"Telecom" is No Longer Seen as a "Natural Monopoly," But Might Some View it as a Functional Monopoly?

Before the 1980s, global telecom regulators universally considered telecommunications to be a “natural monopoly.” 

Nobody uses the term anymore, as it is obvious connectivity services are not a “natural” monopoly. In most countries, an oligopoly tends to exist at the top of the market, though there can be hundreds to thousands of smaller contestants in large continental markets. 

On the other hand, in some markets, we might find policymakers concluding that access services ("telecommunications") is a functional monopoly, if not necessarily "natural."

In the monopoly era, the owner and operator often was a national government. Then began a worldwide shift to deregulation and privatization as a prelude to allowing more competition in formerly-restricted monopoly telecommunications. Often that takes the form of promoting wholesale arrangements that allow retailers to use a single national network. 

Relatively fewer countries have seen significant fixed network competition based on alternate facilities, but facilities-based competition has been the norm for mobile services.

New wireless licenses issued to many smaller firms will be cited as potential new sources of competition as well, since most connectivity services competition has occurred on the mobile networks. 

Of course, access services remain a scale game. There is no contradiction between services provided by hundreds of small firms with small customer bases and domination of the market by three to four providers. 

But several decades of competition at scale have produced a business where profits are hard to come by, while heavy capital investment, if anything, seems to be increasing. The near-term result has been waves of consolidation and a market structure that is oligopolistic.

There always are at least three sets of opinions  in that regard. Some believe monopoly is the ultimate outcome, as fixed networks will simply be too expensive, with too little revenue, in a facilities-based competitive scenario. 

Some who view facilities monopoly as inevitable therefore argument for a robust wholesale monopoly to support retail competition using the one network. 

Others argue  that a duopoly based on facilities ownership might be the best sustainable outcome, and might produce more innovation than a wholesale approach. By definition, in a wholesale-only framework, the capabilities of the network can be purchased by all retailers, at prices that are differentiated only by possible volume discounts. 

Competitive differentiation then mostly occurs when contestants bundle other non-access services or can leverage some operational or marketing advantage. 

Perhaps the easiest way to illustrate that potential is note that telcos using fiber-to-home; cable operators using hybrid fiber coax and mobile operators using distinct physical platforms can create services aimed at different market segments and customers precisely because their platforms are distinct in terms of cost and capabilities. 

So the issue is how much consolidation will happen, and at what point--if at all--supply and demand in the connectivity business are at equilibrium. In other words, what structure will emerge that allows service providers to sustain themselves with adequate profit levels, while still ensuring the benefits of competition for consumers of those services?

If access services are not a "natural monopoly," might they be oligopolies in most cases (mobile and fixed providers able to sustain themselves)? Still, in some cases, policymakers might conclude that either mobility or fixed services are better supported by a monopoly provider, albeit with strong wholesale arrangements.

That might especially be the case in markets where new services beyond bandwidth are likely to be big opportunities.

How Big a Deal is Edge Computing as a Revenue Driver for Connectivity Providers?

Edge computing possibly can grow to generate a minimum of $1 billion in annual new revenues for some tier-one service providers. The same might be said for service-provider-delivered and operated  private networks, internet of things services or virtual private networks. 

But none of those services seem capable of driving the next big wave of revenue growth for connectivity providers, as their total revenue contribution does not seem capable of driving 80 percent of total revenue growth or representing half of the total installed base of revenue. 

In other words, it does not appear that edge computing, IoT, private networks or network slicing can rival the revenue magnitude of voice, texting, video subscriptions, home broadband or mobile subscription revenue. 

It is not clear whether any of those new revenue streams will be as important as MPLS or SD-WAN, dedicated internet access or Ethernet transport services, for example. All of those can be created by enterprises directly, on a do-it-yourself basis, from the network edge. 

The point is that even when some new innovations are substantial generators of revenue and activity, it is not automatically connectivity providers who benefit, in terms of direct revenue. 

One rule of thumb I use for determining whether any proposed new line of business makes sense for tier-one connectivity providers is whether the new line has potential to produce a minimum of $1 billion in annual revenues for a single provider in some definable time span (five years for a specific product. 

