Why Fixed Network Internet Access Features Consumer Speed Tiers, While Mobile Networks Do Not

While it might be difficult to describe the average revenue per account impact of 5G accounts, compared to 4G, it seems far easier to show revenue lift provided by fiber-to-home accounts, compared to copper-based accounts.

Company	FTTH ARPU	Copper ARPU	ARPU Difference
AT&T	$65	$45	$20
Verizon	$70	$50	$20
Lumen	$55	$40	$15
TDS	$60	$45	$15
Brightspeed	$65	$45	$20

In large part, that is likely because FTTH allows sale of higher-priced access plans, which naturally generates higher revenue. Also, unlike the case for mobile account figures, home broadband is “service to a place,” so there is no effect of “multiple users” on recurring price.

In mobile scenarios, multiple users and devices might be supported on a single account, in the form of multi-user accounts, with “lines” or “numbers” ranging from two to some higher number. 

For example, Ericsson suggests 5G boosts average revenue per user by less than five percent, even if many observers suggest ARPA drives significantly higher impact. 

Company	5G ARPA	4G ARPA
AT&T	$100	$80
Verizon	$120	$90
T-Mobile	$90	$70
NTT	$110	$80
Telefonica	$100	$80
Vodafone	$90	$70

It remains unclear whether mobile operators might eventually shift to differentiated access speed plans for their mobile customers, as do home broadband providers. 

There are clear business imperatives underlying each charging principle. The cost of supplying capacity on a fixed network is far lower than on a mobile network, traditionally, largely because fixed networks generally have more ability to add capacity (optical fiber bandwidth is almost infinite; mobile capacity is based on radio frequency resources that are inherently limited). 

Traditionally, fixed networks have huge advantages where it comes to capacity and therefore cost per gigabit of consumption. While we might argue about the precise absolute retail cost of supplying capacity, mobile bandwidth has generally been about two orders of magnitude more expensive than fixed network bandwidth. 

Year	Fixed network (USD/GB)	Mobile network (USD/GB)
2000	100	10,000
2005	10	1,000
2010	1	100
2015	0.1	10
2020	0.01	1
2025	0.001	0.1

Under such circumstances, it makes sense that mobile operators “need” to manage consumption expectations, where fixed network operators can more easily consider “tiers of service” where higher-speed access is sold at higher prices. 

Rarely will mobile operators, with their higher cost to supply capacity, have high incentives to encourage higher customer data consumption by offering higher-speed (and higher-priced) consumption tiers.

Lower Prices are a Feature, Not a Bug, for Policymakers

Lower prices are a policy feature, not a bug. Government policy promoting competiton is designed to create lower retail prices. So it should not be surprising that pressure on average prices per user or customer or account now are a major service provider concern.

What else would you have expected? Lower prices are the intended outcome of competition policy.

To what extent is it correct to characterize legacy service provider revenue trends as “down and to the right?” Obviously legacy services such as fixed network voice, mobile messaging and voice and linear entertainment video generally show that pattern. 

What we often forget is that the very objective of introducing competition for connectivity services, and the government policies to support that objective, are intentionally designed to create lower prices. The objective of policy is a “down and to the right” pattern. 

In other words, “down and to the right” is not a bug, it is a feature. It represents the outcome policy intends, and is not a defect of policy. 

source: United Nations

The other angle is that a proven way of increasing ARPU is to increase speed, despite another clear trend: over time, speeds grow but prices remain relatively flat, or even decline. In other words, the cost of a 300-Mbps connection is the same, or less, than a 512-kbps connection three decades ago. 

Over the short period of 2007 to 2017, for example, U.S. typical speeds grew by two orders of magnitude, while prices dropped. 

At the moment, up to 80 percent of U.S. locations can buy internet access operating at least the gigabit per second level.  

source: Versa Technology 

The global pattern is not always transparent. If one looks only at total service provider revenue, that tends to grow each year, in large part because new customers are added in growing regions (Asia, primarily, but also eventually in Africa). 

So global service provider revenues will grow 14 percent between 2022 and 2027, according to researchers at  Omdia. Monthly average revenue per user will fall by four percent. So the pattern is more customers, each paying less than the “typical customer” used to pay. Revenue might ber “up and to the right” but ARPU clearly is “down and to the right.”

source: Omdia 

Every management team seems to emphasize that value can be enhanced, preventing further commoditization. You can make your own assessment of how effective such efforts have been, or could be. 

Disposable income is among the other limitations. It will be hard to boost ARPU very much in lower-income countries. If policymakers succeed in reducing the cost of connectivity to perhaps two percent of gross national income per capita, that further limits ARPU upside. 

source: S&P Global Market Intelligence 

In other words, the goal of policy in developing countries is to actively reduce ARPU. The objective is precisely “down and to the right” pricing. 

Until recently, home broadband was the major product line producing “up and the right” results, but growth now has slowed in mature markets. That pattern looks more like “flat and to the right.”

