Why Mobile Data Supply Has to Increase 10X

By the end of 2018 average (“mean,” I believe) U.S. household data consumption on fixed networks was 268.7 gigabytes (I believe the figure is “per month”), according to Openvault, up about 19 percent from mid-year and up 33 percent from 2017 levels. So it seems likely that consumption will have grown at least 33 percent in 2019, and most likely higher. 

Assume an average of 2.5 persons per household, each using the average amount of data. That suggests a mobile data consumption per household around 27 GB per month. But spreading use of “unlimited usage” plans likely will push consumption higher. 

Customers on unlimited plans consume 67 percent more mobile data than consumers on usage-based plans in 2017, according to NPD.

Average monthly mobile data usage is around nine gigabytes per month. In a “typical” U.S. household of perhaps 2.5 people, that works out to about 27 GB per month of mobile data consumption, “per household,” on the mobile networks. 

Others believe consumption is lower, at perhaps six gigabytes per user, per month, according to Strategy Analytics.

The issue is how high mobile data consumption might go. Ericsson believes it is possible U.S. mobile customers could be consuming 39 GB each, per month, by about 2024. In a “typical” household of 2.5 people, that implies potential “household” mobile data consumption of nearly 98 GB per month. 

source: Ericsson

The implication is that mobile networks will have to be designed to carry between a quarter and a third of the data now expected in a typical U.S. household. That is a big jump from present levels. If the typical mobile user consumes 9 GB, and a household 268 GB, then mobile traffic is about three percent of fixed network levels.

That, in turn, implies an increase of mobile data capability by about an order of magnitude by 2024. Hence 5G, which will increase capacity by a minimum of 10 times over 4G networks. 

Consumers Confused about 5G?

About 66 percent of a survey of consumers suggests they believe they understand what 5G is, but 20 percent of those consumers thought they already had it, Strategy Analytics reports. That suggests that although some form of 5G awareness is relatively high in the U.S. market, consumers also are confused about it. 

While 25 percent of U.S. consumers surveyed by Strategy Analytics said 5G is an important feature, 20 percent did not see the value, or wanted to wait until the benefits of 5G were proven before purchasing this technology.

source: Strategy Analytics

Whether that is a big problem or only an expected early annoyance is not so clear. Consumers might not find the “”it’s faster” value proposition hard to grasp. What is less clear is whether consumers will find the additional speed useful and obvious, since, with the exception of big file downloads, users might not be able to perceive much difference in their use of most apps. 

Some of us might argue that the perceived benefits will come only when other technologies, such as network slicing and edge computing are widely deployed.

6G?

Nobody knows yet what 6G specifications might be, but some directional insight can be gleaned by looking at where the 5G roadmap leads. 

Release 16, expected about 2020, will include support for NR-Unlicensed (NR-U), Integrated Access and Backhaul (IAB), enhanced Vehicle-to-Everything (eV2X), URLLC and Industrial IoT (IIoT) enhancements, and Service Enabler Architecture Layer (SEAL) for verticals.

Release 17 features might include support for Non-terrestrial Networks (NTNs) (satellites), new frequency bands (7 GHz to 24 GHz and frequencies up to 53 GHz), and NR-Light.

Some believe 6G could offer high-fidelity holograms, multi-sensory communications (touch, taste and/or smell), TeraHertz (THz) communications, and pervasive Artificial Intelligence. Of these, only multi-sensory capabilities seem unlikely. 

Every next-generation mobile network since 2G has reduced latency, improved speed and increased bandwidth efficiency. That is about the only safe prediction we can make about 6G. Use cases and business models have been the areas of greatest forecast error in the 3G and 4G eras, and likely will happen in the 5G era as well.

That suggests high caution about anything other than technical performance predictions for 6G. 

source: Rohde and Schwarz

Much of the work extending the 5G standards will focus on adapting 5G for vertical market applications, including standards for using 5G as a full-fledged local area network in either enterprise or consumer use cases, enabling satellite networks to interwork with 5G, adding connected vehicle and unmanned aerial vehicle support. 

Critical medical use cases, audiovisual production, manufacturing and utility grid apps, location-based services and haptic feedback standards will be created. 

source: Rohde and Schwarz

No "Wrong Choice" of 5G Spectrum, Only Misguided Analysis

Over the past year there have been many shouts of alarm about the “choices” being made for 5G spectrum. Millimeter is the wrong spectrum, it has been said. Mid-band spectrum is the “right choice,” many have argued. 

