Why Google Fiber Has Stumbled, the Lessons to Learn & Why Moving to Fixed Wireless Won't Solve the Problem

From its inception, Google Fiber has owned the gigabit narrative in the media and took on a mythic aura of providing gigabit nirvana.

In the process, it has benefited the entire broadband ecosystem in many important ways, one of which is that it added fuel to the community broadband phenomena.

Let's look at the issues Google Fiber Inc. tapped into and then consider how it may have taken the wrong, albeit traditional, approach.

We know the "digital divide" is real and that it's getting wider. This is true for individuals but also for towns and cities that are seeking the economic benefits that come with the improved infrastructure. Google Fiber helped drive awareness of the promise of gigabit broadband and the dire realization that if you don't have it, the survival of your community cannot be assured: The failure to have the right infrastructure is likely to lead to a steady decline of the community as the young move out and few new businesses move in.

On a lighter note, it was also able to benefit from reduced media scrutiny compared to traditional network service providers with the concept of "up to" -- as in "up to a gigabit" -- disappearing from the narrative. For example, GPON, being the most common FTTH architecture, provides 2.5 Gbit/s downstream to up to 32, 64 or more homes served per gigabit port. Thus, you can get "up to" gigabit performance (maybe symmetric if you're lucky) but only if you're one of a few active homes on the network.

Using an Active Ethernet connection, each port is gigabit-enabled so the technology choice is an important one in terms of what can and will be delivered. Either way, incumbents will no longer have to whisper "up to" in their gigabit advertisements due to Google Fiber: What's good for Google is good for everyone in this case.

However, in all cases, the pipe from the Central Office to the Cloud -- the middle mile -- will be highly oversubscribed. Statistical multiplexing does wonders for business models.

Google Fiber has also made "cherry picking" acceptable and this is important. Regulatory bodies like the FCC in the US loathe cherry picking -- aka "red lining" -- since it implies FTTR (Fiber-to-the-Rich) or, in reality, FTTRE (Fiber-to-the-Rich-Enough). To solve this semantic problem, let's call this approach "Smart Build."

Smart build is the smart thing to do and it is fair and equitable for the entire ecosystem. All that "smart build" means is to let the over-builder, such as Google Fiber, deploy fiber down the street based on demand density. It's quite simple. Assume 'Street A' has eight of ten homes wanting the service and the broadly similar 'Street B' has four of ten homes wanting the service. Let them deploy fiber first down Street A then Street B. If 'Street C' doesn't have any home wanting the service, don't make the over builder deploy fiber down Street C. For challenging locations, fixed wireless solutions provide an alternative where the cost of fiber would be prohibitive. In this way, "redlining" is avoided when dealing with municipalities, as everybody is provided with a solution.

Smart build is important since the fiber company needs to get to financial breakeven as fast as possible to ensure the success of their over-build. It can be the difference between finding a successful business model and simply not being able to deploy, and thus benefit the community. In the US, it can cost $20,000 (and potentially much higher) per mile to deploy fiber-optic cable down the street from pole-to-pole (aerial) and lots more if you want to bury them. Smart build lets the over-builder get the most paying customers as fast as possible. Over time you can "smartly" expand to more and more neighborhoods.

So why has Google Fiber stumbled and what lessons can we learn thus far?
The first lesson learned is that the access network is, to be polite, a "challenge." It's always been the most difficult aspect of a service provider's network, Google included. The "cloud" or data centers are easy compared with the access network. Data centers have nice raised floors and advanced HVAC systems: Central offices have neither. Access network owners also have to contend with miles of wires, many strung pole-to-pole, radon gas in underground vaults, lighting strikes, flooding and "rodent chews."

Then they have onerous local, state and federal regulations to address. The access network is not an electrical engineering problem; it's a civil engineering problem. Moore's Law does not apply. If it did, the same evolution we've seen in performance in transistors would see a bucket truck fit in your hands. Sadly, deploying fiber networks also provides few economies of scale once the early stage learning effects have been resolved. Innovation should help us with some new ways of doing the "old things" and therefore improve the business case but we currently don’t have many ways of reducing this cost per mile.

The second lesson is that Google Fiber proved that today's broadband model is broken. It's broken primarily due to the tight technical and business coupling of the services offered (voice, video and data) to the physical network.