When service provider executives talk about a transition from “telco to techco” they essentially are saying such moves will change the revenue picture to “up and to the right.” 

Service providers in emerging or younger markets have advantages, in that regard. They can still hope to rely on mobile subscription growth and uptake of mobile internet access to fuel their continued “up and to the right” growth prospects.

Mature market executives who own infrastructure assets have no such luxury, and strategic options often hinge on whether mobility revenues exist. Contestants in mobile or fixed businesses who operate using wholesale access, rather than owning infra, have other options. 

Competitors with Infra-based business models complain about higher capital investment requirements and limited abilities to monetize those investments, not without basis in fact. 

Still, we often forget that the whole point of introducing competition in access markets is to drive average costs “down and to the right.” It is a policy feature, not a bug.

Are FTTH Payback Models Sustainable? What's Good for Private Equity Might Not be So Good for Operators

Some of us would admit to being surprised at the payback models  for fiber-to-home deployments, the degree of business moat protection some believe FTTH offers, and therefore the value of such digital infrastructure assets, compared to other assets such as cell tower sites, data center or edge computing assets. 

Looking back over 25 years of business model assumptions, it is quite startling how much the underlying assumptions have changed. Subsidies now play a bigger role, offering in some cases a 20-percent to 30-percent reduction in capital investment in rural markets. 

On the supply side, though demand is not altered, private equity investment now means more capital is available to accelerate build timetables. 

But the most-shocking change are the revenue assumptions for consumer locations. These days, the expected revenue contribution from a home broadband account hovers around $50 per month to $70 per month. Some providers might add linear video, voice or text messaging components to a lesser degree. 

But that is a huge change from revenue expectations in the 1990 to 2015 period, when $150 per customer was the possible revenue target. In some cases, revenue up to $200 per home location was considered feasible. 

Expectations now hinge almost exclusively on consumer home broadband. 

“Our fiber ARPU was $61.65, up 5.3 percent year over year, with gross addition intake ARPU in the $65 to $70 range,” said John Stankey, AT&T CEO, of second quarter 2022 results. “We expect overall fiber ARPU to continue to improve as more customers roll off promotional pricing and on to simplified pricing constructs.”

Lumen reports its fiber-to-home average revenue per user at about $58 per month.

Recent presentations also have shown fiber-to-home home broadband average revenue per user of about $63. 

source: Frontier Communications 

Granted, most larger ISPs believe they can boost ARPU over time, by adding features, adding speed tiers and moving customers to higher-priced plans with higher usage allowances. 

Market share or installed base is the other huge assumption. Can most providers expect to get 20 percent take rates or as much as 50 percent? And what other assumptions about operating cost are necessary to create a sustainable business case at 20-percent share? 

Most incumbent telcos deploying FTTH have been able to get 40-percent market share after several years of marketing. But is a terminal rate around 40 percent to possibly 45 percent (or even 50 percent) reasonable in most cases? If not, where is that possible? 

On the supply side, capital investment benefits from government subsidies and to some extent infra cost declines, though construction costs are stubborn. So while some larger ISPs hint at per-passing network costs as low as $600, others report costs closer to $1,000 per passing, with connection costs of $550 to $600 per customer. 

source: Matt Nicholson Lewis 

The point is that even with subsidies, lower infra gear costs and new investment sources, the demand expectations for consumer services have been slashed as much as two thirds over the past several decades. Many ISPs no longer expect revenue contributions from voice or entertainment video sources, and must build their demand models based solely on internet access. 

The largest ISPs might still expect some revenue contribution from voice or video services, but seem to be modeling higher expectations for business connectivity or contributions to mobile infrastructure cost models. In other words, the cost of small cell infra is aided by the consumer FTTH investment. 

Ultimately, we will see whether  FTTH really underpins a business that provides a competitive moat, while throwing off predictable cash flow. It seems more likely that private equity could succeed in transforming a legacy copper-based access business into a fiber access business, with a boost in equity multiples that will justify the effort.

That might well be a different question than asking whether FTTH really is a real estate or infrastructure asset on par with airports, toll roads, electrical and gas utilities. Much of the bet relies on limited competition. The argument that the first FTTH provider in a suburban or urban market gets 40 percent market share is likely correct. 

That has a reasonable chance of being correct even if two equally competent providers operate their own FTTH networks in a market. 

Some believe the first-mover advantage in a rural market could be substantial enough to support higher market shares. 

But it remains a valuable exercise to ask whether the FTTH business model is sustainable on an operational business on the revenue scales now being seen. 

That is a different question than asking whether a private equity buyer can boost multiples--and then sell the asset--by replacing copper access with optical fiber access. 

As a matter of operating economics, it still seems unclear whether FTTH networks generating only $50 to $70 per month in residential revenue are sustainable, assuming legacy provider cost structures. A small, lean upstart should have an easier time, as embedded costs are lower than those found at legacy firms. 

Does Home Broadband Data Consumption Really Tell Us Anything about Economic Lift?

