Congratulations (coming Sunday) to Denmark. Jonas Vingegaard is one step closer to winning the 2022 Tour de France after he extended his overall lead with a stunning stage 18 victory. A Danish rider last won the Tour de France in 1996.
In a head-to-head battle with defending champion Tadej Pogacar, Vingegaard waited for his rival after Pogacar crashed. Sportsmanship!
Gary Kim has been a digital infra analyst and journalist for more than 30 years, covering the business impact of technology, pre- and post-internet. He sees a similar evolution coming with AI. General-purpose technologies do not come along very often, but when they do, they change life, economies and industries.
Why, over the past couple of decades, have connectivity service providers been so concerned about over-the-top application providers? In large part, because the change from a closed telco value stack to an open internet value stack both devalues traditional telco products and enables new competition.
Put simply, “it is hard to compete with firms that give away what you sell.” It also is hard to compete with providers who are willing to sell “at cost.” Consider the way “triple-play bundling” has worked in the fixed networks business.
While it is true that both incumbents and attackers have used the bundling strategy, they have done so for different reasons. In nearly all cases, some services have been the equivalent of retailer “loss leaders.”
Telcos have been much more concerned with “merchandizing” video services to protect voice service revenues and profit margins, while cable TV operators have been more willing to merchandise voice services or mobile services to protect legacy video revenues.
These days, most fixed network service providers hope to merchandize almost anything to protect home broadband market share, revenue and profits.
Mobile service providers, on the other hand, want to protect broadband data access revenues and merchandize voice, text messaging, phones and video or content.
For communications service providers, the issue has arisen mostly in conjunction with low cost or “free” services such as Skype or WhatsApp that supply voice or messaging services “at no incremental cost,” once a user has suitable devices and Internet access.
That has led many to say the economics of abundance makes new revenue models possible. Some would say “abundance,” a relative term, makes new models essential, in at least some cases. But the implications are startling.
The basic idea is that transistors, storage, computation and bandwidth are so abundant the cost of their use--and digital products built on them-- is a price very close to zero.
The corollary is that businesses based on the use of such resources can be viewed differently from businesses where physical inputs are necessary and relatively expensive.
In other words, businesses based on abundant inputs can "waste" those resources. In a pre-broadband, pre-Internet-Protocol era, unicasting (content on demand) would have been nearly impossible. That is why video entertainment was broadcast or multicast.
The same sorts of economics apply to digital products whose cost of replication is quite low, in comparison to physical goods.
Or consider Facebook Messenger, or Skype or WhatsApp. It is the same general business problem: competitors who can afford to give away valuable apps and features, or price them very low, because their business models are built on some other revenue driver.
Incumbents in the New Zealand home broadband market lost about three percent account market share over the last year ending in March 2022, says IDC.
Notable share gainers include Trustpower, Contact Energy, Electric Kiwi and Nova Energy, energy retailers who now are bundling home broadband with their energy services.
“The energy retailers have a distinct advantage over the telcos; energy retailers don't need to make a profit on their broadband services,” says Monica Collier, IDC New Zealand researcher.
How to ”compete with free" is a major question in the Internet era, where many goods--especially of the non-tangible sort-- can be replicated and produced with low marginal cost.
Gary Kim has been a digital infra analyst and journalist for more than 30 years, covering the business impact of technology, pre- and post-internet. He sees a similar evolution coming with AI. General-purpose technologies do not come along very often, but when they do, they change life, economies and industries.
You might well question the payback model for new fiber-to-home networks which assume recurring revenue between $50 and $70 per account, per month, with little voice revenue and close to zero video revenue; take rates in the 40-percent range; and network capital investment between $800 and $1000 per passing and connection costs of perhaps $300 per customer.
But that is the growing reality. Among the reasons: higher government subsidies; indirect revenue contributions and a different investor base.
All that has shifted fiber-to-home business models in ways that might once have been thought impossible.
In the face of difficult average revenue per account metrics, co-investment and ancillary revenue contributions have become key. Additional subsidies for home broadband also will reduce FTTH deployment costs. All that matters as revenue expectations are far different from assumptions of two decades ago.
“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.”
Mobility postpaid phone ARPU at AT&T was $54.81. According to some studies, fiber-to-home recurring revenue is lower than that. But AT&T appears to be taking market share from key competitors where it has deployed new FTTH facilities.
