Friday, July 27, 2012

Will Google Fiber Economics Work?

“There’s no sense selling a product at a loss,” says Google CFO Patrick Pichette. “But it’s not only about profits, it’s about changing the access costs,” Pichette also has said. 


Assuming you believe that Google is serious about "making money" selling symmetrical 1-Gbps connections at $70 a month, and video service starting at $50 a month, the issue is what Google can do that cable operators or telcos have not been able to do to get capital investment and construction and other costs down to a point where retail prices at such levels still turn a profit.


Sonic.net might agree that it is possible, under some circumstances, to offer very high speed broadband access at shocking prices. Sonic.net already offers consumers 1-Gbps service for $70 a month. But Sonic.net also notes that its construction cost is about $500 for each home passed. And since Sonic.net gets about 33 percent take rates, the effective network cost for each customers is about $1500.


If Google gets similar economics for the network, and few observers are likely to think Google has found some magical way to avoid the actual costs of installing cabling, and also gets about 33 percent penetration, it should be possible to make money at $70 a month for a 1-Gbps service. 


Of course, operating costs will have to be kept in check as well, and that is an area of potential friction for Google, which arguably will not have the infrastructure a service provider might normally be expected to support. 


On the other hand, significant portion of the cost of delivering service is not the actual backbone network, but the drop network and then customer premises equipment. Google's "$300 connection fee" suggests the cost of activating a drop is that amount.


Then there is the cost of the customer premises equipment. Some might argue Google has a cost advantage in that area. To be sure, building custom boxes, in low volume, is not generally the key to low costs. 


But perhaps Google has built a really simple box, using its new Motorola expertise, that dramatically lowers CPE investment. On the other hand, some observers will note that Google actually supplies three separate boxes, plus a Nexus 7 tablet, for a customer buying the 1-Gbps plus video entertainment service.


Some will argue it is hard to see significant cost savings when using a discrete approach such as Google is employing, but perhaps that saves significant money. 


Others might argue that Google will save on marketing costs or other overhead, and that might be a more-reasonable argument. Sonic.net probably does not spend as much money on marketing as Comcast, Time Warner Cable or Verizon does. Google might be able to do as well as Sonic.net

No comments:

"Tokens" are the New "FLOPS," "MIPS" or "Gbps"

Modern computing has some virtually-universal reference metrics. For Gemini 1.5 and other large language models, tokens are a basic measure...