Two reports, with more on the way, from US securities analyst Bernstein Research, the research arm of Sanford Bernstein & Co, shine a positive light on the fiber machinations of Google. Like most observers of Google in this matter, Faultline has sat mesmerized by Google fiber build outs, unable to put any meat on the bones and explain them away. Bernstein now attempts this.
The two reports, one out on May 6th, followed up quickly on May 7th, were shown to us at Faultline, and start with Bernstein employing a local research company in Kansas and going door to door to interview customers of the Kansas Fiber build out ‘ have they heard of it, do they subscribe, do they like it, would they go back to their incumbents, which options do they have and what do they pay etc..
The reports kick off saying ‘We don’t think Google Fiber is an experiment, a regulatory ploy, or just a bluff to keep incumbents in check or get them to upgrade their networks. We believe the potential to build a large, profitable business is one of the main motivations for Google. We acknowledge the challenge Google faces scaling up the business profitably and the risks from aggressive competitive response.’
The reports then proceed to impute a variety of likely US-wide penetrations given how happy consumers are with the service, and building a cost model for installing fiber across in each of 3,400 US cities in a manner that it says reflects the microeconomics of building fiber-to-the home network infrastructure. This suggests that in the long term, over a 4 to 5 year period, Google may suddenly decide to build out to as many as 177 million people in almost 67 million households along with 17 million businesses and concludes that, ‘ We believe there are material chances that Google could build a network passing 20 or 30 million US homes and small businesses in the next several years profitably.’
There are some worrying issues about the reports. For instance it claims that Google is currently capturing 75% of all homes it passes and to assume that going forwards across all of the US would be madness, since other operators will fight back, but Bernstein appears to have factored all this in. Under one part of the model it calculates what if Google ignored all the towns where Verizon’s FiOS is built out, and then shows how this would make each home more expensive to reach, but that way there would be less competition.
It then says that under the right conditions Google could achieve an internal rate of return of 12%, which it says is break even for cities where the average
cost per location passed is roughly $1,150 or below. When it avoids all FiOS cities this goes down to $700, and that means that there are far less places where it can build out and break even. But there is no attempt to see how much more freedom the core Google business might gain from such a move, and there is no attempt to justify an unprofitable build out, which may well suit Google.
Instead it opts for a middle way between the two build out figures, targeting a cost of around $900 per home passed saying that to reach 40 million homes passed, Google would have to build 1,100 cities, which it says is coincidentally the number of cities which have expressed to Google a desire to have Google Fiber.
The deployment approach the papers describe is also not knew, the idea of describing neighborhood (or fiberhoods) and then getting a deposit paid by enough people to make the deployment cash positive before it is even built. This has been tried in most parts of the world on initial broadband deployment. Using offers of free broadband to schools and parts of the local municipalities reeks of its free WiFi work of 5 and 6 years ago, and is a great way to get permissions to build out, but again it’s nothing new.
The truth is that if the operators were not tired older companies with cynical worn down help desk employees, many operators would achieve this type of result, both in terms of getting an extra almost 11% who like the company, but also in going the extra mile to charge less than the market rate, for a superior product.
The worry is that the survey was based only on 195 people, and we cannot take it seriously. There would be damage to Google in terms of brand, but also its employee mix, if it developed an operator service culture which pays it to scrimp and save on costs, reduce its customer care to a minimum in the style of AT&T and Comcast, and leave equipment in the ground until it is old and obsolete. Google is one of the companies brave enough to think outside the square, and rebuilding a better AT&T doesn’t feel like the right way for it to spend $20 billion of more of its cash.
However we will know what’s going on if Google suddenly embarks upon another 20 or so fiber build outs in the next few years, although by then it will have to discuss them as offering a significant impact on future earnings, something that it can really argue is not the case with the handful of build out it has gone in for so far.