By that rule of thumb, tier-one service providers might be able to create edge computing revenue streams that amount to as much as $1 billion in annual revenue for some service providers. But most will fail to achieve that level of return in the next five to seven years.

That is not to say "computing at the edge" will be a small business. Indeed, it is likely to account for a growing part of public cloud computing revenues, eventually. And that is a big global business, already representing more than $400 billion in annual revenues, including both public cloud revenues as well as  infrastructure spending to support cloud computing; the value of business applications and associated consulting and services to implement cloud computing.

The leading public cloud computing hyperscalers themselves represent about $72 billion or more in annual revenues already. All the rest of the revenue in the ecosystem comes from sales of software, hardware and services to enable cloud computing, both public and private.

source: IoT Analytics

It is likely a reasonable assumption that most public edge computing revenue is eventually earned by the same firms leading public cloud computing as a service.

Perhaps service provider revenues from edge computing could reach at least $20 billion, in about five years. By that standard, multi-access edge computing barely qualifies as "something worth pursuing," at least for tier-one connectivity service providers.

In other words, MEC is within the category of products that offers reasonable hope of payback, but is not yet in the category of “big winners” that add at least $100 billion to $200 billion in global service provider revenues. 

In other words, MEC is not “mobile phone service; home broadband. Perhaps it will be as big as MPLS or SD-WAN. For tier-one connectivity providers, perhaps MEC is more important than business voice (unified communications as a service). 

source: STL, KBV Research 

As with many other products, including Wi-Fi, SD-WAN, MPLS, 4G or 5G private networks, local area networks in general and  enterprise voice services, most of the money is earned by suppliers of software (business functionality) and hardware platforms, not end-user-facing services. 

The reason is that such solutions can be implemented on a do-it-yourself basis, directly by enterprises and system integrators, without needing to buy anything from tier-one connectivity providers but bandwidth or capacity. 

So one reason why I believe that other new connectivity services enabled by 5G likely do not have the potential to substantially move the industry to the next major revenue model is that none of those innovations are very likely to produce much more than perhaps one percent of total service revenues for the typical tier-one service provider. 

The opportunity for big public connectivity providers lies in use cases related to the wide area network rather than the domain of indoor and private networks. That is why the local area networks industry has always been dominated by infra providers (hardware platforms) and users who build and own their own networks (both enterprise and consumer). 

And most of the proposed “new revenue sources” for 5G are oriented towards private networks, such as private enterprise local area networks. Many of the other proposed revenue generators can be done by enterprises on a DIY basis (edge computing, internet of things). Some WAN network services--such as network slicing--attack problems that can be solved with DIY solutions.

Edge computing is a solution for some problems network slicing is said to solve, for example. 

None of the new 5G services--or new services in aggregate-- is believed capable of replacing half of all current mobile operator revenues, for example. And that would be the definition of a “new service” that transforms the industry. 

All of which suggests there is something else, yet to be discovered, that eventually drives industry revenue forward once mobility and home broadband have saturated. So far, nobody has a plausible candidate for that new service.

Edge computing might be helpful. So might network slicing, private networks or internet of things. But not even all of them together are a solution for industry revenue drivers once home broadband and mobile service begin to decline as producers of at least half of industry revenues.

It already seems clear that others in the edge computing ecosystem--including digital infra providers and hyperscale cloud computing as a service suppliers--will profit most from edge computing.

Power Users Aren't What They Used to Be

“Power users,” defined as accounts using far more data than the typical home broadband user, are not necessarily what you might think. Though we might traditionally have thought of such power users as major content creators or users with extraordinary downloading behavior, that arguably is no longer the case. 

The phrase “yesterday’s power user is today’s typical user” is apt. Where perhaps 11 percent of home broadband users in the third quarter of 2022 were power users, consuming at least a terabyte of data each month, the typical or average account used perhaps 496 gigabytes, with a median consumption of perhaps 324 gigabytes, 

According to Openvault, about 16 percent of accounts in the third quarter of 2022 were “power users” consuming at least a terabyte of data per month. Perhaps we once thought of power users as people with much higher than average computing skills, perhaps including software code writers, very-active content creators and sharers or online gaming enthusiasts. 