Different investors also are becoming important for access infrastructure. Retail connectivity providers are judged by their ability to generate cash flows, but hampered by the huge capital investments they must make to do so. Institutional investors, on the other hand, have longer payback horizons. They value the predictable cash flow just as much as do telcos, but can afford to be more patient on payback.
Ongoing reductions in operating costs and complexity also play some role in lower breakeven points for connectivity provider access investments. Also, government support mechanisms can reduce deployment costs by as much as 30 percent, in some cases.
Lumen reports its fiber-to-home average revenue per user at about $58 per month. For those of you who have followed fiber-to-home payback models for any length of time, and especially for those of you who have followed FTTH for many decades, that level of ARPU might come as a shock.
Though some honest--and typically off the record--evaluations by some telco executives 25 years ago would have predicated the FTTH business model as “you get to keep your business” rather than revenue increases.
Few financial analysts would have been impressed.
The theory was that upgrading to FTTH would allow incumbent telcos to essentially trade market share with cable companies: gaining video subscription market share from cable as cable took voice share. The assumption was that home broadband share would remain about where it was.
The thinking was that per-home revenue could range as high as $130 to $200 per month, even as overall market share was gained by cable and lost by telco providers.
So the business case remains challenging, and especially so in less-dense areas. The main point is that firms deploying fiber to home facilities must do so with radically-reduced expectations for ARPU.
In recent investor presentations, Frontier Communications has made three points about its prospects for revenue growth based on optical fiber deployments: the number of consumer broadband accounts; the number of businesses within 250 feet of existing fiber assets and the number of cell towers within one mile of Frontier fiber assets.
Recent presentations also have shown fiber-to-home home broadband average revenue per user of about $63.
source: Frontier Communications
Two decades ago, the business model might have assumed far higher revenue per account. A telco or cable TV customer with a voice line generating $30 a month, plus internet plus video could have been worth about $100 a month in revenue.
Now Frontier says ARPU for an FTTH customer is about $63 a month. Assume that figure includes some amount of voice revenue and zero video revenue. The change in revenue expectation (not adjusted for inflation) per potential customer is roughly 40 percent lower than might have been the case in 1995.
Lower-density areas might only be upgraded with fixed wireless, though higher subsidy levels will increase deployment of FTTH in rural and lower-density areas.
For at least some observers, the change in FTTH business model asumptions is stunning.
Gary Kim has been a digital infra analyst and journalist for more than 30 years, covering the business impact of technology, pre- and post-internet. He sees a similar evolution coming with AI. General-purpose technologies do not come along very often, but when they do, they change life, economies and industries.
Incumbents in the New Zealand home broadband market lost about three percent account market share over the last year ending in March 2022, says IDC.
Notable share gainers include Trustpower, Contact Energy, Electric Kiwi and Nova Energy, energy retailers who now are bundling home broadband with their energy services.
“The energy retailers have a distinct advantage over the telcos; energy retailers don't need to make a profit on their broadband services,” says Monica Collier, IDC New Zealand researcher.
Where have you seen this story before? Fairly broadly in the consumer internet app and consumer access services business. The scary implication: “it is hard to compete with firms that give away what you sell.”
In other words, can you compete with "free?" That complaint lies at the heart of issues connectivity service providers have had with over the top edge providers or app providers.
How to ”compete with free" is a major question in the Internet era, where many goods--especially of the non-tangible sort-- can be replicated and produced with low marginal cost.
For communications service providers, the issue has arisen mostly in conjunction with low cost or “free” services such as Skype or WhatsApp that supply voice or messaging services “at no incremental cost,” once a user has suitable devices and Internet access.
That has led many to say the economics of abundance makes new revenue models possible. Some would say “abundance,” a relative term, makes new models essential, in at least some cases. But the implications are startling.
The basic idea is that transistors, storage, computation and bandwidth are so abundant the cost of their use--and digital products built on them-- is a price very close to zero.
The corollary is that businesses based on the use of such resources can be viewed differently from businesses where physical inputs are necessary and relatively expensive.
In other words, businesses based on abundant inputs can "waste" those resources. In a pre-broadband, pre-Internet-Protocol era, unicasting (content on demand) would have been nearly impossible. That is why video entertainment was broadcast or multicast.
