Too Early to Assess 5G Impact on ARPU

It is way too early to determine what impact 5G might have on service provider revenues, as account volume still is insignificant on a global scale. And while there are signs 5G might be boosting average revenue per account in South Korea, pricing policies still are developing in other markets. 

In markets were pricing is usage based, there might be some revenue lift from greater usage. In other markets, where there is no 5G price increment, ARPU might be affected in any number of ways. ARPU might drop, if competitive pressures are high.

In other cases, ARPU might remain flat, if simple substitution happens (5G replaces 4G). In yet other cases, ARPU could grow if usage-based pricing means 5G data consumption is meaningfully higher. 

source: Strategy Analytics

The meaningful revenue upside might come in consumer services at all, but in new higher-value services for businesses incorporating internet of things, edge computing, virtual private networks, artificial or augmented reality.

Price and Margin Compression Seem to be Permanent Conditions

If the global capacity business is an example, connectivity providers are having a very-difficult time adding enough value to raise prices. As Grant Kirkwood, Unitas Global founder quips, “prices always go down; they never go up.” 

That has been true for long distance voice calls, mobile minutes of use, text messaging or undersea transport. 

Source: Grant Kirkwood, Unitas Global

source: Recon Analytics

Likewise, look at European Union mobile roaming charges between 2013 and 2017. 

source: IREF

The other problem is that demand for legacy products is falling, with demand for new products rising. In the past, that has not been too difficult to manage. Data products sold to enterprises, for example, have gradually evolved over time, and continue to do so. But that is not an existential threat. 

Obviously, it is a bigger danger if markets evolve in the direction of rivals elsewhere in the value chain being able to supply the functions telcos traditionally have provided, themselves. And that was among the points made by Sean Bergin, AP Telecom co-founder. New markets like the internet of things will grow, but the value and share of revenue earned by connectivity providers is expected to decline from perhaps 23 percent of total ecosystem revenue to as little as five percent by 2025.

In part, that is because platform and application revenue will inevitably drive more of the total value of using IoT, as most of the economic value of electricity availability is driven by products using electricity; as most of the economic value of internal combustion engines is reaped by firms supplying products that use gasoline. 

Still, that issue of competition from over the top apps and services can be an opportunity or a threat, said Tony Mosley of Ocean Specialists. Partnering with an OTT sometimes can work, he said. Or, firms can move into other areas of the value chain beyond connectivity services. 

The PTC Academy provides management training to rising industry leaders across the Asia Pacific region and is held at various locales yearly, including the annual September event in Bangkok.

What Impact Might 5G Have on Wi-Fi Offload?

Smartphone data usage never is as simple and transparent as one would guess, since most users likely use Wi-Fi for most of the activity. And much hinges on the nature of data plans.

Where consumers routinely buy unlimited usage plans, Wi-Fi might represent half, or perhaps less, of total device data activity. That seems to be the case in the U.S. market, as all the major service providers have moved to support some form of “unlimited” usage on some plans. 

source: Opensignal

The issue has been what might happen in the 5G era, when mobile access speeds and latency performance are equal to or better than fixed access. So long as the tariff structure is similar, it seems inevitable that some market share will shift from fixed to mobile. 

On the other hand, at least some service providers might still see advantages to encouraging users to switch to Wi-Fi. The prime example is mobile virtural network operators who must buy wholesale capacity. In those cases, the service provider has a vested interest in seeing its customers switch to Wi-Fi whenever possible.

So switching effects will likely vary within any single market and between markets, depending on MVNO market share.

That might reverse the historic pattern of users switching to Wi-Fi from the mobile network when they can. In the earlier days of 4G, when tariffs were higher, switching to Wi-Fi was logical and saved users money. 

source: Mobidia

If 5G is available in unlimited use plans, so most users effectively never have to worry about how much data they are consuming, we are likely to see a reduction in use of Wi-Fi, compared to staying on the mobile network all the time.

Many Once Believed Millimeter Wave Would Not Work

Durga Malladi, Qualcomm SVP of engineering, explains how the once-ignored millimeter wave bands are going to enable many 5G use cases and applications.

