I understand data mining only from a very high level. I'd like to see a description of how it works as a revenue model, including how much a third party might be willing to pay for this or that snippet of information.

Also, does data mining's revenue potential vary by demographic quality (e.g., income)? For example, are some third parties more interested in -- and thus more willing to pay to learn about -- people who have a lot of money to spend?

yeah, a simple example would be Amazon showing which products sell in your town, they probably sell data down to zipcode.  i'm not exactly sure what google plans to do with the data but other isp's are already mining the connection and selling stats. google is far more advanced with forensic data mining and has the channels in place to sell data so will probably go much further than what isp's make.  given that google already makes billions just from ad clicks and data mining search data, they'll likely find ways to make magnitudes more now that they can mine the entire connection.

it gets questionable when isp's peek into the packets ('deep packet inspection')... ie, they can easily know what sites are visited but could peek inside the packets to see what product you're interested in or which type of porn you like.  it won't surprise me if google pushes those boundaries further than any isp and it would get worse if they tie to a personal id or ip address.  privacy groups claim that google already records web searches with an ip address logged.  of course the feds love logged data so it will be hard to fight this.  the feds could turn out to be google's largest customer if not already.

I would expect them to say that, as there's no need to deflate the price. And, I don't think either of those things is untrue. Google is a business. I'm sure they want to profit from the fiber initiative. I think your data mining point is sound. But I also expect they are figuring out the business model as they go and prepared to absorb some short-term losses, regardless of how they handle their accounting. If there wasn't an experimental aspect to this, why wouldn't they roll out everywhere at once? Is Google committed to getting in the business of being an ISP?

I don't think they are committed, but it appears to be the intent.  The KC rollout is probably both, an R&D experiment and a test on how to make money being an ISP that makes most revenue by data mining rather than by most revenue coming from the subscription, as traditional ISPs work.  Google has also expressed concern on the direction of ISPs capping data and wants to prove it isn't necessary and in fact will slow down innovation of Inet services..

I could be wrong, but as I understand, the manner in which Google is going about putting the lines in is really unprecidented.  With access to right of ways and poles on the cheap or for almost nothing, they don't have to follow the much more expensive cabel co model.  Google has existing lines and can build out a significant start to the service within a year.  That is not something a new cable company could or would do, frankly.  In the past, investing in laying cable and fiber was like a giant game of Risk.  Once a cable co plants its flag in a community, cable co #2 is going to have to spend a lot of money to overtake their original investment, making it a less attractive place due to decreased revenue from competition.  Google is entering enemy territory with cheaper deployment strategies, lack of interest in television sets, and a superior product by speed.  All huge advantages.

Much has been stated about the ISP price points like they are charging exactly the right amount based on their investment.  I would contend that they charge what they charge because they can charge it.  Obviously there are investments and operating expenses to account for, but where is the competition?  Hasn't AT&T U-Verse done very well in the markets they have reached with high speed?  I've heard that TWC prices go down when they enter a market.  But even AT&T who is largely the second to market is treating customers hostily and throttling speeds and enacting usage caps.  I think it will be very refreshing to have single fee, a single option, and for that option to be the best quality.  That is what made Costco so successful.  It sounds quite obvious what you will get with Google's service (remains to be seen, but still), and I think the simplicity of not having to read through the code of RoadRunner, RoadRunner Turbo, etc, and why you need to buy PowerBoost to get the speed you are already paying for, will be worth it.

One big wild card is the amount of fiber that Google plans to lay versus how much it will lease from other operators. If Google leases a lot, then the operators they buy from would benefit because Google is subsidizing their infrastructure upgrades. For example, if TWC lays a bunch of fiber to meet Google's needs, it then has more infrastructure over which to offer its own, competing services.

A similar scenario has been playing out for years in cellular: Telcos, MSOs, etc. build fiber to cell sites because carriers are replacing T1 backhaul. Every piece of fiber laid to a cell site means the operator now has fiber infrastructure closer to the homes and businesses along that route. 

Or leasing fiber to Google.

Never mind..... apparently big yellow reels of conduit are actually replacement gas lines

http://www.pressdemocrat.com/article/20110607/ARTICLES/110609527?p=all&tc=pgall

Looks like 700 lucky people in Sonoma County, CA will get residential fiber before anyone in KC.  The plan is way more affordable than any competition for the speed; $40 for 100 Mbps, and $70 for 1 Gbps + a few phone lines.  If Google can come in at this level in KC, or even cheaper, I think it will benefit every consumer in the metro, regardless of if they jump ship or not.  It will be interesting to see.  It does not appear that the scale of this trial is anything close to what will be offered next year.  Google did use Sonic to manage its fiber network at Stanford University, so it is possible that this is a pricing trial in preparation for KC.

Am hoping Google can price 1Gbps at $50/month and $20/month for 50Mbps. That would truly be a game changer. Could be doable if they data mine the connection.

Some possible uses for super high speed internet... basically nextgen cloud services and virtual desktops...

http://www.heraldonline.com/2011/06/30/3185511/internet-forecast-cloudy-with.html

The Google fiber at Stanford is free the first year.  Doesn't mean it will be in KC but that would be a pretty awesome way to enter the market.  They are using Sonic Net to manage the network... they have their own fiber rollout and it's $70/month.

http://www.kansascity.com/2011/07/20/30 ... rvice.html

cool. my son's off to Stanford this fall to study nuclear physics. He'll be happy.

Did he get a perfect SAT?

Google lit its fiber test-bed in the neighborhood around Stanford University today. ...

It's only getting 150Mbps instead of 1Gbps but I'll take it.  I wonder if they're using the same equipment/architecture that is planned for KC.  It's free the first year but there are pricing rumors of $40 for 50Mbps or $70 for 100Mbps+.  I'd go for that.

Google helps makes KC an emerging innovation center. 



http://www.bizjournals.com/kansascity/news/2011/08/22/google-helps-make-kansas-city-an.html?ed=2011-08-22&s=article_du&ana=e_du_pub

Nice enough article.  I was disappointed to see the Star discount the positive assessment in their coverage.  Only in Kansas City.

Well the WSJ article is a bit misleading.  KC doesn't have 2900 tech companies, it has 2900 companies that have an IT shop or are a tech company.  Still not too bad.  KC ranks 18th per cap in tech employment, which is not bad for a metro ranked about 30th in  pops.  The original source also says KC ranks 8th for telcom employment.