Do ISPs Really Suffer From High Data Demand?

When access providers complain about their revenues not keeping pace with the infrastructure they must build to support their own customer requirements, that argument arguably makes more sense for fixed network home broadband than mobile internet access, looking at mobile data consumption and mobile operator service revenues as analyzed by Omdia. 

Granted, mobile networks are less capital intensive than fixed access networks, but mobile data consumption and mobile data revenue growth are not inversely related. Granted, traffic growth is happening at a faster rate than revenue growth, but the revenue trend still is positive. 

source: LightReading, Omdia

Some of us might argue that even that trend is misleading, in the sense that it is mobile operator customers who are invoking delivery of nearly all that data, as they watch streaming video, listen to streaming audio or interact with image-rich social media sources.  

To the extent that networks are supposed to compensate each other for use of facilities, the asymmetrical flow of data between customers on one network--who request content delivery from sources on other networks--might be viewed as an application of the “calling party pays” arrangement. 

In such content sessions, it is the calling party that creates the data demand by invoking delivery of content from a remote network. Only by ignoring that reality can it be claimed that the content delivery network is the “calling party.”

Assume the internet value chain represented about $6.7 trillion in annual revenues, as estimated by Kearney and GSMA. Assume internet access revenue was about 15 percent of that total amount, or about $1 trillion earned each year providing internet access. 

Assume global revenue earned by “telcos” was about $1.5 trillion, as estimated by IDC. That implies that as much as 66 percent of total telco and ISP revenue earned annually was generated by internet access.

Even if that is a high estimate, it suggests the importance of internet access for telco and ISP revenue and profits. 

It is impressionistic, but even if data demand grows at faster rates than access revenue, logic would suggest that profit margins for internet access are likely higher than for many other services including business data networking, video services or mobile messaging and voice. 

Perhaps only legacy voice services, which generally speaking are harvested, requiring minimal new capital investment, might have margins higher than internet access. 

Some 40 years ago, linear video might have produced margins in the 40-percent range, where today most providers would be lucky to see 10-percent margins. 

By some estimates voice service margins in 1980 were as high as 50 percent. Mobile voice margins might have been in the 30-percent range in 1980 and might be as low as five percent today. 

But even if we use a blended rate of 10 percent for ISP and telco service margins, internet access, as the largest product category, still produces the greatest volume of profits. 

Again, it might only be illustrative, but ISPs might well be making average profits on their internet access services, but up to two thirds of gross revenues on those services. 

We can argue about the cost of delivering capacity now, compared to 49 years ago, but nobody would question that the cost to deliver a bit has declined dramatically over that period. 

In that sense, the total capacity demand generated by an ISP’s customers might not matter as much as often portrayed. What matters more is the contribution internet access makes to total revenues and profit margins. 

To the extent that traffic asymmetries exist between access providers in different regions, those traffic flows are mostly dictated by the location of content servers and the end user locations where people are requesting delivery of content. 

So whether one agrees that content delivery is a remote response to a local customer’s requests, or is an unrelated part of a single session, it is not so clear that ISPs literally have a broken business model. As content servers are deployed “closer to the edge” over time, asymmetrical data flows arguably could be reduced.

Video Calling Now Among the Top-three Mobile Phone Activities

Technology forecasts can be notoriously incorred, but on timing and adoption. Consider video calling, something that consumers do fairly regularly now.

Despite the many predictions about video calling, that use case was not a mass market and routine use case until recently, when the Covid pandemic forced people to start using it. 

AT&T, for example, demonstrated Picturephone at the 1964 World’s Fair, but adoption never happened. 

source: Time

Not until the Covid pandemic forced people to work from home and students to learn from home was there mass adoption of video calling services such as Zoom. 

A recent survey of user behavior in 2021, however, shows that video calling follows only instant messaging and voice calling as an activity on a mobile phone.   

source: GSMA Intelligence

How Much is Home Broadband About Physical Media?

Knowing what physical media is used by an access network does not necessarily tell one much about actual capacity or expected customer speed experiences, on any access network. Nor does physical media necessarily drive customer choices in an exclusive way. 

Personally, I’d buy a gigabit service provided by any network compared to an FTTH network supplying less capacity than that. Media does not matter, in that regard. Of course, price, upstream capacity and other issues play a part in such decisions. 