FCC Asks for Input on CBRS PAL Bidding Rules

The Federal Communications Commission has opened a comment period on Citizens Broadband Radio Service application and bidding procedures. Of particular concern are rules for auctioning priority access licenses (PALs) in the 3550-3650 MHz portion of the 3.5 GHz band. The PAL licenses would offer quality of service protection for licensed users. 

Bidding in this auction, which is designated as Auction 105, is scheduled to commence on June 25, 2020. The FCC proposes to offer seven PALs in each county-based license area, for a total of 22,631 PALs nationwide. 

Each PAL will consist of a 10-megahertz unpaired channel assigned by automated frequency coordinators, known as Spectrum Access Systems. Consistent with the Commission’s rules, Priority Access Licensees would be permitted to aggregate up to four PALs in any license area.

Both private networks and carrier 5G services are expected to use CBRS, but it seems likely many private enterprise networks could use best-effort access, and even some service provider implementations may do so. 

Mobile Experts believes such private enterprise networks will use CBRS, as will service providers, to support their 5G services. Although the early thinking was that CBRS would use 4G transmission platforms, service providers now will be able to use CBRS spectrum to support their 5G networks. 

source: Mobile Experts

Consumer 5G May Not Require Edge Computing; Enterprise Use Cases Might

Edge computing, 5G and artificial intelligence are fundamentally inseparable, many would argue. The reason is that the advantages of ultra-low latency and ultra-high bandwidth apps and use cases hinges ultimately on use of all three underlying technologies, plus use of the new 5G virtualized core networks. 

Though 3GPP Release 15 standards are aimed primarily at supporting consumer smartphone use cases, Release 16 and 17 will extend 5G frameworks to private networks, IIoT (industrial IoT), and Internet of Vehicles (IoV), according to Jim Thompson, Qualcomm CTO. 

Those new enterprise applications will often require AI and edge computing, he said. 

CEOs of Qualcomm, Verizon and IBM discuss 5G

5G and edge computing are fundamentally intertwined, where it comes to creating enterprise customer value from mobile connectivity services. 

Will FCC Approve Some Form of C-Band Alliance Spectrum Sale?

One might conclude from remarks by Michael O’Rielly, Federal Communications Commission commissioner, that some form of the C-Band Alliance proposal for auctioning as much as 300 MHz of mid-band spectrum will be allowed to proceed. 

“I am so pleased to report that the FCC is near completion of its review process and is finalizing details for its reallocation, which should come later this fall,” he said. Reading between the lines, O’Rielly seems to suggest that a private auction will proceed.

“Given what is already in the pipeline and how long it takes for the Commission to set up and operate an auction, we are talking years, and I mean years, before completion,” he said.  

Significantly, O’Reilly said “we can certainly ensure transparency, accountability, fairness, and openness without having to run the auction ourselves.”

The argument has been made that the FCC should conduct a public auction for these frequencies rather than allowing the private sector to do it. Moreover, the argument has been made that it is unfair for these private, foreign satellite companies to receive all of the proceeds from any spectrum auction, private or public. 

“In the grand scheme of things, if it is a contest between speed and the government trying to extract a significant piece of the transaction through a lengthy process, I’ll take the speedy resolution,” he said. 

As George Ford, Phoenix Center chief economist has argued, a private transaction such as this is in the public interest. 

The largest, most promising source of U.S. mid-band spectrum suitable for repurposing to mobile wireless use is the C-Band, a 500 MHz swath of mid-band spectrum (3.7-4.2 GHz band) presently allocated for the provision of satellite communications, say T. Randolph Beard,  George S. Ford and Michael Stern of the Phoenix Center for Advanced Legal and Economic Public Policy Studies.

Satellite service providers, operating through the C-Band Alliance (“CBA”) consortium, have indicated that a large portion of this band could be quickly repurposed through a private sale, a proposal now being evaluated by the Federal Communications Commission. 

Perhaps the most compelling argument for a private transaction is the CBA’s proposed timeline, the researchers argue.  Assuming the FCC adopts, largely, the CBA proposal by year-end 2019, the Alliance claims it can conduct a private sale in the first-half of 2020 and clear at least 60 MHz for terrestrial operations in as little as eighteen months. 

Is such a private sale in the public interest? The researchers say it is. “It makes sense for the government to allow the consortium to serve as an agent in conducting the sale, thereby ensuring the rapid and efficacious repurposing of the band,” they argue. “For many reasons, a public auction is a less desirable approach” than a public auction.

