.
3) The 28% of people who do not have broadband tend to have lower incomes and less education. Therefore, their value to Google is less per person than the people already online.
So, the bottom line is that this will not simply pay for itself by new broadband users clicking ads. But, who knows what this could lead to in the future. Maybe the way they get users (and politicians) to jump on board is by offering it for free now and once they prove the value of the service then they find a way to monatize it.
| 3:09 am on Jun 26, 2006 (gmt 0)|
I'm not disagreeing with your argument that Google might have difficulty recovering their costs, but I do think your numbers are way off:
|Let's see, they probably spent around $2,000 per node (on the CHEAP side) - that's $700,000 in capital. |
I've done quite a bit of wireless installations, even ones that cover large areas (though not metro-sized yet). I would say that $200 would be closer to the cheap side, and $2,000 would be more on the "oh my god how could you possibly spend that much on a node" side.
|350 - T1 lines at $150 / month = $52,500 / month |
Are you suggesting that they run a T1 line to each node? Why would they do that? Systems like this usually have bandwidth at a central location, or spread out across a few locations. The nodes then communicate with each other to share bandwidth. It's likely that they have a few fiber lines feeding a or a couple nodes that all get bandwidth from the same place. Their cost this bandwidth is probably much closer to $5,000/month than to $52,500/month (depending on the speeds they're offering).
| 10:52 am on Jun 30, 2006 (gmt 0)|
|I've done quite a bit of wireless installations, even ones that cover large areas (though not metro-sized yet). I would say that $200 would be closer to the cheap side, and $2,000 would be more on the "oh my god how could you possibly spend that much on a node" side. |
Are we talking about the same thing? These devices need to deal with rain, snow (in some areas of the country), heat, humidity, power surges and lightning strikes. You also have to get power to each device. They have to be mounted so they don't blow away in storms. The installation time alone will cost you $200 - $500 because you can't have just anyone install this stuff.
You can't get the truck to the street for $200, I'm sorry. (And I've met with three manufacturers of this equipment, I know too...)
| 11:06 am on Jun 30, 2006 (gmt 0)|
|Are you suggesting that they run a T1 line to each node? Why would they do that? Systems like this usually have bandwidth at a central location, or spread out across a few locations. |
I agree with you - it can be done the way you're suggesting. Then in your business case, all you need to do is add the cost of the fiber you're going to run everywhere, buy some land to put a centralized location in, build your centralized location (including the equipment you need inside the building) and add T3 at $10,000 per month.
You're just trading O&M for capital.