It all looked so certain a couple of years ago. Google Fiber, the company’s gigabit consumer fiber network, had been extended from its initial deployment in the Kansas City metropolitan area into Austin, Texas, and Provo, Utah. In 2014 and 2015 this was rapidly followed by announcements and rollouts in Atlanta, Charlotte, Raleigh–Durham, Nashville and Salt Lake City. Another 15 cities were planned and announced through to June 2016. Google (by this time Alphabet) then when on to acquire high-speed wireless specialist Webpass in October 2016. Within three years Google had a direct-to-consumer access network with over 400,000 subscribers across eight major US cities. What chairman Eric Schmidt had described as ‘not an experiment’ in 2012 actually looked like it wasn’t an experiment. Now Alphabet Access, the group running Google Fiber, plans to lay off 9% of its staff and “pause” all expansions and rollouts, while the Access group’s CEO has stepped down.
Just around the corner in Menlo Park, Facebook’s infrastructure ownership plans were also dealt a blow in September, quite literally. The SpaceX rocket carrying the AMOS 6 satellite exploded on the ground, taking with it Facebook’s planned 18 gigabit-per-second internet capacity over Africa, intended to roll out Internet.org. While the two stories share little, they do highlight a common theme—that owning networks is high in cost, risk and work. For the best part of a century, traditional telcos have been pouring capital investment into wires and nodes across huge disaggregated networks. Where Facebook and Google excelled is in providing hugely scaled, centralised platforms run out of data centers in California and reaching across the world on networks that others had to worry about. Out of frustration at the lack of investment and also from hubris as to their ability to scale all businesses, both invested in owning the internet, pipes and all. For now, those bets look like they might fail.
As always, there are market dynamics and mitigating circumstances that make any analysis more difficult and less polarising. Google’s choice of market to roll out Google Fiber and the broader attempt to bridge the digital divide by offering free internet access to lower income areas did not resonate, and uptake was muted. In addition, incumbents AT&T and Verizon both responded with more aggressive fiber rollout and wireless high-speed internet access through AT&T’s AirGig and Verizon’s “5G” fixed wireless access.
Even if Google Fiber may be on hold, however, it is focusing on point-to-point fixed wireless technology acquired through Webpass. Alphabet is still investing in high-altitude internet drones and balloons, as well as satellites. For its part, Facebook is looking for alternative ways to get internet access over Africa while also investing in internet drones.
Both Alphabet and Facebook investing in infrastructure is symptomatic of the same desire—vertical integration of the access and content. Both offer consumer platform services in the form of search, social, communication, email or maps. Both monetise these platforms at global scale, using targeted advertising based on detailed user and usage analytics. Both thought they could extend reach or increase engagement on these services by operating the networks through which they traffic their services—implicitly suggesting that the incumbent telecoms industry isn’t doing enough to provide consumers with access to their content.
The economics of networks and platforms are very different. Facebook makes around $10-12 on average per year per user. It takes 1.7 billion active users to make a $17 billion business, which means Facebook needs to scale to around 1 in 4 people in the world to make around $20 billion. Alphabet’s $75 billion in mostly advertising revenue is a little harder to contextualise, but involve around 2 billion Android devices, 1 billion Gmail accounts and 3.5 billion daily Google searches, as well as the billions of ad views on affiliate sites through DoubleClick and other Google subsidiaries.
By contrast, businesses that sell the network with a service, such as broadband, mobile or TV, can be in orders of magnitudes smaller to possess the same revenues. DirecTV in the US and Sky UK makes similar amounts of money as Facebook from just 30 million and 20 million subscriptions, respectively—mostly from TV. T-Mobile US makes $24 billion from 60 million mobile subscriptions. Comcast makes a similar amount as Alphabet from around 55 million subscriptions across broadband, telephony and TV; and Verizon does the same from 130 million mobile users. In many cases telecom providers offer multiple services, and so the revenue per home is actually much higher as broadband, mobile, TV and telephony services are added together. This means that 10 million homes can easily generate access and related service revenue equal to Facebook’s total global income. There are more than 10 million people in either Belgium or Greece.