The same sorts of economics apply to digital products whose cost of replication is quite low, in comparison to physical goods.
Gary Kim has been a digital infra analyst and journalist for more than 30 years, covering the business impact of technology, pre- and post-internet. He sees a similar evolution coming with AI. General-purpose technologies do not come along very often, but when they do, they change life, economies and industries.
As the European Union considers new laws that would make a few hyperscale app providers pay fees to providers of internet access infrastructure, seven EU countries warned the European Commission against any possible hasty decisions.
The letter was sent by the Netherlands, Denmark, Estonia, Finland, Ireland, Sweden, and Germany.
Telefónica, Deutsche Telekom, Vodafone and Orange are pushing the EC to mandate the new payments.
The European Union declaration on digital rights act calls for a framework where "all market players benefiting from the digital transformation…make a fair and proportionate contribution to the costs of public goods, services and infrastructures".
In other words, big app providers who impose most of the traffic demand on access networks should help pay for the access infrastructure.
It is not surprising that infrastructure cost burden sharing is sought by internet access providers, who argue they must invest based on demand created by third parties. Says Orange, “the investment burden must be shared in a more proportionate way.”
“Today, video streaming, gaming and social media originated by a few digital content platforms accounts for over 70 percent of all traffic running over the networks,” Orange argues. “ Digital platforms are profiting from hyper scaling business models at little cost while network operators shoulder the required investments in connectivity.”
It is not a new argument. Infrastructure cost sharing has been talked about since at least 2006.
The document also calls for “developing adequate frameworks so that all market actors benefiting from the digital transformation assume their social responsibilities and make a fair and proportionate contribution to the costs of public goods, services and infrastructures, for the benefit of all Europeans”
In other words, the argument is that access providers need some form of revenue sharing with the app providers using the networks.
We might take that as evidence that the connectivity service provider market is broken, or breaking.
If so, what is the reason and what are the solutions? Those questions are inevitable as the European Telecommunications Network Operators’ Association continues to argue before EU regulatory bodies that some form of revenue sharing (taxes, essentially) need to be levied on hyperscale app providers to make the service provider business model work.
The basic argument is that hyperscalers impose most of the costs of building and operating an access network, but that internet service providers cannot “recover their costs” because of asymmetric bargaining power with the hyperscalers.
Not to be flippant, but that is akin to electricity providers arguing that they do not have the power to levy fees on the manufacturers of air conditioning or heating systems, light bulbs or lamps to support the electricity delivery business model.
Electrical utilities recover such fees--and support their business models--by charging customers for consumption of electricity. Historically, so did telcos, though the role of government subsidies and price regulation also was key.
As always, the problem is complex and not simple to resolve. Large ISPs argue that their business models are unworkable without direct revenue contributions from third parties. That is the gist of the argument.
So the issue is how to remedy the problem. And ISPs might not like the eventual answers.
One potential answer is to nationalize the former incumbent telcos, making them government-owned entities again. Another answer is to allow the markets to return to monopoly structures by allowing most of the competing firms to fail.
Nor can big ISPs avoid questions about the soundness of their own cost structures; marketing practices and efficiency. Ignoring for the moment differences in market scale, U.S. ISPs between 2014 and 2020 reduced the amount of capital investment; European telcos saw capex burdens grow.
To be sure, some of the change might be transitory. Japan’s costs climbed while South Korea’s capex first climbed, then dropped, presumably because of the spike caused by early 5G deployment.
Still, one “easy” solution is for ISPs to charge their own customers based on volume consumed, as remains the case for electricity, fresh water, natural gas. In fact, nearly every commodity purchased by consumers is based on volume of consumption, whether that is groceries, gasoline or clothing.
Eventually, that basic line of reasoning might eventually lead back to price controls and rate regulation, which ISPs almost always oppose. But inability to charge based on volume of consumption is among the root causes of business model instability.
Gary Kim has been a digital infra analyst and journalist for more than 30 years, covering the business impact of technology, pre- and post-internet. He sees a similar evolution coming with AI. General-purpose technologies do not come along very often, but when they do, they change life, economies and industries.
Why are telcos and other connectivity providers potentially just like AOL? AOL, you might recall, was a walled garden provider of first party or “end user” services and features including chat, groups and email, as well as internet access itself, in the dial-up days.
