Thursday, March 23, 2017

Internet Ecosystem Power Already Has Shifted to App Providers; Net Neutrality Does Not Matter, in That Sense

The arguments for strong forms of network neutrality have assumed that, in the absence of rules barring any levels of consumer internet access other than “best effort,” access providers would exercise market power in ways that would stifle innovation on the part of app and content providers.

Ignoring for the moment the countervailing argument that some apps and services might actually require quality of service mechanisms, or that consumers should have the right to choose such QoS-based services, if they choose, it has never been completely clear that innovation or business success--for any app provider--is fundamentally conditioned by the nature of internet access policies.

The fortunes of competitors to Google, Facebook and others is logically not dependent on access rules that might include optional quality of service mechanisms, but on the end users’ preference for those leading apps, and the ability of those firms to keep innovating.

It never has been so clear to some of us that the availability of managed services (under conditions when best effort access is protected by weak forms of net neutrality rules, such as access by all consumers to any lawful app) is some sort of major inhibitor of innovation in the apps space.

In fact, virtually all observers would note that value and pricing power within the internet ecosystem  now decisively rests with the application layer providers, which is why access providers always are concerned about their roles as providers of “dumb pipe” internet access.

So long as best effort access is the norm for consumer services, it never has seemed so likely that managed service availability, in and of itself, would shape app provider fortunes. Widespread use of content delivery networks (QoS services in the backbone) do not seem to have affected innovation in the apps business, for example.

Though a study might exist that quantifies the benefits or harm strong network neutrality rules have caused to aps markets, I have not seen such a study. Granted, it would be hard to “prove” a negative, such as an argument that with, or without strong net neutrality rules, the app and content markets would look much as they presently do, now.

It seems logical to conclude that Netflix would be the force it is under any set of rules, because it is a product consumers like, and value. In fact, Netflix now argues that even the end of strong network neutrality rules should not affect its business.

Nor, some would argue, has it been proven that consumer-friendly policies often said to be network neutrality issues, such as zero rating of internet access, are detrimental to innovation. The software business now is said to have a “winner take all” pattern. If that is the case, then net neutrality rules, or the absence of such rules, are not going to be decisive.

The other issue is that zero rating might be a necessary policy as former linear video subscription services become managed services. Linear video users, voice and text messaging users have never directly paid a separate “access charge.” Instead, access is simply an enabler of the retail service. That same pattern will have to hold as streaming becomes the next-generation version of the video subscription service. People will pay for the content. They will not pay for the access, in addition to the content (in a direct way).

And, if has to be noted, if consumers are paying for a subscription video service, they are going to expect a level of quality consistency that will, at least on occasion, require QoS mechanisms to operate. In other words, the business model itself requires zero rating of bandwidth.

The point is that is not clear that presence or absence of strong network neutrality rules fundamentally or even significantly shapes app and content provider market results.  So long as basic network neutrality rules remain in place (consumer access to all lawful apps; no app blocking or interference based on ownership of the apps or content), it has not yet been demonstrated that any actual harm to app providers can be shown, based on lack of strong net neutrality rules.

It would be hard to do so, granted. But it just seems logical that it is the broader appeal of apps and services (demand) that drives app provider success. So long as access providers cannot block lawful content, or engage in other actions that would be criminal under standard antitrust or restraint of trade rules, it has yet to be proven that QoS mechanisms themselves cause harm to app providers.

Nor, it might be argued, do ISPs have so much market power they can do so, even if they wanted to do so. The evidence from distributor negotiations with linear video service content providers provides some insight. ISPs have not proven so powerful they can harm content providers who own valuable content.

In part, one might say, that is because access providers do not have the power to annoy their customers and business partners, as every legacy service is losing value, as customers shift to substitutes, or other providers.

According to a Deloitte survey, 74 percent of U.S. consumers “still subscribe to pay TV such as cable or satellite, but 66 percent of subscribers say they keep their pay TV because it is bundled with their internet.”

If that is an accurate guide to potential behavior, as much as 66 percent of the linear TV subscription business is in danger, and maintained mostly because of price breaks. As in the case of bundled voice and internet access, so video-plus-internet bundles might well disguise huge dissatisfaction with linear video services that would surface quickly if the bundling did not also represent cost savings.

That is one example of the reason why even leading access providers in the U.S. market cannot afford to annoy their customers or abuse leading app providers. Not only is there growing competition, but the basic products themselves are losing relevance.

