Google Fiber’s recent announcement about Austin, Texas and Provo, Utah has sparked a flurry of reactions from other service providers in the US. AT&T and Century made (somewhat dubious) assertions that they would match Google’s Gigabit. Comcast, Time Warner and Frontier on the other hand insisted that no one needs a Gigabit.

I wonder what IBM's CEO of 1943 would say to that, he who famously asserted, "I think there is a world market for maybe five computers." Making assumptions on the needs of end-users, especially in technology is a very good way of being ridiculed on social networks a few years down the line. A visionary leader is the one who creates the need rather than wait for it to happen.

Still, the CEO of Frontier Communications boldly stated that at Frontier, "we don't focus on megabits, we don't focus on gigabits, we focus on activities." She then added that 80% of her clients were using 6Mbps subscriptions even though they could get 20Mbps by paying more. She was implying, I think, that if the masses are content with 6Mbps then surely that’s a decent target for everyone.

A friend of mine recently riffed off this concept of determining the need or the supply on the basis of ‘what most people do’. Wouldn’t it be great, he argued, if everything was determined that way? The average speed on highways is 45mph? Let’s build cars that go to 60 tops! The median household income in America is $50k per year. Surely no one needs more than that. Let’s tax the rest! I’m sure the same CEOs who have been arguing that their customers didn’t need more than what they’re currently provided would suddenly argue that they need more than the median income…

No, that’s clearly a nonsensical way of looking at it. I understand why these business leaders are worried: Google is in effect creating a demand that they cannot currently supply for, and would have to invest considerable amounts of money to respond to. Clearly not a palatable scenario for them.

We only have to go back 15 years though to remember that the telecom market isn’t demand driven but supply driven. When the first companies to offer mobile phones came onto the market in the 90s, the demand was estimated at a few hundred thousand subscriptions at best in large countries. Until someone had a business model epiphany: the cost of providing mobile phones wasn’t in providing the devices or managing the service. It was in buying the spectrum, setting up the cell-sites and buying the equipment to run it. All fixed costs, albeit very high fixed costs. So selling cheaper phones to turn it into a mass market made sense. And it worked, probably even better than anticipated.

Guess what? The major cost component of offering a Gigabit is the deployment of a fiber network. Once that's in the ground, the incremental cost of offering 1Gbps as opposed to say 100Mbps comes down to slightly more expensive ports, slightly more expensive routers or CPEs and bandwidth provisioning. Various regulators in Europe and North America have estimated the latter cost to be in the 10s of cents per subscriber per month. Let’s assume Gigabit customers consume 5 times as much bandwidth as ADSL customers (which is far from being a given: FTTH operators I’ve spoken to report a usage for 100Mbps that’s double on average that of ADSL customers). The cost of provisioning the transit for a Gigabit would still likely be sub $1 per month per subscriber. Admittedly that might not be true everywhere and in places where the backbone providers are monopolistic ISPs will likely be charged more. Still, the amounts being discussed are ridiculously low both in light of the cost to deploy the infrastructure and, more importantly, in light of the revenue potential.

So even if we accept the idea that customers don’t need a Gigabit service today, does that mean it shouldn't be offered to them? Does that mean they wouldn't want to purchase it? A cursory examination of existing Gigabit offers throughout the world suggests the opposite. Operators around the world have been using Gigabit offerings very effectively in essentially two different ways:

• some, like Turkcell Superonline in Turkey or Chattanooga’s EPB are using it as a marketing argument, a strong and visible differentiation of their network compared to what the competition offers. Sure, these Gigabit offers are prohibitively priced, at least for residential customers (Superonline charges $1000 per month for a Gbps, and EPB charges around $300 for the same), and sell very little. But they allow them to market “the only network that supports a Gigabit ». The City of Chattanooga has even turned it into an economic attractiveness slogan, labeling itself "The Gig City";
• others are actually positioning Gigabit offers as core to their acquisition strategies. It’s the case of Hong-Kong Broadband Network in Hong-Kong with its famous Gigabit for $23/m offer, and of course it’s how Google is positioning its own product in Kansas City (Gigabit and associated services for $70/m). The effectiveness of any infrastructure business model is predicated on acquisition, a point that is not lost on either of these companies.

Considering that, if you’d gone through the pain of deploying a fiber to the home network, why wouldn’t you offer a Gigabit?

Who cares if the customers don’t need it as long as they want it?