The point is that we sometimes fetishize FTTH, when we should be looking also at speed and other elements of the customer experience. Before FTTH became available, I’d assumed most people would prefer to buy it. In the abstract, that makes a good deal of sense: it’s the better network, right?

But price-value relationships matter. FTTH availability is one matter; buying decisions are driven by a much-wider set of considerations. 

Even though we conventionally assume fiber to home is much faster than copper access, with other platforms such as geostationary satellite, low earth orbit satellite, fixed wireless or hybrid fiber coax somewhere between copper and fiber home broadband platforms, FTTH networks can be activated at a range of speeds. In some cases, FTTH might not represent the fastest-available home broadband choice. 

So comparisons and targets are, in one sense, better evaluated in terms of speed capabilities and price-value relationships, matched by consumer buying behavior. What a policymaker wants is gigabit speeds or multi-gigabit per second speeds, not access media as such. 

There always seems a gap between customer preferences and internet service provider offers. In markets with strong cable operator competition, for example, FTTH tends to get between 40 percent penetration and 45 percent adoption after about three years of marketing. Some FTTH ISPs hope to reach a terminal adoption rate of 50 percent, but that is about the extent of expectations. 

source: IDATE, TelecomTV

Data from other European markets shows similar gaps between facilities deployment and take rates, where take rates hover between 45 percent and 47 percent. And that is a view of physical media choices, not necessarily speed tiers chosen by customers. 

In the U.S. markets, as well, many consumers choose not to buy the “fastest” tiers, but rather tiers someplace in the middle between fastest and slowest. 

source: OpenVault

The point is that enabling fast home broadband networks is one matter; customer demand is another matter. At any given point in time, it is likely that a majority of customers buy services in the middle ranges of capability; not the fastest and not the slowest. 

Consider U.K. fiber to premises networks, where “superfast” networks, by definition, operate at a minimum of 24 Mbps to 30 Mbps. Perhaps 42 percent of U.K. premises can buy FTTH-supplied home broadband. 

source: Uswitch 

Project Gigabit is a UK Government program aimed to bring £5 billion worth of investment to the country’s home broadband infrastructure. The aim is to bring gigabit-capable coverage to 85 percent of the U.K., and maximize coverage in the 20 percent of  hardest-to-reach locations by 2025. 

Based on past experience, it is safe to predict that, at some point, most customers will buy services at the gigabit per second level, just as most now buy services operating at about 30 Mbps. Just as safely, we can predict that, at some point, most customers will buy multi-gigabit per second services as well. 

We sometimes forget that during the dial-up era, people bought services topping out at perhaps 56 kbps in 1997. By 2000, typical speeds had climbed to 256 kbps; by 2002 reaching 2.544 Mbps. 

source: NCTA 

By 2005, typical speeds were in the 8 Mbps range; by 2007 speeds had climbed to about 16 Mbps. By about 2015 we began seeing advertised speeds of 1 Gbps. 

In all those eras save the dial-up period, the top speeds were not purchased by most people. Capabilities are important, to be sure. But consumer demand also matters. 

It is not necessarily a policy failure if most customers choose not to buy a particular product. 

source: Uswitch 

In competitive markets where gigabit alternatives are available on other platforms, FTTH take rates often hover around 40 percent of locations passed. If FTTH were clearly the superior choice, in terms of price-value, take rates would be higher. 

How that changes in the future is a reasonable question, especially in markets with facilities-based competition. In markets with but a single network provider, but multiple retail competitors using one network, FTTH take rates could be much higher, even if market share held by any single contestant 

Why Industry Rebranding Will Mostly Fail

"The more things change, the more they stay the same" might well apply to the connectivity business, despite all efforts to rebrand and reposition the industry as having a value proposition based on something more than "we connect you."

Consider the notion that 6G mobile networks will be about “experience,” as suggested by Interdigital and analysts at Omdia. At some level, this is the sort of advice and thinking we see all the time in business, where suppliers emphasize “solutions” rather than products and virtually all suppliers seek to position them as providers of higher-order value. 

The whole point of home broadband or smartphones with fast internet access is that those capabilities support the user experience of applications. 

And many have been talking about that concept for a while. “The end of communications services as we know them” is the way analysts at the IBM Institute for Business Value talk about 5G and edge computing, for example. 

To be sure, connectivity is not the only business where practitioners are advised to focus on end user benefits, solutions or experiences. But connectivity is among businesses where perceived value, though always said to be “essential” to modern life, also is challenged by robust competition and the ability to create product substitutes. 