Those criticisms are rather off point. Different countries are making different immediate choices largely because--of a range of permissible frequencies--mid-band is available. 

For historical reasons, the mid-band is not immediately available in the U.S. market, forcing early movers to rely on millimeter wave spectrum sooner than they might have preferred, even if the 5G standards clearly point to millimeter wave as the future of 5G, and subsequent platforms as well, simply because that is where most of the unencumbered spectrum exists. 

The U.S. Federal Communications Commission is not ignorant. It knows what resources can be made available now, and what has to happen to clear more mid-band spectrum. It is doing so. Clearing part of the C-band is among the actions the FCC is taking. But that takes time. 

Verizon, in particular, has had to rely on millimeter wave. Among the top four national carriers, it has the least spectrum, per customer. And after weighing its options for bandwidth, Verizon concluded that upgrading its terrestrial network with dense optical fiber, enabling small cells and hence millimeter wave radio networks, made more sense than bidding for lots of new spectrum. 

That does not mean Verizon or the other service providers will be shy about bidding on additional spectrum. “More” is always the answer, longer term. 

Still, Verizon believes the cost of its dense fiber network approach will work for capacity expansion. Verizon surely will rely on mid-band for coverage. AT&T initially relied on millimeter wave for its business-focused services, especially for fixed network substitution. 

Its consumer 5G will use 850 MHz low-band spectrum, and AT&T will acquire more mid-band spectrum when it is made available. 

T-Mobile, with relative plentiful new 600-MHz assets, will rely on low-band for its 5G launch. Sprint has lots of mid-band spectrum licenses, which will eventually be put to work by whatever company winds up owning it. 

The larger point is that no dangerous or wrong spectrum choices have been made by the FCC or service providers. Specific firms have made choices congruent with their assets and strategies. Long term, there is no “choice” to be made between millimeter wave and mid-band spectrum or low-band. All will be used.

But different service providers in different countries have differential access to low-band or mid-band spectrum. So the initial strategies and deployments will reflect those immediate realities. Longer term, all the choices will be in play.

Mobile Cost Per Bit Has to Keep Dropping

Mobile networks increasingly must be built to support cost per bit metrics very close to fixed networks, something that might in past decades have been thought impossible. And they must do so because the dominant driver of bandwidth consumption now is video, the media type with the lowest revenue per bit implications for most mobile operators. 

In other words, as a practical matter, lower costs per bit are a necessity as traffic shifts overwhelmingly to video content, forcing operators to supply ever-more bandwidth, but earning little additional revenue, if any, as consumer propensity and ability to pay are constrained. 

source: Ericsson

In past years, mobile cost per bit has been an order of magnitude to two orders of magnitude higher than on fixed networks. That becomes a big business issue as consumption on mobile networks begins to resemble fixed network levels (especially for 5G fixed networks that compete directly with fixed networks). 

Verizon executives have noted that they have been slashing mobile cost per bit by as much as 40 percent per year. At such rates, it is conceivable that mobile data costs, in many cases, could be as little as 10 percent of 2018 costs, clearly an order of magnitude reduction. 

In 2010, for example, U.S. mobile data costs in some cases were as high as $25 per gigabyte. By 2015, mobile data costs had dropped to perhaps $5 per gigabyte. Since then, costs arguably have declined further, and should continue to drop in the 5G era as well. 

source: 5G Americas

There long has been a huge disparity of orders of magnitude between revenue bit bit earned for texting, voice, internet access or video subscription services. 

Text messaging has the highest revenue per bit, possibly six orders of magnitude higher than internet access, in some instances, in Europe. Mobile voice revenue per bit is perhaps two to three orders of magnitude higher than internet access, in Europe. 

Subscription video sold as a managed service might have revenue that in principle is higher than internet access at modest levels of usage, or lower revenue per bit at high levels of usage. 

source: 3G4G blog

The big takeaway is that the relentless drive to reduce mobile data cost per bit is mandatory if mobile operators are to support growing mobile data consumption under conditions where consumers really cannot spend much more as they consume more data.

Price Premium for 5G? Perhaps Not

Early evidence suggests 5G might carry a price premium up to 20 percent, according to the Ericsson Mobility Report. That might be the case in some markets, including South Korea, where there has been at least a little revenue lift since 5G was launched. 