The different ways that network business like Comcast and Verizon, as well as platform business like Facebook and Google, scale goes even further. Facebook and Google both have around 30-40% margin and require around 1,000 employees per billion dollars of revenue. The world’s largest network operators have margins of around 15-20% and closer to 2,000 employees per billion dollars of revenue. While no company necessarily enacts an identical strategy, the consistency in structure suggests there is a chasm in operation structure and expectations—twice as much cost and half the profit is a hard sell for investors and managers alike.
But coming at it from the other direction seems like a much better deal. The same numbers for traditional media and content companies such as studios and broadcasters typically lie in between that of network operators and the big internet platforms. And network access can be sold more easily bundled with content, as has been proven by pay TV for many years; pay TV typically commands revenue per subscriber around the same as average internet, mobile and telephony subscriptions combined. So there are two good reasons to buy into content and platforms if you already have the access networks, but there are no equivalent good reasons to buy or build an access network if you’re already a successful content producer or platform owner. That’s not always been the case—it certainly wasn’t when broadcasters built transmission networks—but it is the case today due to the economics of free internet distribution.
The biggest telecoms deals of the recent years have been network owners moving closer to content and platforms. Broadly matching this trend was Comcast buying NBCUniversal in 2011—although NBCUniversal profiles at around 20% margin and 2,000 employees per billion dollars in revenue, almost identically to Comcast at the time of acquisition—and then Comcast acquiring DreamWorks in 2016. Financially the new and old Comcast look very similar, but have the strategic benefit of a vertical integration at a time of high uncertainty to both traditional business models—that of content and of networks. Making a wide bet means the combined company can respond to a wider range of dynamics. The deal looks good and reduces long-term risk. AT&T’s acquisition of DirecTV in 2014 similarly took a step from access toward pay TV, and the recent announcement to acquire Time Warner is even more attractive: AT&T with DirecTV have around 15 % margin, while Time Warner has closer to 25%. Either way it’s an easy sell to investors.
Which takes us back to the internet giants. Whether you’re Alphabet, Apple, Facebook or Tencent, the economics of networks don’t stack up within the current business. More cost, more people, more risk, lower margins. Networks are necessarily geographically limited, while these companies all sell products or support platforms that are necessarily global. The business case is in fundamental conflict. On the other side, traditional telcos can easily absorb a margin cut on buying content makers and balance it with margin increases on their localised subscriber businesses. Exclusive content deals in the countries they operate in help to drive up the lower margins on their core access and pay TV businesses—there is overhead in the finances.
There is good reason why network owners are vertically integrating content right now, because it makes financial sense for them to do so. But the corollary does not necessarily make sense: global advertising platforms don’t really benefit from buying fiber or satellites—certainly not if they can just piggyback on someone else’s network. But potentially the elephant in the room is another of the internet era’s winners. Amazon has a 2% margin on $100 billion revenue, supported by over 250,000 employees. They make even the most lumbering telco look like a gold mine. Once again, there is good reason why Amazon so heartily invests in the modest margins in cloud storage and services— it’s a better business than their core business. But where Amazon differs from the platform players is that Amazon sees itself as a network much like a telco, distributing physical goods in an internet world. Perhaps it is also some kind of telco.
The financial imperatives of a wide range of companies perhaps disguises the key underlying risks associated with letting companies own the means of consumer access and the content which consumers wish to access. Facebook’s Internet.org and Free Basic essentially offer a constrained, Facebook-approved and -centric version of the internet in exchange for free managed infrastructure. Regulators frequently intervene to ensure content and channels are priced and brokered fairly where telcos own key content creators. Both suggest that some form of internet or content neutrality is required when it comes to access networks. So in this sense, perhaps telcos are the companies that take responsibility for the provision of access in the public interest.