One of the realities of the internet era is that although end user data consumption keeps climbing, monetization of that usage by communications service providers is problematic. Higher usage might lead to incremental revenue growth, but at a rate far less than the consumption rate of growth.

That is the opposite of the relationship between consumption and revenue in the voice era, when linear consumption growth automatically entailed linear revenue growth. Though there was some flat-rate charging, most of the revenue was generated by usage of long-distance calling services. 

“On the surface, exponential increases in scale would seem like a good thing for CSPs, but only if pricing keeps pace with the rate of expansion,” the institute says. “History and data suggest it will not.”

Indeed, Nokia Bell Labs researchers have been saying for some time that "creating time" 

is one way of illustrating the difference between core value propositions in the legacy and today’s market. “We connect you” has been the traditional value prop. But that could shift to something else as “connectivity” continues to face commoditization pressures. 

source: Nokia Bell Labs  

The growing business model issue in the internet era is that conventional illustrations of the computing stack refer only to the seven or so layers that pertain to the “computing” or “software” functions. Human beings have experiences at some level above the “applications” layer of the software stack, and business models reside above that. 

Likewise, physical layer communication networks and devices make up a layer zero that underpins the use of software that requires internet or IP network access. The typical illustration of how software is created using layers only pertains to software as a product. 

Software has to run on appliances or machines, and products are assembled or created using software, at additional layers above the seven-layer OSI or TCP/IP models, in other words.

So we might call physical infra and connectivity services as a “layer zero” that supports software layers one to seven. And software itself supports products, services and revenue models above layer seven of the software stack. 

Some humorously refer to “layer eight” as the human factors that shape the usefulness of software, for example. Obviously, whole whole business operating models can be envisioned using eight to 10 layers as well.  

source: Twinstate 

The point is that the OSI software stack only applies to the architecture for creating software. It does not claim to show the additional ways disaggregated and layered concepts apply to firms and industries. 

source: Dragon1 

Many have been using this general framework for many decades, in the sense of business outcomes driving information and computing architecture, platforms, hardware, software and communications requirements. 

source: Wikipedia

 

In a nutshell,  the connectivity industry’s core problem in the internet era is the relative commoditization of connectivity, compared to the perceived value create at other layers. 

Layer zero is bedeviled by nearly-free or very-low-cost core services that have developed over the last several decades as both competition and the internet have come to dominate the business context. 

Note that the Bell Labs illustration based the software stack on the use of “free Wi-Fi.” To be sure, Wi-Fi is not actually free. And internet connectivity still is required. But you get the idea: the whole stack (software and business) rests, in part, on connectivity that has become quite inexpensive, on an absolute basis or in terms of cost per bit.  

Hence the language shift from “we connect you” to other values. That might include productivity or experience, possibly shifting beyond sight and sound to other dimensions such as taste and touch. The point is that the industry will be searching for better ways to position its value beyond “we connect you.”

And all that speaks to the relative perception of the value of “connections.” As foundational and essential as that might be, “mere” connectivity is not viewed as an attractive value proposition going forward. 

It remains to be seen how effective such efforts will be. The other argument is that, to be viewed as supplying more value, firms must actually become suppliers of products and solutions with higher perceived value.

And that tends to mean getting into other parts of the value chain recognized to supply such value. If applications generally are deemed to drive higher financial valuations, for example, then firms have to migrate into those areas, if higher valuations are desired.

If applications are viewed as scaling faster, and producing more new revenue than connectivity, and if suppliers want to be such businesses, then ways should be sought to create more ownership of such assets. 

The core problem, as some might present it, is that the “experience” benefits are going to be supplied by the apps themselves, not the network transporting the bits. It is fine to suggest that value propositions extend beyond “connectivity.” 

source: Interdigital, Omdia

The recurring problem is that, in a world where layers exist, where functions are disaggregated, connectivity providers are hard pressed to become the suppliers of app and business layer value. So long as connectivity exists, value and experience drivers will reside at higher layers of the business stack. 

Unless connectivity providers become asset owners at the higher levels, they will not be able to produce the sensory and experience value that produces business benefits. Without such ownership, the value proposition remains what it has always been: “we connect you.”

If so, the rebranding will fail. Repositioning within the value chain, even if difficult, is required, if different outcomes are to be produced.