How Much Churn, Revenue Loss Does Bundling Prevent?

Bundling of consumer services (phone, internet, TV, mobile) has, for several decades, been a foundational strategy in U.S. consumer markets. In simple terms, the strategy has been justified by the need for economies of scope, when economies of scale are dwindling. In other words, if the total number of customers if limited, or falling, one way to boost revenues is to sell more things to the customer base, where in the past it had been able to sell one thing to more accounts.


That is why the term “units sold” now is relevant. It is no longer the number of accounts, but also the number of services sold to each account, that drive revenues.


But bundling now also seems to hinge on “tie in” sales. In the United Kingdom, customers who want to buy internet access on the Openreach platform must also buy voice services. Similarly, to get the best rates, U.S. telco consumers often must buy voice to get internet access. Cable TV customers often find that the price for bundled voice, video and internet is priced so that it makes sense to buy three services instead of two, or two instead of one.


Service providers are well aware of the danger they face. If there were no price inducements, it is likely that voice subscriptions would fall faster and further, as the one product nearly every household wishes to buy is internet access. And even with all the inducements, only a half--or fewer--homes choose to buy voice on the fixed network.


There is evidence service providers are correct in believing bundling prevents significant revenue loss. According to a Deloitte survey, 74 percent of U.S. consumers “still subscribe to pay TV such as cable or satellite, but 66 percent of subscribers say they keep their pay TV because it is bundled with their internet.”

If that is an accurate guide to potential behavior, as much as 66 percent of the linear TV subscription business is in danger, and maintained mostly because of price breaks. As in the case of bundled voice and internet access, so video-plus-internet bundles might well disguise huge dissatisfaction with linear video services that would surface quickly if the bundling did not also represent cost savings.

The point is that internet access is the lead product for a fixed network selling to consumers, while voice is the least-wanted product. Video sits someplace in the middle, apparently. And even if the internet-plus-video is the strongest two-product bundle, demand is the issue. As Google Fiber found, far fewer than expected consumers bought the Google Fiber video service, than was expected.

So there now are indications that demand for linear video is far lower than once was the case. That is why experimentation with "skinny bundles" is so widespread. One element that can be adjusted is the retail price, to change the value-price relationship. And the easiest way to change retail price is to lower the "cost of goods sold," which means lower wholesale content costs.

Inevitably, that means buying fewer channels at wholesale.

Bundling does support gross revenue and unit sales beyond the levels that likely would occur without the bundling. But bundling also seems to work largely because it represents lower retail prices for a product the consumer really does want (internet access).

What Will Drive IoT Value? Algorithms or Data?

It remains to be what relative contributions of value will be made in the communications business by internet of things and machine learning (artificial intelligence). Almost by definition, huge arrays of sensors will create lots of data, from lots of connections. But making sense of all that data is where the actual business value will be created.

In that sense, AI is the bigger trend, compared to IoT. It is only illustrative, but some have estimated that, eventually, transistors used by some individual computing devices will vastly outnumber the equivalent number of a single person’s brain cells. It is an inexact comparison, but illustrative.

Some believe that, in the future, it is not algorithms but data stores that will drive value. The thinking, by some, is that algorithms will become commoditized (widely available at lowish cost), while it will be data on human behavior that becomes valuable because it is less commoditized.

That would be a complete inversion of the present pattern, where some argue algorithms drive value. That switch, where value lies in data stores is seen by many as happening in the internet of things areas.



Virgin Media Sets 100 Mbps Minimum Speed

Virgin Media will set 100 Mbps as the minimum consumer internet access speed on its network.

Not so long ago, the UK. government set a speed of 24 Mbps, then 30 Mbps as the minimum standard for a “superfast” service. Now, it has added new terms. “Ultrafast is defined as 100 Mbps by the European Commission.

Ofcom changed its definition of “ultrafast” from 100 Mbps to 300 Mbps.

is to become the first widely-available UK broadband provider to offer ultrafast speeds of 100Mbps and above as standard as it revamps its bundles and launches the Virgin fibre brand.

This move reaffirms Virgin Media’s position as the UK’s ultrafast broadband provider with a top speed of 300Mbps.   As the need for fast, reliable broadband increases a speed of 300Mbps is four times faster than Virgin Media’s main competitors’ top speeds and gives households more bandwidth to stream, game, chat and work all at the same time on multiple connected devices.