“While the government may increase its revenue through a public auction, it does so only at the cost of an outsized reduction in total economic welfare,” they say. 

The satellite operators possess a significant information advantage over federal regulators with respect to how much spectrum can be repurposed, how much investment is required to accommodate the change, how to avoid interference problems, among other important considerations. 

This private information and the rapid pace of redeployment are both valuable and warrant compensation. Payment for information and efficiency is not “unjust.” 

Compensation to the agent (owners of the spectrum licenses) for its private information is efficient and not “unjust enrichment,” they argue. 

Third, considering the possibility that the principal may demand compensation from the agent, any such compensation (i.e., a regulatory payment) should take the form of a fixed payment rather than a share of auction proceeds. 

Fourth, while a public auction may be used to increase the government’s proceeds from repurposing the band, such enrichment is more than matched by a reduction in total economic welfare, they argue. 

The Commission is also actively pursuing allowing unlicensed wireless services in the 6 GHz band, which presenting provides C-Band uplinks.  

“This is a prime location for unlicensed spectrum because it can be combined with the 5 GHz and 5.9 GHz bands to provide the large spectrum channels needed to achieve 5G-like results,” said O’Reilly. “Opening this band to additional uses is doable as long as the incumbent protections are reasonable, but this is an example where clearing may not be feasible.”

But O’Reilly also suggests the FCC is looking at other bands, such as the 3.7 GHz to 4.2 GHz band, and perhaps 7.125 to 8.5 GHz, for mobile, indoor wireless or other use cases.  

End of Spectrum Scarcity in Latin America?

It often is hard to visualize the massive increase in spectrum 5G will bring. Consider today’s reality. In Latin America, for example, mobile spectrum available in each country.

Total spectrum for six markets (Argentina, Brazil, Chile, Colombia, Mexico and Peru) totaled 2,893 MHz at the end of 2018, up by 120 MHz compared to 2017. 

On average, each country has about 482 MHz each for mobile services by all providers in their countries. If one assumes three carriers in each market, a typical mobile operator might have access to about 161 MHz of capacity. 

source: S&P Global Market Intelligence

Additional low-band and mid-band spectrum auctions are scheduled, primarily of additional assets in 700 MHz, 2 GHz and 3.5 GHz regions. But the amount of raw bandwidth varies dramatically by frequency. Where in Chile an additional 20 MHz of capacity will be added at 700 MHz, and 30 MHz in the 2 GHz region, 50 MHz will be added at 3.5 GHz and 850 MHz in the 28 GHz region. 

source: S&P Global Market Intelligence

The reason millimeter wave spectrum is becoming more important is the vast amount of spectrum available, compared to all assets in the low-band and mid-band regions. Globally, millimeter wave millimeter wave will vastly increase capacity, by one or more orders of magnitude initially, by perhaps two orders of magnitude in the medium term. 

Staggering amounts of new spectrum are coming. In this illustration, note the present mobile assets at far left. Thickness of the bars indicates relative capacity. Then look at the thicker blue bands in the millimeter region, both proposed licensed and unlicensed spectrum assets the U.S. Federal Communications Commission plans to release for commercial use. Huge. Just huge. 

source: Anokiwave

Scarcity of spectrum, and therefore bandwidth, always has been a constraint in the mobile business. That might not be so much the case in the 5G and subsequent eras. 

5G Gets Better in the Real World

As with any major new platform, 5G--especially using millimeter wave spectrum--will progress. Handsets will perform better, radios will work better, operational knowledge will improve wtih experience. 

Millimeter Wave and Mid-Band are Focus of 5G Spectrum Growth, Everywhere

Any way one looks at the matter, huge amounts of new capacity are going to be made available for mobile and wireless communications in the 5G and subsequent eras. The reason is pretty simple: people keep using more data, which means more spectrum has to be freed up to support that ever-growing usage. 

Most of the activity right now is around the new millimeter wave bands and mid-band spectrum, as shown by spectrum moves in North and South America. 

source: GSA

How Long to 50% 5G Use?

Projections for 5G adoption generally assume demand will resemble 4G to a great extent. If so, we might expect 5G adoption to reach 50 percent adoption by about four to five years after full commercial launch in any single area or country.