The difficulty is that 5G average revenue per unit might be affected by any number of other consumer choices. Perhaps consumers upgrade to unlimited usage plans, which normally are the most expensive. In such cases, it is not necessarily the 5G pricing premium but the higher-priced unlimited plans which drive revenue increases.

It is not an easy matter to predict whether 5G will produce higher average revenue per account in the U.S. market. For starters, not all the leading service providers will charge a price premium for 5G. T-Mobile US continues to say it will not charge any price premium for 5G. AT&T has not yet unveiled consumer 5G pricing. 

And Verizon and Sprint both charge a $10-a-month premium for 5G access, as part of their unlimited plans, including all Verizon unlimited data usage plans and Sprint’s top unlimited usage plan. 

Though T-Mobile US says it will not charge a price premium for customers of its 5G network, both Verizon and Sprint are charging $10 a month more for 5G, on the top-end unlimited plans, in Sprint’s case, or on any unlimited data usage plan in Verizon’s case. That means there is some potential for revenue increases on 5G plans, for Sprint and Verizon, to the extent that consumers choose to buy unlimited-usage plans. 

Sprint’s 5G service is reserved only for its $80-per-month "unlimited premium" customers. The 5G network will not be available for use to Sprint customers buying $70-a-month or $60-a-month unlimited plans. 

Access to Verizon’s 5G Ultra Wideband network is $10 per month with any of Verizon’s unlimited usage plans. 

Beyond that, AT&T provides access to the 5G network as a feature of its most-expensive unlimited usage plans. So a consumer switch to one of the unlimited-usage plans creates higher revenue, though it is not necessarily 5G that drives the decision. In fact, according to AT&T, there is no price premium for 5G, in a direct sense, on such plans. Consumers pay more for unlimited usage. 

Early experience with 4G probably is relevant. At first, service providers thought they might be able to charge a small premium for 4G. In the U.S. market, that rapidly became impossible, as competition heated up quickly, eliminating the chance to create premium pricing for 4G. It quickly became the case that 4G was cheaper than 3G, at least on a cost per gigabyte of usage basis. 

So one might argue that 5G, as such, will not generally create opportunity for a direct increase in prices. Instead, users might find they pay more for unlimited usage plans, getting 5G as a feature of such plans.

Mobile Bandwidth Consumption Increased 78% Last Year

Though there are other reasons why 5G might be advantageous for mobile service providers, customers and app providers, the practical reason for 5G is simply that, with global bandwidth consumption growing 78 percent, year over year, 4G networks simply will be unable to keep up.

source: Ericsson

Though 4G can be, and will be, improved, in terms of bandwidth capabilities, there are some clear limits. A few technologies, better radios and small cell architectures are key to that effort.

Among the technology improvements, many would point to MIMO radios, 256 QAM for more intense modulation and carrier aggregation (for 60 MHz bandwidth or more) as ways to get to gigabit 4G.

The 3GPP standards for 4G support aggregating up to 32 carriers for LTE. Not only does that create bigger channels that support more bandwidth, and therefore higher speeds, but doing so also increases efficiency, eliminating bandwidth that otherwise would be wasted for guard bands. 

Operators are deploying massive MIMO antenna configurations with up to 128 antenna elements (64 transmit and 64 receive), which increases bandwidth by essentially creating multiple parallel channels where a single-antenna system has but one channel. 

Radios also are using more heavily sectorized radios, which supply more effective bandwidth by reducing interference. Where traditionally cell tower radios have used three sectors, 4G cell sites now use six sectors or nine sectors.

Denser networks using smaller cells also are part of the solution, creating more effective bandwidth by more intensively reusing existing spectrum. 

So one might ask why not simply use 4G and add more new spectrum. Part of the issue is that the 4G standard only includes frequencies up to 5 GHz, not any frequencies above that range. Standards-compliant network elements and platforms therefore cannot be purchased to support spectrum above 5 GHz.

Mid-Band 5G Spectrum Allows U.K. Speeds 4X to 6.7X Faster than 4G, It Appears

Mid-band spectrum is touted for 5G as it offers better coverage than millimeter wave spectrum but higher bandwidth than low band frequencies. The latest Speedtest data on U.K. 5G show the results. Compared to “typical” speeds on all other U.K. mobile networks taken as a whole, 5G provides downstream speeds about 450 percent to 618 percent faster, according to Speedtest. 