Ceilings and floors both are important for analysts, industry executives and regulators, but a good argument can be made that it is the floors (minimum and universally available speeds) that matter most. In the U.K. market, so far, Virgin Media has been leading the market. The same trend can be noted in the U.S. market, where, since 2007, cable operators have disproportionately provided the faster speed connections.


Tier
Speeds (download/upload)
VIVID 300
300Mbps/20Mbps
VIVID Gamer*
200Mbps/20Mbps
VIVID 200
200Mbps/16Mbps
VIVID 100
100Mbps/12Mbps






Tuesday, March 21, 2017

Many Mobile Execs Likely Already Understand They Might Not Win Big with 5G

It is probably fair to note that not every mobile operator today sees a clear 5G business model. Probably just as certain: even firms who do see a business model do not have a fully-understood strategy already in place for various models that might exist. Even its biggest supporters might readily agree that 5G is not likely going to be the best network for every important application.

It is not so much that there is confusion about potential 5G business models as there likely is an accurate understanding that incremental revenue opportunities will not be ubiquitous, equally substantial or transformative in every case. In the colloquial, there likely will be 5G winners and losers, with the greatest odds of success lying with the largest tier-one carriers with the biggest internal markets, deepest pockets and other assets that allow them to be significant providers of big internet of things applications and solutions.

MTS in 2016, in its core Russian market, saw slightly negative revenue growth, year over year, in both its fixed network and mobile businesses. That might be one reason why MTS executives are skeptical about the 5G business case, for example. MTS likely already understands that 5G mostly represents a faster network with lower latency, but not necessarily an obvious driver of new revenues at the application layer. In other words, 5G will be a better network than 4G, but might not drive all that much incremental new revenue from expected internet of things applications and devices.

Also, MTS, having largely finished its 4G network builds, would prefer to spend less on network capital, not more, which 5G will require.

As always, statements questioning business models have to be parsed. When any executive or scientist says “something cannot be done,” there are unstated qualifiers, whether the speaker acknowledges those qualifiers or does not.

When it is said that there is “no business model for 5G,” that can mean many things. It might mean that “at the present moment, before settled standards, available transmission gear and customer devices such as phones are available, there is no existing business model.” That is correct.

Sometimes, as with earlier generations of technology, some companies, having capital issues, and having just made a big platform investment, have downplayed the importance of a coming platform, to protect the value and revenue of their “just built” platform. That is reasonable, and provides yet another set of qualifiers: “we cannot, at the moment, having just finished our latest next-generation platform, afford to invest in yet another next-generation platform.” That also would be reasonable.

In other cases, the qualifiers might relate to potential market opportunities, as in “we do not see, in our current markets, serving current customers, incremental revenue opportunities of sufficient size to justify making the investment in 5G.” That also would be a rational set of qualifiers.

All of those sorts of qualifiers, and others you might think of, are reasonable enough ways to understand what actually is being said when one hears an exec say “there is no 5G business model.”

What is meant is that “there is no 5G business model for our assets, geographies, customer bases, financial resources, human and financial capital assets and markets, at this time, and with our current understanding of incremental revenue opportunities.”

That never should be taken as a statement that “no other company, in any market, and with any different set of resources, can create one or more new business models sufficient to support investment in 5G, at some point when standards, network platforms and consumer terminals are available.”

There is a difference between “we cannot do so” and “nobody can do so.”

How Much Do New Platforms Boost Telco Revenue Potential?

One obvious fact about the competitive telecom business that makes it quite different from the old monopoly business (pre-1990s) is that better technology often has its greatest impact in the areas of operating cost or asset efficiency, even if the advantages are said to include ability to offer new services.

Although new technology more frequently reduce some forms of capital investment, new platforms sometimes drive big new revenue opportunities. Signaling system 7 improved telco operations, but also created the ability to offer text messaging as a retail service.

Optical fiber access networks add enough bandwidth that telcos could offer video entertainment services.

With a few exceptions, though, new platforms in the internet era have had very mixed implications for service revenues. The main problem is simply that the newer platforms (with the exception of enabling video entertainment), do not necessarily represent a chance to grow legacy revenues. Instead, new platforms mostly enable different (and arguably better) ways to deliver current services. The point is that, on a net basis, the better platforms do not necessarily represent “new” revenues, just better ways to supply “old” functions.