These days, the popularity of video streaming adds a more mundane class of users: people who watch lots of entertainment video. Perhaps a working definition is a person or household that streams at least eight hours of video each day. Do that and it is easy to top a terabyte of usage in a month. 

source: T-Mobile 

It should then come as no surprise that Openvault data shows a continued increase in gigabit service plan adoption, as well as migration of subscribers to speeds of 200 Mbps or higher. Though there is no linear casual relationship between access speeds and total data consumption, the two phenomena are correlated. 

Faster speeds allow more to be done in any X amount of time, so more data can be consumed in X amount of time, for example. Over time, applications also are designed to take advantage of higher bandwidths (speed), such as embedding autoplay full-motion video into apps. That increases “involuntary” data consumption. 

Some 15 percent of U.S. households purchased gigabit tier plans in the third quarter of 2022, an increase of 35 percent  over the 11.4 percent market share 12 months prior, says Openvault. 

As always, typical speeds also increased for typical accounts. The percentage of accounts buying service in the 200 Mbps to 400 Mbps range doubled to 54.8 percent  from 27.4 percent over the last year. 

At the end of the third quarter, only 4.7 percent  of all subscribers were provisioned for speeds of less than 50 Mbps, a reduction of more than half  from the third quarter 2021  figure of 9.8 percent. .

Average monthly usage of 495.5 GB was up 13.9 percent from 3Q21’s average of 434.9 GB, and represented a slight increase over 2Q22’s 490.7 GB. Median broadband was up 14.3 percent year over year, representing broader growth across all subscribers.

Year-over-year growth of power users of 1TB or more was 18 percent, to 13.7 percent of all subscribers, while the super power user category of consumers of 2 TB or more rose almost 50 percent during the same time frame. 

source: Openvault  

No 5G New Revenue Source Matches Fixed Wireless

It is virtually impossible to argue with the proposition that the first new revenue stream of any significance enabled by 5G is fixed wireless internet access. Revenue from fixed wireless far exceeds revenues from other potential new revenue sources. 

Revenues from 5G fixed wireless, in the near term, will dwarf internet of things, private networks, network slicing  and edge computing, for example. 5G fixed wireless might, in some markets, represent as much as eight percent of home broadband revenues, for example. 

None of the other potential revenue sources--network slicing, edge computing, internet of things, private networks--is likely to hit as much as one percent of total service provider or mobile service provider revenues in the near term. 

As important as edge computing might be, as a revenue growth driver for mobile operators, revenue contributions might be relatively slight for some time. The same might be said for the revenue contributions made by internet of things services as well.

In 2024, it is conceivable that IoT connectivity revenues for mobile operators globally could  be in the low millions to tens of millions of dollars, according to Machina Research. Millions, not billions. 

In 2026 the global multi-access edge computing market might generate $1.72 billion. Even if one assumes all that revenue is connectivity revenue booked by mobile operators, it still is a far smaller new revenue stream than fixed wireless represents. 

If the home broadband market If the home broadband market generates $134 billion in service provider revenue in 2026, then 5G fixed wireless would represent perhaps eight percent of home broadband revenue. 

Do you believe U.S. mobile operators will make more than $14 billion to $24 billion in revenues from edge computing, IoT or private networks?

Nor might private networks or edge computing revenues be especially important as components of total revenue. It is almost certain that global service provider revenues from multi-access edge computing, for example, will be in the single-digit billions ($ billion) range over the next few years. 

Though a growing number of home broadband subscribers should have access to gigabit speed fixed wireless service eventually, present coverage is relatively nil. 