Keep in mind: this does not mean 5G is that much faster than 4G. 5G is that much faster than the average of downstream data on all other networks in use. On the other hand, 5G arguably is just about that much faster than 4G. 

source: Speedtest

On EE, it is conceivable that 5G is about 400 percent faster than 4G. On 02, 5G might be 760 percent faster than 4G. 

source: Speedtest

5G Gains: More Supplier Push than Consumer Pull

Supplier push rather than customer demand is going to drive early 5G subscriptions, since the experience advantages are going to be quite intangible in most cases. In many markets where 5G is launched using low-band spectrum, speed will not, in fact, be much different than fast 4G.

In other markets, where mid-band or millimeter wave spectrum also is used, consumer use cases will not be able to take advantage of the additional speed, with a few exceptions such as big file downloads, virtual reality or augmented reality. But how much time do consumers spend downloading big files? How many use AR or VR already?

And yet, 5G will be adopted. We do not know whether adoption will be faster, slower or at the same pace as 4G, but 5G will be purchased. One big reason is that 5G--though having eventual advantages for most consumers--has bigger advantages for mobile operators. 

Even if consumers might not experience much benefit at first, mobile operators will. 5G features lower cost per bit than 4G, which helps prop up the business model when more capacity has to be supplied at a relatively fixed price. 

Over the long term, consumers can only spend so much money on their communication services, so wallets budge only slowly, if at all. 

5G also has network features helpful either for cost containment or new service creation. 5G networks will be virtualized, which creates the ability to turn up virtual private networks (network slicing) rather easily, compared to legacy methods. 

In the radio network, virtualization allows mobile operators to contain radio network costs, since equipment from different suppliers increasingly will be mixed and matched with core networks. 

Longer term, 5G latency performance opens up potential space for a role in edge computing networks, vertical market applications and partnerships. 

And, as always, for some suppliers in each market, 5G represents a chance to preserve or upset existing market shares, if competitors cannot easily keep pace. 

Hence the paradox: Though 5G is, by design, supposed to have performance advantages over 4G for mobile service provides and customers, the early advantages will rarely be sufficient to drive consumer demand. Instead, supplier push will be at work. 

For customers, the advantages include faster top speeds and lower latency.  For mobile operators, the advantages include faster top speeds, lower latency, virtualized networks, which promise new features and lower cost, higher device density (supporting lots of sensors and internet of things devices) and lower cost per bit. 

That list is a clue to why 5G will be adopted, perhaps even more rapidly than 4G: it benefits mobile operators more than consumers. In fact, an argument might be made that 5G benefits mobile operators almost to the exclusion of consumers, at first. In the early going, using low-band spectrum, 5G might not be noticeably faster than 4G. 

When mid-band and millimeter wave spectrum is used, speeds will be much faster than 4G, but in ways that do not actually benefit the use cases most people have for their phones. Big file downloads will be faster, but what percentage of time do people spend downloading big files? 

At first, the early adopter desire to "get it first," where the driver is as much status as anything else, will be the driver. But there will be indirect drivers as well, such as the value of "5G comes with your new service plan." In many cases, 5G will be a "nice to have" attribute of a service plan, but it is the service plan that drives the switch. 

Perhaps that will change over time, as new use cases develop. But it might also be the case that 5G gets adopted because it provides value for mobile operators, who will create incentives to adopt 5G, even if the actual experience advantages might be hard to demonstrate. 

This can be seen in recent Opensignal measurements of speeds on new 5G networks. Where early tests of U.S. 5G rely exclusively on millimeter wave spectrum, the U.S. has the highest speed. 

Aside from big file downloads, the experience advantages are almost impossible to demonstrate. Faster might be better, but actual consumer smartphone apps cannot actually take advantage of the faster speeds, yet, since virtual reality and augmented reality are not yet widely used. 

Conversely, not that in some markets where low-band 5G has been launched, the speeds are almost identical to 4G, so there is no actual experience advantage. 

source: Opensignal

Compared to existing 4G, 5G in early days has doubled to nearly tripled real-world speeds in some cases, but had almost no impact on speed in a few cases. Again, the choice of spectrum, or availability, really do matter. 

Where low-band spectrum is the 5G choice, it sometimes does not life speeds very much. Where millimeter or mid-band spectrum is used, 5G speed advantages are clear, but app performance and user experience do not change much, with the exception of big file downloads.

source: Opensignal

AT&T 5G Launch Illustrates Demand-Creation Issue

AT&T plans to launch consumer 5G services in the first half of 2020, after a year of selling only to businesses. For mass market customers, AT&T plans to rely on low-band spectrum, using millimeter wave high-band spectrum for capacity reinforcement at large venues. 