Consider voice over LTE (VoLTE). There are now 165 operators in 73 countries investing in VoLTE, including 102 operators that have commercially launched an high definition voice service using VoLTE in 54 countries, according to Juniper Research.

Some might interpret that development as representing a new revenue stream, in the form of high-definition voice services that are sold at a different price point. It remains to be seen if that actually happens. Some might argue HD voice only makes mobile voice usable where it frequently now is very sub-par in terms of audio quality.

In other words, mobile voice is “broken” and HD voice just fixes the problem, making mobile voice usable.

Likewise, some argue that over the top voice represents a revenue opportunity for mobile service or fixed service providers. The same might be said for voice over Wi-Fi. In most cases, though, such OTT voice mechanisms mostly provide indirect business value, operating cost savings or more-efficient use of existing assets.

For example, VoLTE means operators can avoid using 3G for voice services, freeing up capacity for other purposes, plus allowing simpler operations. Voice over Wi-Fi allows mobile operators to offload voice access operations in a more seamless way.

Still, for the most part, new platforms do not represent a net gain in service revenues. Part of the reason is simply that VoLTE replaces 3G voice, for example. To some extent, even investments in FTTH only allow telcos to keep pace with other competitors who are driving higher speeds, though FTTH does make high-quality entertainment video possible for a telco.

But since many alternatives exist for every legacy product, OTT voice and messaging allow new competitors to offer those services, with some ability to add features to a carrier service. That has a net negative impact on carrier revenue potential.

The larger point is that, in a competitive market, even better access platforms do not necessarily represent a huge net gain--if any--in carrier capabilities, where it comes to providing new services.



Monday, March 20, 2017

FTTH Economics Might Now Largely Depend on Monopoly Conditions

The economics of fiber to the home now seem to hinge very much on whether a monopoly deployment is possible, typically when one national network is built that is used by all, or most, retail competitors. Think of Australia’s National Broadband Network; Openreach in the United Kingdom, New Zealand’s network or Singapore’s model, or nations of the European Union.

In markets where facilities-based fixed network competition exists, entirely or in part (United States and Canada being examples of the former; the Netherlands, United Kingdom and France examples of the latter), market dynamics are quite different.

In principle, a monopoly fixed network provider could hope to have market share up to 90 percent (some satellite, some independent ISPs, some cable operators could take some share). In a competitive, facilities-based market, any single competitor might hope to get 40 percent to 50 percent market share. The lower figure might happen where there are significant third party suppliers, the latter figure happening when there are just two participants competing.

CenturyLink, for example, now earns 76 percent of revenue solely from business customers. The implication is that no matter what CenturyLink does, its consumer fixed networks drive, at best, 24 percent of results. That is a practical illustration of the “80/20 rules,” or Pareto Principle, which means 80 percent of results come from 20 percent of activities. For CenturyLink, one might say that 20 percent of the assets (those serving business customers) drive 80 percent of results.

Over recent decades, telcos have had to replace half of revenues every decade. The big shifts so far have been fixed network long distance to mobile; then mobile messaging replacing voice; then mobile data replacing mobile messaging. For many fixed network providers, internet access replaced voice, and now entertainment video is replacing internet access.

The point is that highly-competitive markets and diminution of all legacy revenue streams now mean the potential financial return from any network investment is limited. In essence, telcos and other competing service providers have been running in place, losing some revenue sources and replacing them with others.

For many access providers, the business problem is that new networks cost more, but support new revenues at a lesser level, in part because the amount of stranded assets (assets that do not generate revenue) grows in a competitive market, and in part because all legacy revenue sources are maturing.

That has huge implications. It means the economics of FTTH now are not driven principally by asset cost, but by the nature of the markets where the assets are deployed. Monopoly, or near-monopoly scenarios, are most favorable. Highly-competitive, facilities-based markets are where the payback is most tenuous.

In other words, even Verizon finds that up to 60 percent of its FTTH assets are stranded, and not generating revenue. That is not unusual, in any U.S. market.

So the new questions are whether FTTH actually makes sense, as a ubiquitous network, in any market where there is robust, facilities-based competition.

Many Winners and Losers from Generative AI

Perhaps there is no contradiction between low historical total factor annual productivity gains and high expected generative artificial inte...