T-Mobile and Verizon are expected to have 11 to 13 million total fixed wireless customers by the end of 2025. If total U.S. internet  accounts are somewhere on the order of 111 milliion accounts, and if small business users account for 11 million of those accounts, then home users might amount to about 100 million accounts. 

source: Ooma, Independence Research 

If Verizon and T-Mobile hit those targets, their share of the home broadband market--counting only fixed wireless accounts-- would be about 10 percent. Significantly, most of those accounts will be gained “outside of region” for Verizon. That is significant as Verizon’s fixed network only reaches about 20 percent of U.S. households. Fixed wireless allows Verizon to grow its account base among the 80 percent of U.S. home locations that cannot buy Verizon fixed network service. 

For T-Mobile, which in the past has had zero percent market share in home broadband, all of the growth is incremental new revenue. If those accounts add $600 per year in added revenue, then the 11 percent share of home broadband supplied by fixed wireless represents perhaps $6.6 billion in new revenue for the two firms. 

That is a big deal, considering how hard it is for either firm to create a brand-new line of business that generates at least $1 billion in new revenue. 

The other apparent takeaway is the size of the market segment that cares more about price than performance. 

Segments exist in the home broadband business, as they do in many parts of the digital infrastructure and digital services businesses. In other words, even if some customers want faster speeds at the higher end of commercial availability, up to 20 percent of the market cares more about affordable service that is “good enough.”

The center of gravity of demand for 5G fixed wireless is households In the U.S. market who will not buy speeds above 300 Mbps, or pay much more than $50 a month, at least in the early going. T-Mobile targets speeds up to 200 Mbps. 

During the third quarter, about 22 percent of U.S. customers bought service at speeds of 200 Mbps or below. In other words, perhaps a fifth of the home broadband market is willing to buy service at speeds supported by fixed wireless. 

source: Openvault  

The other takeaway is that home broadband net account additions over the past year have disproportionately come on the fixed wireless platform, representing about 78 percent of all net new accounts. 

source: T-Mobile 

To hang on to those accounts, Verizon and T-Mobile will have to scale speeds upwards, as the whole market moves to gigabit and multi-gigabit speeds. Some percentage of those upgraded accounts could come on some fiber-to-home platform. Even out of region, both firms could strike deals for use of wholesale assets. 

But the bigger part of the retention battle is going to center on ways of increasing fixed wireless speeds to keep pace with the ever-faster “average” speed purchased in the home broadband market. 

And that almost certainly means using millimeter wave spectrum to a greater degree. In the end, even using small cell architectures, there is only so much capacity one can wring out of low-band or mid-band spectrum. 

As a practical matter, almost all the future bandwidth to support mobility services or fixed wireless is to be found in the millimeter wave regions. 

In the WAN, the Next Generation Network Can Take Quite a While to Make Sense

As 400 Gbps wide area network systems now are touted, one thinks back about 20 years to another time when service providers were weighing different solutions for their WANs. 

Back then, when the WAN standard for long-haul optical transmission was 2.5 Gbps, WAN operators were pondering the cost and value of upgrading to either 10 Gbps or 40 Gbps.

As I recall, back around 2000 the cost of upgrading to 10 Gbps was about 2.5 times the cost of greenfield 2.5 Gbps networks.

Again, as I recall, 40 Gbps networks cost more than four times the cost of 10 Gbps. That made the decisions in favor of 10 Gbps--especially given the amount of dark fiber availability--logical. 

Assuming the same pattern holds, WAN operators needing to add more capacity will opt for 100 Gbps networks once the price premium over 40 Gbps is about 2.5 times. 

How long it will take for the economics of 400 Gbps to reach levels where it makes more sense, for most service providers, rather than a 100 Gbps upgrade is the issue. 

As always, the economics are easier for metro networks and within data centers. But long haul upgrade economics are more difficult. 

As always, “how long can we afford to wait?” affects the decision. Recalling the huge amount of dark fiber put into place around the turn of the century, the other issue is whether it makes more sense to continue using 10 Gbps on multiple fibers rather than upgrading either to 40 Gbps or 100 Gbps in the WAN. 

Unlike baseband standards, which tend to increase by an order of magnitude each major generation, optical transport systems often do  not, in part because of the technical issues for optical waveguides such as  chromatic dispersion and cost issues such as port density. 