So long as humans view the primary communications industry value as "connections," all rebranding focusing on higher-order value will fail. It is the apps that will be given credit for supplying all sorts of new value.

How Much can Meta Compress its Hierarchy?

Meta’s effort to flatten management does raise logical questions about effective span of control in technology organizations. The span of control refers to the number of direct reports any single manager can effectively handle.  

In this simple illustration, a larger organization has multiple layers of direct reports, while a simple organization might have only one layer of direct reports. Any large organization is going to have a “tall” structure. 

source: OrgChart 

But the span of control is always limited to a small number of direct reports, usually described as topping out around seven people. Any large organization, therefore, is going to have lots of people who are essentially "managing managers." 

To be sure, spans might differ from firm to firm, based on the functional activities each engages in. The military span of control, for example, might be different from that of technology organizations.

The actual span of control arguably varies by the type of tasks to be managed. Highly unstructured work might have a limit of three to five direct reports. 

Highly-structured work, such as in call centers, might allow spans as large as 15 or more. The point is that the span of control is always sharply limited. So any large organization is going to have many people who are essentially managing other managers.

Can

How Much Can Tech Firms Flatten Span of Control?

Meta’s effort to flatten management does raise logical questions about effective span of control in technology organizations. The span of control refers to the number of direct reports any single manager can effectively work with.

To be sure, spans might differ from firm to firm, based on the functional activities each engages in. The military span of control, for example, might be different from that of technology organizations.

But span of control always is limited to a small number of direct reports, usually described as topping out around seven people. Any large organization, therefore, is going to have lots of people who are essentially "managing managers." 

The actual span of control arguably varies by the type of tasks to be managed. Highly unstructured work might have a limit of three to five direct reports. 

Highly-structured work, such as in call centers, might allow spans as large as 15 or more. The point is that spabn of control is always sharply limited. So any large organization is going to have many people who are essentially managing other managers.

Classic Platform Business Model Revenue Still a Science Project for Most Big Connectivity and Cloud Computing Firms

Network as a service, computing, storage or infrastructure as a service might easily be confused with a platform business model. After all, platform business models tend to involve use of remote or cloud computing, an application for ordering, provisioning, payments and customer service. So do many XaaS offerings. 

XaaS can provide value including reduced cost; greater agility and security that is maintained at industry leading levels. Sometimes XaaS also can provide advantages in terms of innovation or potential customer scale. 

But the difference in business models is not “buy versus build” or “virtualized” access, scale or innovation but the mechanism by which revenue is earned. A virtualized service offered by a “pipeline” business model provider is still an example of a traditional pipeline model: the seller creates the service and sells it to the customer. 

Amazon Web Services computing and storage functions, for example, are “sold as a service,” but that does not make those AWS products part of a platform business model. The Amazon Web Services Marketplace, on the other hand, is an example of a platform business model.

The marketplace supports transactions between third-party sellers to Amazon customers where Amazon earns a commission on each sale.

The general observation is that, at this point, though many firms are trying to add platform business model operations, those operations remain at a low level, compared to traditional pipeline operations. 

Classically, a platform earns revenue by earning a commission for arranging a match between buyer and seller. AT&T’s online marketplace, for example, allows third parties to offer internet of things products available for purchase from AT&T customers. 

AT&T also once hosted its own advertising platform Xander, which was sold to Microsoft. It allowed firms to place advertising on AT&T’s websites and apps. 

So far, revenue contributions have been small enough not to identify as distinct revenue streams. 

Likewise, Verizon once operated Verizon Media that placed ads on Verizon content assets, but that business was sold to Apollo Funds.

Some might consider the use of application programming interfaces evidence that a platform business model is in operation, but that is incorrect. APIs might be used to support a platform business model, but use of APIs, in and of itself, does not change the business model. 

APIs, though, are often a capability exploited by business model platforms, to connect users of the platform; to allow third-party developers to contribute value; to collect user data or to create revenue by charging fees for use of the APIs.

The GSMA Open Gateway initiative supporting APIs usable across networks supports a traditional pipeline model, where the firms create, support and sell their products directly to customers. 

So at least so far, few tier-one connectivity providers have shifted a significant portion of their operations to platform business models, or made it the key strategic direction. Recent asset dispositions by AT&T and Verizon suggest that approach remains experimental and non-core.