 And some might note that adoption might not be especially driven by demand for higher speeds or new services, but simply by handset replacement cycles. Five years is long enough for just about any user to need to replace a device. And it is logical to expect service providers to push 5G-capable models and service plans as those devices are purchased. 

The 4G (Long Term Evolution) commercial launch began in 2008 with early deployments of data cards by TeliaSonera in Sweden in December 2009. Deployments in the United States and Europe started in 2010. Globally, it is worth noting, by 2019, nearly a decade later, adoption of 4G had reached about 53 percent of global connections. 

source: GSA

In any single market, 4G reached 50 percent adoption after about 3.5 years in the United States and Japan. Adoption elsewhere has taken a bit longer, perhaps five years in the United Kingdom, for example, and longer than that in other European countries. In Singapore 4G got higher adoption than 3G in about 17 months. 

source: Recon Analytics

The point is that it is reasonable to believe 5G might reach the 50-percent level of commercial use in any single country by about four to five years after full commercial launch. 

What Happens to Mobile Operator Capex as 5G Gets Built?

It is a truism that the next generation of mobile networks coexists with the prior generations for quite some time. That observation is true also for the “next generation” of fixed networks as well. Some might argue that often, if not always, leads to a “cap and grow” strategy, where investment in the legacy networks are curtailed in favor of building the next-generation platform.

Such transitions always pose a concrete problem for infrastructure suppliers, service and device suppliers. All would like to maximize revenue of the legacy products while new product sales volumes scale. Some might argue that especially is true for 5G, which actively builds on many investments in 4G, ranging from near-term use of the 4G core network and tower sites, as well as dynamic spectrum sharing allowing a 5G radio to use 4G spectrum. 

On the other hand, suppliers making such transitions often argue that the legacy products still offer high value. While logical and in supplier interest, projections for mobile network capital investment suggest the “cap and grow” strategy is precisely what will happen. 

GSMA Intelligence predicts that, on a global basis, mobile infrastructure spending is going to remain mostly flat through 2020. And history suggests that will not change post 2020, as more 5G networks are built.

source: GSMA Intelligence

source: IHS Markit

Many forecasts predict a decline in 4G investment, logical in light of the completion of most 4G network builds. 

source: Edge of the Cloud

And many rightly predict that 5G capex will start to grow by about 2021. 

source: Edge of the Cloud

The big takeaway is that if mobile infrastructure capex remains relatively flat, but 5G networks must be built, there has to be a shift from 4G capex to 5G capex. That is not to say 4G is going to lose its status on the driver of the majority of retail revenue. 

But in a flat capex environment, if funds are shifted to build 5G, 4G investment has to drop.

5G Use Cases Will Develop, as Did 3G and 4G Use Cases

It can be argued that 3G and 4G followed a path to commercialization that reflects the value proposition at various points along the adoption cycle. Early on, for both 3G and 4G, there were no devices capable of using the new network other than data dongles or sticks. So the first use case was mobile internet access for PCs and other devices. 

AT&T is following that playbook exactly, selling its early 5G services only to business customers, and only for fixed wireless data access, typically as a backup to a fixed connection. 5G will add one important new wrinkle: data rates high enough to function as a substitute for fixed network connections (5G fixed wireless). 

As the networks launch nationwide, and new 4G-capable devices are made available, the biggest use case will be consumer smartphones. Only later will potential applications develop that take advantage of the bleeding edge of 5G capabilities (ultra-low latency and gigabit speeds). 

source: Sierra Wireless

Stand-Alone 5G Might Reduce 3G, 4G Capacity

Analyst Zahid Ghadialy discusses 5G supplied in standalone (SA) and non-standalone (NSA) modes. He points out that refarming of 3G and 4G could reduce capacity, and therefore possibly performance, of 3G and 4G networks.

Can Fixed Wireless Get 15% to 20% Consumer Internet Access Share?

Verizon now says its fixed wireless offer will be available in all 30 markets nationally where Verizon is introducing its mobile 5G service.  “You should expect that every market that opens a 5G Mobility market will in due course be a 5G fixed wireless because it is one network,” said Ronan Dunne, Verizon Communication EVP. 