That was an issue 20 years ago when network operators were looking to upgrade capacity to 10 Gbps or further to 40 Gbps, for example. Typically, backwards compatibility tends to be a bigger issue for long-haul and access network operators than for operators of data centers. 

And a jump from 1 Gbps to 10 Gbps was easier to finesse than a leap to 40 Gbps. The consideration often involves an upgrade over existing optical cabling networks that disturbs existing operations the least possible amount. 

Often, an order of magnitude leap in bandwidth requires quite a lot of network element replacement, and therefore higher cost. Local networks in the past used multimode fiber, while long-haul networks use single mode fiber. 

That has cost implications. The cost of using a four-by-10 Gbps solution is roughly four times as much as a 10 Gbps solution, for example, on multimode networks. On a multimode, short-range network, a  10X solution (10 Gbps upgrade to 100 Gbps) costs about 10X more. 

On a single mode, long haul network, the cost of upgrading from 10 Gbps to 40 Gbps is 4X. But the cost of upgrading to 100 Gbps is far more than 20X. 

source: Cisco

The reason, at a high level, is that the upgrade by 4X generally uses some form of multiplexing the older existing standard, but using more fibers in a cable. A multimode fiber network upgrade might involve only a switch of line cards. 

On such local networks, where a 10-Gbps uses two fibers, transmission on a 40-Gbps multimode fiber network uses as many as 12 fibers in a cable, for example. An upgrade from 10-Gbps to 100 Gbps means upgrading from a two-fiber cable to a 24-fiber cable. 

A 10X upgrade tends to be less of an issue for local users such as data centers or even metro networking suppliers but a much-greater issue for connectivity providers with comparatively greater sunk investments in optical cabling infrastructure. 

Cost issues always matter, which is why, at transition points, service providers often are asked to choose between one option generally available now and a higher-performance option expected in a few years. 

At least, that is what I seem to recall from past evaluations made by WAN operators about core network upgrades. 

Is Reliance on Public Cloud Dangerous for Telcos?

How big a threat are hyperscale cloud computing suppliers to the connectivity service provider customers they serve? Not much, if at all, some argue. Others think the danger is significant. 

"Let's be honest, they could run the whole network for probably half the cost," said John Giere, Optiva CEO. Giere seems to be referring only to the cost of server resources, though. "Sixty-five percent of servers are bought by five companies in the world.”

The implication is that the sheer cost of compute infrastructure could be much lower than any single connectivity provider could obtain. 

As always, the analysis of total cost is much more complicated than the cost of acquiring server capacity to run software. There are personnel costs, software costs, maintenance, power and redundancy costs, for example. 

And the complex analysis arguably turns on value. Would you prefer a 50 percent return on investment on an investment that improves 5 percent of your cost structure, or a 25 percent return on investment on an investment that improves 20 percent of your cost structure? 

In other words, connectivity providers have to understand where their costs are created. 

Consider thinking about 5G-based edge computing networks. Some might see costs centered in the core network while others believe it is at the edge. As networks become more virtualized and disaggregated, those perceptions might prove equally correct or incorrect, no matter which view presently holds. 

Perhaps the biggest difference right now is thinking about return on investment in core networks, as, by definition, much network investment for edge computing must happen “at the edge.” 

source: IBM Institute for Business Value 

At some volume level, the economics of using public cloud services on an outsourced basis rather than running owned compute infrastructure might even reverse. For enterprises, at some level of volume, it virtually always makes sense to buy and own rather than lease and pay for services. 

But there are other considerations. Some might say the task of managing the network is not where the value lies. Rather, it is in the ability to tap the latest and most-advanced compute capabilities to build new services at the customer-facing level. 

If you think about the problem as a matter of operations support systems--not strictly “compute resources”--you get a glimpse of the difference between a platform to “run the network” and a platform to create new services. ” 

Simply all operations support and compute platforms exist to support business outcomes, but those outcomes also hinge on go-to-market skill.

Simply, one cannot create a customer-facing service unless the underlying network can support it. But creating a new customer-facing service relies on much more than the compute or connectivity platforms. 