To illustrate this, imagine that your town only had FedEx and UPS wanted to deliver packages. If they acted in today's incumbent-centric world, they would have to build their own roads to every house they wanted to deliver to. If UPS's new roads impacted FedEx's existing roads, UPS would have to pay FedEx to alter their roads before they could start building theirs (Make-Ready). To make this model even more inane, if you want to switch roads, you have to go all the way to nearest major city for the closest intersection. The current model is also troubling because whoever builds the roads gets complete control over who gets to drive on them.

Net neutrality only applies to Internet access not to the video and voice services. In the future, Internet access will be just one service and cloud companies will want alternative ways to access their customers. Thus, if a cloud company (e.g., Amazon, Microsoft, Facebook, et al.) wanted to serve subscribers directly they too would have to build their own roads.

However, any municipality/development/smart city wanting to ensure choice of both service providers and services for residential and business customers, needs to have control of the fiber asset rather than surrendering this to someone else. This then ensures competition and choice for end users -- which drives innovation and also the take-up in an area.

Google Fiber has certainly learned the first lesson. Yet, it doesn't appear to have learned the second one. It failed, or stalled, because it used yesterday's business model. This business model is exactly the FedEx-UPS model described above. Change FedEx to any telco or cable incumbent and change UPS to Google Fiber.

Google Fiber's parent company (Alphabet) benefits with gigabit broadband regardless of who deploys the last mile. Alphabet sells advertisements, and a 4K-Video ad will generate more revenue than a 1K ad. Why it would want to be the last mile provider in select disparate geographies is an enigma: If its strategy was to force incumbents to act, it succeeded. However, no action means no reaction: Incumbents have already indicated a slowdown in gigabit rollouts.

So, is Google Fiber just another over-builder with a cool brand name? It faced the same initial challenges and decisions as any other. Our belief is Google Fiber would have preferred to offer Internet access only, as the clear trend is towards streaming. However, it quickly and accurately determined that there aren't enough cord-cutters now, or in the short to medium term, to reach break even.

The biggest barriers to massive cord cutting are demographics and large segments of the subscriber base still want a traditional TV channel line-up. Thus, Google Fiber moved to a "Double Play" of video and data. Voice was initially off the table for two reasons. First, it believed the traditional "home phone" was going away and second, they wanted to avoid the mountain of regulations associated with legacy voice services. Here too, demographics demand these voice services and Google Fiber began the climb up the mountain of regulations.

It didn't take them long to realize how unprofitable and time consuming these traditional video services can be. Kings have power and content truly is king. Here they get the double whammy. Video content costs continue to rise each year and cord cutting is a real phenomenon. So each year they need to amortize the rising video costs over an ever decreasing number of subscribers.

The narrative is now that fixed wireless will save the day. It won't. Even if Google Fiber is a real long-term business for Alphabet, changing the physical layer won't fix the broken business model. The danger to the ecosystem is that Google does "well enough" with fixed wireless regarding effective and sustainable throughput and take rates. This would lead the masses astray by solving the wrong problem.

There is tremendous amount of innovation occurring in the mobile market, much being driven by the R&D around 5G. Keep in mind that basestations of all types, including WiFi access points, work best with fiber backhaul. All this is great and both mobile and fixed wireless solutions need to play an integral part in ensuring broadband for all. But no technology can fix a broken business model.

Let's consider the problem with the economic model of the fiber over-builder -- Google Fiber in this case. The traditional, closed network architecture where services are tightly coupled with the network infrastructure (as in delivered by the same entity who owns the physical network) is unsustainable in the 21st century. No one, including Alphabet/Google, has enough capital to profitably "wire up" enough communities to make a difference on a national scale. But it's Google Fiber's insistence on owning the network and on offering the triple-play consumer services that has perhaps doomed them; not their choice of physical layer topologies.

The skills and expertise required to deploy and operate a vast fiber access network are drastically different than those required to offer services at large scale. Trying to do both, particularly as a newcomer, is a recipe for failure.

In the next article we will consider the different operator approaches that municipalities, developments and smart cities will come across. Some are imposed on them: Some they are mistakenly choosing. As with Google Fiber, the right business strategy and model require careful consideration and an understanding of the pitfalls of wrong choices.

— Greg Whelan, Principal, Greywale Insights and Richard Jones, Chief Commercial Officer and Co-Founder, Ventura Next