When Google announced last summer it was building a fiber network in Kansas City, it was easy to believe that Google Fiber was just an experiment.
But Google has recently announced it's building a similar network in Austin, Texas, and Tuesday night the city council in Provo, Utah voted to sell its $39 million fiber optic network to Google, making Provo the third Google Fiber city.
Some people think that this fiber thing is the beginning of one of Google's biggest undertakings yet, that the company--after getting into the mobile OS and phone business, and maybe the automotive industry--may actually be thinking about becoming a nationwide ISP.
The great disruptor
Google plans to offer consumers gigabyte-fast Internet access for $70 per month, with the option of buying TV service (200 HD channels) for another $50. In Kansas City, you can get 5 mbps broadband for free if you pay a $300 "construction fee" or $25/month. That option will be available in Austin and Provo too, but the fee is only $30.
Google Fiber Internet service is way faster than what we're used to. The gigabit per second Internet service Google will sell is about 56 times faster than the average home broadband we have today (according to Ookla/Speedtest.net stats).
The Google Fiber service is also a far better value. Americans pay an average of $4.25 per month per megabit/second of throughput speed. Google will be charging about 7 cents per megabit/second of speed.
The fastest broadband in the U.S. today is Verizon's FiOS, with a top download speed of 300 mbps. That's probably all the speed most people need, but it's not cheap at $205 a month. That's about 68 cents per megabit/second of speed compared to Google Fiber's 7 cents.
The ISP market Google would enter is ripe for disruption. It's characterized by inflated prices, big, slow-moving incumbents, old technology, and not much competition. Google is good at rethinking the way a service is sold and delivered, then offering a better product at a lower price (sometimes free) than what the incumbents are offering. That's what the company would try to do as an ISP.
The cost of doubling down
But it's going to cost Google a lot to offer the service on a large scale.
One analyst firm estimates that were Google to build and operate a fiber network that passed 20 million homes (not connect, just bring the fiber within reach) for five years, the cost would be between $10-$15 billion. The firm points out that the estimate assumes Google will be able to find new markets where it can reach the same number of homes with the same footage of fiber (or less) as it did in Kansas City.
And even after that big spend and time commitment, Google would be only a mid-size operator. Charter passes 12 million homes today, Time Warner Cable 20 million, and Comcast 53 million.
Then there's the process of going city-by-city and suburb-by-suburb trying to get permission to dig up the streets to drop the fiber and sell the services. This might remind some Googlers of the headaches of getting the people of San Francisco to sign on to the free Wi-Fi plan back in 2005.
Google already gave up the telephone service part of its planned triple-play offering because satisfying the regulations and tax rules involved in selling that service was just too much.
Then there's the political aspect. If the experience of the people who built the UTOPIA municipal fiber network in Utah is any guide, Google may face all manner of obstructionist maneuverings by big cable and telco ISP lawyers and lobbyists in city halls, county courthouses, and statehouses. At every turn, they would attempt to slow down or increase the costs of the fiber build.
Perhaps most seriously, will people know what to do with gigabit-per-second speed? It's likely that Google will be playing the role of "first mover" by first planting the fiber then hoping that developers design Web apps that fully exploit its speed.
And as ARS Technica points out, the actual experience of using Google Fiber Internet service today isn't much different from regular old copper-based broadband service. That's because most of the infrastructure that makes up the Internet is still geared for slower broadband speeds. Just because the last mile, the on-ramp, is superfast doesn't mean the superhighway is all that super yet.
Surely Google knows all this, so why would it want to do it? There must be some pretty important reasons, and I have an idea of what they might be.
Google has always been an over-the-top player, meaning that it has relied on broadband ISPs like Comcast and AT&T to reach its customers with its services. In the beginning, that service was search. Then came advertising. Then came other services like Gmail and documents. Now it's video and apps and music and social networking and a hundred other things. And all of those things, in one way or another, are designed to keep people on Google properties and bolster the company's ad business.
But if even one big ISP suddenly banned Google's services, Google's business model would fail. No ISP would be crazy enough to do that, but the ISPs could do other things short of banning to make it more expensive for Google to bring its services to consumers, depending on what the regulatory environment would allow. They could limit bandwidth for certain types of Google services, or charge a tariff for delivering large files like movies. Larry and Sergey no doubt see it as a potential threat to the business, and one that could materialize in a number of different ways.
So naturally, Google would like to take more control of its own destiny. It would like to take ownership of the network that connects it to its customers. That's what this fiber business is all about.
Speed inspires usage
But it's more than that. Google isn't selling fiber broadband that's just "competitive" with existing services brought to you by Verizon, AT&T, and Comcast. It's way faster.
As we've seen clearly in the wireless world, when you give people more speed, they tend to both use more online services and want more speed. With all that fiber speed, people might watch more YouTube videos (and view more ads), do more searches (and view more ads), and read and send more Gmail (and view more ads).
Who knows, Google might try to launch services inside Google+ that fully exploit the fast fiber connection; this might give the fledgling social network a much needed membership bump. Gathering more of the personal and preference data that's the lifeblood of social networks could help Google bring more ad-targeting value to its advertising partners. Facebook will not kill Google, as some have said, but Google needs a viable social network.
A brand new channel
Google has a much bigger appetite for television than Google TV can satisfy. The biggest problem with that product is that is relied on cable networks for the content. With Google Fiber, Google will own the pipe that brings in the Internet service and the TV service. This no doubt sets the Google people's imaginations in motion.
Big advertisers still spend much more on TV advertising than they do on Web ads. Google may experiment with new ways to target television ads at specific demographics, and then measure the results. Right now, TV networks can only hope to match ads with the demographics they believe are watching the shows. Google may find a way to target ads at viewers in real time, as web advertisers have learned to do using cookies.
In for the long haul?
Google has been known to bail on projects and products that aren't working out. Is it possible that Google would double down on the fiber project for a few years, then get bogged down in the realities of the business (see above), then wash its hands of it? That's precisely what happened when Google tried to build a free Wi-Fi network in San Francisco back in 2005.
And that project was small potatoes compared to what Google's trying now. Expanding Google Fiber nationwide, if that's what the company chooses to do, is going to entail more trench warfare than any project Google has ever committed to before. It's going to take a lot of single-minded determination, starting at the very top of the chain, with Larry and Sergey.
Google Fiber is not just some new caprice thought up by the Google founders; they have been thinking about this problem for a long time. After the San Francisco experience, Google invested in Spanish company Fon, whose software turns Wi-Fi routers into a hotspot, allowing you to share some of your bandwidth with other Fon users. In 2009, Google worked with Tropos Networks to provide free Wi-Fi to 118,000 people in Mountain View, California. In January of this year the company turned on free Wi-Fi for people in the Chelsea neighborhood in New York City.
One pundit pointed out, wisely, that Larry and Sergey are big thinkers; being remembered as just the guys who made Gmail and Google Docs won't be enough for them. Being remembered as the guys who made the first self-driving car, then the guys who made access to the Web faster and better by using fiber to bust the cable/telco ISP duopoly--they'd probably like the sound of that.