Much of the effort and skill centers on domain knowledge, code-writing capabilities, marketing skill and an internal organization that does not slow down or block such development. Add capital resources, depth of ecosystem partners, a lack of distracting other issues and organizational agility to that list of requirements. 

In other words, using a hyperscale partner as the compute platform is less about the cost of managing the core network and more about  leveraging a platform for creating new services that require the use of the network, at the “my platform can support that” level. 

Such capabilities arguably apply most to services enterprise customers value, deploy and buy. That applies to capabilities enterprises themselves require for internal consumption as well as the products they create and sell to consumers. 

Perhaps the biggest connectivity provider fear is losing control of key capabilities in the network technology or customer-facing markets. That is not an insignificant risk. 

Still, in an increasingly open, disaggregated, layered ecosystem of value creation, revenue models and functions, it is hard to argue against the proposition that the hyperscalers will always be able to innovate at the platform support level faster than connectivity providers can accomplish.

Telcos used to build their own consumer devices; their own switches (special purpose computers), their own operating systems, billing systems and so forth. They rarely do so anymore. 

As the architecture of computing changes, connectivity providers simply are changing with those evolutions.

"Blaming the Victim" When Surveys Don't Work as Expected

Some might call the effort people put into their survey responses as “satisficing.” As applied to survey response data, the term means some people are not thinking too much about the actual responses they are giving to the poll questions. That might be akin to "blaming the victim" of a crime for crime's commission.

Some of us might argue the term "satisficing" is quite misapplied. To the extent "satisficing" can be said to apply, most of it already has been applied in the design of the polls or surveys.

To be sure, the definition os “satisfice” is to “pursue the minimum satisfactory condition or outcome.” 

As used to describe survey respondent behavior, it connotes “choosing an alternative which is not the optimal solution but is a solution which is good enough.” 

source: FourweekMBA 

But that is precisely what multiple choice survey instruments require. As often stated, respondents are told to “pick the answer that most represents your views.” As most of us can attest, oftentimes none of the available options actually represents our “true” opinions. No matter. 

Also, unlike Simon’s search for understanding of decision making, he challenged the notion that human thinking actually could encompass all possible solutions. The whole point is that humans cannot do so, so a “good enough” solution always is chosen. 

In a multiple-choice survey instrument, the designers already have eliminated all but a set of choices. Respondents do not have to choose the “best possible” response, only that response presented to them, which is a handful of choices. 

The “satisficing” already has occurred, but it does not represent  respondent behavior: it represents all the simplifying decisions made by the designers of the survey instrument. 

One must indicate which answer “best” fits one’s views. The term “satisficing” was created in 1947 by Professor Herbert Simon  n his 1947 book Administrative Behavior. 

His argument was that humans cannot be fully rational when making decisions. So-called  rational choice theory, which asserts that this is how decisions are made, is unrealistic, Simon argued. 

Instead, what humans actually use is a process he called bounded rationality. What humans actually do, since they have limited data, limited time and limited capabilities, is seek a workable solution to a problem, not technically the “best possible” solution. 

The concept is that humans do not have unlimited time, resources or capability to rationally consider all possible solutions to any problem, and then choose the optimal solution. Given all the constraints, they search for a limited number of solutions that will work, that are “good enough” and proceed. 

As applied to survey design or survey response, bounded rationality--known as “satisficing”--already has been employed. Survey designers already have chosen a very-finite set of “solutions” or “answers” to problems, issues, attitudes or choices that might possibly be made in real life. 

Perhaps the real answer--from any respondent--is that they would choose “none of the above,” all of the time, for reasons they have no way to communicate to the survey design team. 

Perhaps it is understandable that survey instrument designers fault their respondents for providing “bad data.” Some of us would submit that is not the problem. The problem is the faulty architecture of thinking about the issues for which answers are sought; the explicit choices offered to respondents; the forcing of responses into a predetermined framework; using language not nuanced enough to capture actual choices, beliefs, preferences or possible actions. 

If the data does not fit one’s assumptions or existing beliefs, whose fault is that?