Is it really possible to quantify the economic uplift from internet access or home broadband operating at 100 Mbps or 500 Mbps? One might argue there is a difference between access at 100 Mbps or 1,000 Mbps. But do we actually know that? 

And consider data consumption as a “good thing” if it is higher. My own account recently surpassed 1 Tbyte per month (and continues to climb). One might argue that is evidence of some sort of “productivity” advantage. It is not.

The reason more data is being consumed is because the household watches streaming video, often in 4K. Consumption of linear video (which does not increase data consumption has dropped to the point that only some live news and sports are ever viewed.

So in my case, higher data consumption has nothing to do with productivity, or learning or work. It is entertainment video consumption, pure and simple. If there were any heavy gamers on the account, that might also drive higher consumption. 

But productivity? The increased amount of consumed data has nothing to do with it.

Many industry trade groups have to walk a fine line when addressing usage, take rates, revenue and speeds, in relation to societal or economic benefits.

Proponents must argue their industries create lots of value for economies and society; are well positioned for growth and at the same time, still need some help. The connectivity industry seems always to be in that position.

A new report on digital communications issued by the European Telecommunications Network Operators’ Association points out that fiber-to-home coverage has passed 50 percent of locations and that 5G coverage likewise has doubled over a single year from about 30 percent to 60 percent. 

On the other hand, ETNO says, peers are doing much better. “Uptake of 5G in Europe has been lagging behind,” says ETNO. “Despite being available to 62 percent of the population, 5G in Europe constitutes only 2.8 percent of the total mobile connections, compared to 13.4 percent  in the US and 29.3 percent in South Korea.”

That implies that uptake is a problem. 

On one hand, ETNO argues that “telcos” have increased their  commercial activities in edge computing, Open radio access networks, internet of things, “big data” and security, making the argument that telcos are innovating and investing. 

ETNO also notes that  average mobile data usage per capita per month, in 2020, was 8.52 GB in Europe, 10.62 GB in the US and 12.52 GB in South Korea.The implication there is that usage volume is a problem. 

Average spend per capita on communications in Europe is forecasted to be €33.8 per month, lower than global peers (€71.7 in the US, € 36.1 in South Korea). So lagging average revenue per account is deemed to be a problem. 

source: ETNO

Service provider revenues in Europe are also lower than in other geographies. Mobile average revenue per user (ARPU) was €14.4 in Europe, compared to €37.9 in USA and €25 in South Korea. Again, this is viewed as a problem, ETNO notes. 

source: ETNO

The same general pattern holds for home broadband revenue, ETNO says. 

source: ETNO

ETNO also argues that average home broadband downlink speeds are higher in some peer markets. 

source: ETNO

The message is equal parts “we are vital contributors to society and economy” and “our financial survival is imperiled.” Says ETNO, “networks are vital, but the financial outlook remains unclear for the telecoms sector.”

This is the sort of argument any industry would make when it wants to show it is important for the government to support the industry, and also arguing why that support is required. 

Aside from arguments over which other participants in the ecosystem should be compelled to contribute additional support (capital investment, usage fees, support payments), how one thinks about usage, customer revenue magnitude, ISP revenue magnitude and prices are an areas where more debate is possible.

For example, the explicit assumption is that higher data consumption by customers is better than lower data consumption. Higher customer spending is assumed to be better than lower spending. 

The assumption is that uptake of the newest networks is good, in and of itself. Hence, higher 5G adoption rates are good; lower rates are bad. 

ETNO notes that average public market equity values for ISPs are lower than for other categories of firms, the typical argument being that content and app providers earn higher multiples of revenue and therefore are valued more highly. 

That is obviously true, but also ignores the fact that each industry can have a different market valuation, for reasons the market assigns. Growth companies are valued differently from value assets. Retailers are valued differently than software and information technology companies. 

Different parts of the financial services business are valued differently as well. The point is that markets assign valuations. The mere existence of differences only indicates that the market values some firms and industries higher or lower, for reasons related to growth potential, business moats, revenue consistency or any number of other reasons. 

The implicit argument made by suppliers and proponents of information or communications technologies is always that society and the economy profit when such adoption happens. Most often, the claims go further and argue that economic growth actually is driven by the rate of new technology adoption. 

All of those assumptions can be challenged. At some level, we might all agree that universal availability of electricity is correlated with economic growth. But correlation often is not evidence of causation. If areas with the same degree of access to electricity still have different outcomes and growth rates and magnitudes, something else is at work,. 

Also, most policymakers embrace lower consumer prices as a positive good. So some would argue lower prices are a good thing, not a “problem.” 

Also, demand is different from supply. In arguing that consumption is an issue in Europe, ETNO essentially argues that lower demand is a problem. As much of a problem as that might be for ISPs, it is not so clear that demand actually is a “problem.” 

We need to separate two different issues. One issue is making sure home broadband and high-quality mobile service is ubiquitous. But a separate issue is how consumers avail themselves of those resources. 

But data consumption, data rates, average revenue per account or average cost per access might not have much to do with social or economic uplift.