And then we will see whether forecasts of significant adoption are proven correct. The question is whether fixed wireless might eventually represent 15 percent to 20 percent of active subscriptions supplied by fixed network ISPs, a figure some have suggested is a possibility. 

That is deemed implausible by some. Maybe not. A significant role for fixed wireless is more likely if the actual universe of fixed network accounts is smaller than generally assumed. 

By some estimates had U.S. fixed network broadband accounts were as high as 110 million in 2017. Other estimates peg the number of connected locations at about 100 million. By other estimates, there are a total of about 97 million U.S. internet access connections supplied by fixed networks (not including mobile internet access connections to phones and other devices).

Those figures appear to include business broadband accounts as well as consumer accounts, however. On the other hand, it is possible that business use of fixed wireless might grow the addressable market to the extent that fixed wireless is viewed as a suitable backup service. AT&T, in fact, has been marketing its 5G fixed wireless service exclusively to business customers, who do seem to use it as a backup to their fixed access connection. 

The potential universe for consumer fixed network internet might be smaller than often assumed. Some households do not wish to buy the service. Estimates by Pew Research suggest only 89 percent of U.S. residents use the internet at all. 

And mobile substitution is a growing reality for 15 percent to 20 percent of U.S. households, according to Deloitte. The point is that the total addressable market for fixed network internet access might be as small as 82 million homes or as large as 88 million. 

source: Pew Research Center

source: Deloitte

No all housing units are occupied. Vacancy rates can range from more than one percent for owned housing and up to seven percent for rental units. Some “homes” are not used all year, as a matter of course.  

U.S. housing locations number somewhere between 129 million and perhaps 135 million. In 2014 the U.S. Census Bureau estimated housing units at a bit over 133 million, so the 135 million estimate seems reasonable. 

Assume 10 percent of locations choose not to buy fixed network internet access because they do not use the internet. That suggests the addressable market is 121.5 million locations. 

And perhaps 10 percent of U.S. homes use the internet, but exclusively from a mobile device. Remove the mobile-only households and 13.5 million more units get removed from the potential universe of buyers, reducing the market to a potential of about 108 locations. 

But also keep in mind that some units are rooms in houses, boats, mobile homes or other units possibly not suitable for fixed internet access. Mobile homes represent about 6.4 percent of housing locations. About three percent of people live in group quarters, vans, RVs, and boats, according to 2014 U.S. Census Bureau data. So remove another six percent of locations from the addressable universe. That subtracts about eight million more locations, for a total addressable market of 100 residential locations. 

If fixed wireless might eventually represent 15 percent to 20 percent of active subscriptions supplied by fixed network ISPs, and the addressable market is 15 million to 20 million locations, then fixed wireless could grab 15 percent to 20 percent market share. 

Skeptics say Verizon’s fixed wireless effort so far is quite miniscule. True, but it is worth remembering that Verizon has deliberately chosen to restrict its roll-out, partly for marketing reasons (it wants to test demand), partly for technology reasons (it is waiting for full standards-based customer premises equipment and network gear), partly for operational reasons (Verizon will deploy, logically, where it faces the greatest capacity demand, and that typically is urban core areas, not suburban areas where many homes are located). 

Also, Verizon’s use of millimeter wave spectrum for mass market services is quite novel. As with any radically-new technology, there is a learning curve. 

source: Business Insider

So the spotty coverage is not too surprising. The careful roll-out is not unusual. And Verizon is coming up the learning curve. The early deployments of fixed wireless were careful and over-engineered, with install costs too high to sustain as a mass market service, using pre-standard, almost custom network elements. 

But what has Verizon learned? Where it had to install external antennas early on, it now finds that “almost 80 percent of the antenna are indoors rather than outdoors,” said Dunne. “And that's critical for them, the ability to self-provision.” Install costs will plummet, as a result. 

The early customer premises equipment was essentially using a cellphone chipset with low power output. The production-level gear now coming uses higher-power chipsets, available in the first half of 2020, will feature higher-power, meaning easier self-installs in a wider range of settings where the low-power gear might have encountered issues. 

“We're building a mobility network that enables a 5G residential (fixed) offering, we're now in a position where we can lean into that,” said Dunne. The fixed wireless network, in other words, is not an overlay network on the 5G mobility platform, but part of it. And that also will help manage costs as the fixed internet access service becomes a mass market offering.