AT&T initially is focusing on coverage, rather than capacity, so although consumers in the new markets might see 5G speed boosts over 4G levels, they will not see the highest speeds available on the millimeter wave networks. 

In fact, AT&T executives downplay the speed differences. “You’re going to see some marginal differences,” says Igal Elbaz, AT&T SVP. 

So one of the emerging angles on 5G is that 5G using low-band should not be expected to offer speeds much faster than advanced 4G. 5G using mid-band or millimeter wave spectrum potentially can boost speeds far above 4G, but few consumer use cases exist. 

It all adds up to a demand stimulation issue. Service providers will have to convince customers there are reasons to switch to 5G, even when the value proposition is questionable. On low-band networks, speed is not very different from advanced 4G. On mid-band and millimeter wave networks, speed is much higher, but few use cases now exist. 

If I had to guess right now, I’d predict that much early consumer adoption of mobile 5G will be driven indirectly, the result of choices made about service plans and high-end phones, not 5G service specifically. 

There are, of course, value drivers unrelated to technical performance, namely the early adopter desire to first use a new product with the best specs. 

In fact, in some markets, where 5G is based on low-band spectrum, the speed advantage of 5G might be quite muted and hardly better than 4G.

In the United Kingdom and Spain, where 5G initially was launched using low-band spectrum, there was only a slight improvement in experienced internet access speeds. Under those conditions, it is hard to say what the direct value proposition might be. 

source: Opensignal

In markets where 5G initially is launched using mid-band spectrum, consumers might have a relatively straightforward value proposition: slightly higher cost, significantly-higher bandwidth, even if the use cases remain few. 

The point is that, in all markets, most consumers might struggle to find a direct value proposition unrelated to status or bragging rights, in the early going, as there are very-few consumer mobile applications or use cases that benefit from near-gigabit-per-second speeds, or even average speeds in the 100-Mbps range. 

Consumers who do lots of downloading on their mobiles are about the only users who will--at first--be able to take advantage of higher 5G speeds on mid-band or millimeter wave networks. 

Ignoring for the moment the few use cases where 5G speeds in the 100-Mbps to near-gigabit range are relevant, indirect drivers are likely going to be more important than direct value propositions, for most customers. 

The consumer direct value proposition might be low-band 5G, which offers slight or possibly even no real advantage over 4G; mid-band 5G, which offers quantifiable speed advantages, but few experience advantages; or high-band 5G, which offers huge speed advantages, but few experience benefits over 4G. 

Beyond that, some customers, including those in the U.S. market, will have to choose between 5G phones supporting low-band 5G or millimeter wave 5G. The early devices will not support both low-band and high-band. That requires a device using the Qualcomm X55 modem, expected to be available in late 2019, but not available in the first generation of 5G devices. 

All of that suggests to me that indirect value will drive most 5G subscriptions. Customers will get 5G as a byproduct of their service plan or device choices. 

5G is Going to Create High Volatility

While low-band spectrum is universally acknowledged as the best “coverage” spectrum, millimeter wave is the best when “capacity” is required. And though many say “mid-band” is the best balance of coverage and capacity, even mid-band assets will not affect overall bandwidth capabilities as millimeter wave spectrum. 

This can be seen in recent Opensignal measurements of speeds on new 5G networks. Where early tests of U.S. 5G rely exclusively on millimeter wave spectrum, the U.S. has the highest speed.

Australia, Switzerland and South Korea use mid-band spectrum and show gigabit speeds, but not as fast as millimeter, as theory suggests will be the case. 

source: Opensignal

Countries relying on low-band spectrum have lower speeds. 

Compared to existing 4G, 5G in early days has doubled to nearly tripled real-world speeds in some cases, but had almost no impact on speed in a few cases. Again, the choice of spectrum, or availability, really do matter. Where low-band spectrum is the 5G choice, it sometimes does not life speeds very much. 

source: Opensignal

Such choices should be kept in mind as we enter a period of potentially highly-volatile changes in average speed differentials between countries, based on whether 5G is adopted, or not, and which spectrum assets are widely used. 

Old rules of thumb are likely to be upended in the first several years of the transition to 5G, as some nations move earlier, while others wait, and as reliance on millimeter wave varies from place to place. 

As always, we are reminded to take global rankings with a bit of skepticism. Things change, where it comes to internet access adoption, speed and cost. In that regard, 5G is going to be highly disruptive.