But AOL’s business model fell apart once third-party apps and services were easily accessible on third-party home broadband networks.
Telcos used to be walled gardens as well, offering their customers fully-owned first-party apps and services including dial tone for voice; messaging and video entertainment. Early on, telcos also offered walled garden experiences as well.
But all that falls apart when any lawful app or service can be provided by a third party and consumed by any user with internet access.
And that shift to third-party, open apps in a loosely-coupled ecosystem explains why access providers worry about their business models, profit margins and revenues.
No more than happened to AOL’s business model, openness itself leads to third-party app creation and demand in the internet era. The winners in the scale app business are no imposing costs and limiting access provider revenue. The value of those apps creates the demand for internet access.
Access providers can sometimes acquire additional roles in the app area or other parts of the value chain (data centers, content ownership, cloud computing roles, for example), though it is not easy to do so.
Organic efforts are possible but hard to scale. Acquisitions provide scale but generally at the cost of high debt burdens.
But the backdrop is important. Just as walled garden AOL lost its business model when first-party experiences could be supplied by any other entity on the internet, so telcos faced the same problem when their own walled gardens were challenged by third-party experience providers.
There is no easy way to return to the walled gardens of the past, for most apps and services.
Gary Kim has been a digital infra analyst and journalist for more than 30 years, covering the business impact of technology, pre- and post-internet. He sees a similar evolution coming with AI. General-purpose technologies do not come along very often, but when they do, they change life, economies and industries.
Everyone expects that the working definition of broadband will continue to change over time. In the U.S. market, for example, the official national government definition is a minimum of 25 Mbps downstream and 3 Mbps upstream, with proposals to boost that minimum to 100 Mbps.
The real pushback has been against an immediate shift to symmetrical 100-Mbps speeds, which would imperil most ISPs in the upstream direction.
Of course, most people use internet access operating vastly faster than the current minimums.
If 98 percent of U.S. homes can buy internet access at the defined minimum from a cabled terrestrial network, that still leaves two percent that cannot, of course. Satellite and fixed wireless services reach virtually al lthe rest of the locations with service at the required minimums.
Changing the definition of “broadband” will likely increase the percentage of homes unable to purchase “broadband.” And that might boost the percentage of U.S. customers unable to purchase broadband access to as much as 16 percent or so of locations.
Of course, supply and demand are always changing.
About half of U.S. internet access customers buy services running between 200 Mbps and 400 Mbps as of June 2022. That is a shift. Until recently, about half of the customers purchased services running between 100 Mbps and 200 Mbps.
Roughly 70 percent of fixed network broadband customers purchase service at speeds of 200 Mbps or higher. Customers who buy gigabit or faster service have reached 13 percent, while customers of services operating between 500 Mbps and 900 Mbps are six percent of total.
The definition change largely will affect any ISPs who currently sell service that will not match the 100-Mbps definition, especially some satellite and telco ISPs who use copper access networks. Adjusting the definitions might mean some providers, at some locations, do not qualify for broadband subsidies.
In the U.S. market, for all the complaints we hear, gigabit speeds now are available to more than 88 percent of all U.S. homes, according to the Federal Communications Commission. Even if one disagrees with that estimate, most consumers in the U.S. market actually buy services operating far faster than the minimum.
Other estimates peg the percentage of homes with cable high-speed access at 90 percent. And reported uptake of gigabit speeds in rural areas is far higher than what most likely believe.
Consider rural telco networks. “Respondents to this year’s survey report an average of 4,467 residential and 469 business fixed broadband connections in service,” NTCA says, with an average of 7,581 serviceable locations.
“On average, three-quarters (75 percent) of serviceable locations are served by fiber to the home (FTTH) in 2021; this is an increase of 5.1 percentage points from the prior year’s survey, the latest Broadband/Internet Availability report issued by NTCA says.
The point is that “broadband” definitions have been changing for decades. 50 years ago, broadband was defined as any data rate of 1.5 Mbps or faster. 30 years ago the working definition had moved up to perhaps 10 Mbps. The current Federal Communications Commission definition is 25 Mbps.
But very few customers buy services operating that slow.
Gary Kim has been a digital infra analyst and journalist for more than 30 years, covering the business impact of technology, pre- and post-internet. He sees a similar evolution coming with AI. General-purpose technologies do not come along very often, but when they do, they change life, economies and industries.
