Thursday, September 21, 2017

What to Do About Industry Challenges? "Take the Package," One Exec Quips

Displaying Screenshot 2017-09-22 at 8.40.11 AM.png
“Take the package” (early retirement) quipped Tony Mosley, Ocean Specialists director of business development, after a review of major trends in the global telecom business at the latest PTC Academy program in Bangkok, Thailand.

Mosley's playful retort came just before students developed a list of key challenges they would face as new CEOs of their own retail businesses.

The work teams came up with a list of six major issues they would have to confront:
  • Margin compression
  • Regulation
  • Over the top services
  • Differentiation
  • Spectrum
  • Convergence

As part of the three-day program, students (mid-career telecom professionals) are exposed to the business challenges leaders of businesses confront, and how they work to overcome those obstacles.

As always is the case, there was debate about whether it is possible to “move up the stack,” adding value and perhaps occupying new niches in the business ecosystem, to boost revenue and raise margins.

At the concluding session, students were immersed in thinking about whether such moves up the stack are possible, for which firms and industry segments, and where to look for such opportunities.

PTC Academy is an annual training program created by PTC to provide promising mid-career professionals with an introduction to the sorts of business challenges they will face as CEOs.

The event was co-sponsored by APTelecom and Loxley Broadband, and held at the facilities of Loxley, the Thai conglomerate, which also owns Loxley Broadband.







Wednesday, September 20, 2017

Is the Mobile Business Model Nearing an End? What Replaces It?

There are many reasons to expect a rather massive consolidation in the global telecom industry over the next decade. The cost of networks keeps rising, with every mobile generation. We soon will reach market saturation, when virtually every person who wants to use mobile phone service, for example, already does so.

Essentially, we are reaching a point where customers are spending about as much as they feel appropriate for the set of services now offered (business or consumer).

There are hopefully major new revenue opportunities coming, but we should also expect most mobile operators and fixed operators to need to replace about half their revenue every decade, from now on.

Those are a few of the big changes to come. There are others.

The 5G era is going to be different from all others in the history of telecom, for several reasons. Traditionally, scarcity has been the paramount business constraint.


Bandwidth was scarce and therefore expensive. Regulatory strategies were designed to keep matters that way, using monopoly or oligopoly market structures; expensive licensed spectrum; franchise or other market entry rules.


In the competitive era, where in most countries mobile competition has been the primary expression of competition overall, scarcity has remained a fundamental assumption. In most markets, less than 500 MHz of total spectrum has been available for all mobile operators to use.


Wi-Fi bandwidth helps, but even including all mobile and Wi-Fi bandwidth, less than a gigaHertz generally has been available for terrestrial applications.


All of that is going to change in the 5G era, which has to have mobile executives in many markets worried about the outcome. The reason is that multiple tools--small cells; new millimeter wave spectrum; spectrum sharing; new unlicensed spectrum; multiple input multiple output radios and spectrum aggregation across licensed and unlicensed bands are going to essentially replace scarcity with abundance.


The difference likely ranges between an order of magnitude to three orders of magnitude more usable spectrum being made available.  


Source: Intel


As always, a fundamental change in supply will affect demand and prices. Not only will the price per gigabyte drop, but we also are likely to see more use of private mobile networks and business models that essentially shrink the size of the addressable mobile market for the first time.


As we already have seen in the undersea capacity markets, significant demand actually is removed from the addressable market as enterprises (Google, Facebook, others) build, own and operate their own global networks. Where in the past they might have been expected to buy capacity from telecom providers, enterprises now increasingly simply operate their own networks, on their own behalf.


Look only at the impact of the shift to small cell architectures, compared to traditional macrocell network designs. Under any circumstances (low or high load), small cells boost total capacity as much as four to six times.


And that is before we start adding capacity implications from an order of magnitude or two orders of magnitude more total spectrum; spectrum sharing and spectrum aggregation of all sorts, that increases the efficiency of use of that spectrum.

Business model and market structure changes are virtually inevitable.

Abundance is Going to be a Scary Thing

The 5G era is going to be different from all others in the history of telecom, for several reasons. Traditionally, scarcity has been the paramount business constraint.

Bandwidth was scarce and therefore expensive. Regulatory strategies were designed to keep matters that way, using monopoly or oligopoly market structures; expensive licensed spectrum; franchise or other market entry rules.

In the competitive era, where in most countries mobile competition has been the primary expression of competition overall, scarcity has remained a fundamental assumption. In most markets, less than 500 MHz of total spectrum has been available for all mobile operators to use.

Wi-Fi bandwidth helps, but even including all mobile and Wi-Fi bandwidth, less than a gigaHertz generally has been available for terrestrial applications.

All of that is going to change in the 5G era, which has to have mobile executives in many markets worried about the outcome. The reason is that multiple tools--small cells; new millimeter wave spectrum; spectrum sharing; new unlicensed spectrum; multiple input multiple output radios and spectrum aggregation across licensed and unlicensed bands are going to essentially replace scarcity with abundance.

The difference likely ranges between an order of magnitude to three orders of magnitude more usable spectrum being made available.  

Source: Intel

As always, a fundamental change in supply will affect demand and prices. Not only will the price per gigabyte drop, but we also are likely to see more use of private mobile networks and business models that essentially shrink the size of the addressable mobile market for the first time.

As we already have seen in the undersea capacity markets, significant demand actually is removed from the addressable market as enterprises (Google, Facebook, others) build, own and operate their own global networks. Where in the past they might have been expected to buy capacity from telecom providers, enterprises now increasingly simply operate their own networks, on their own behalf.

Look only at the impact of the shift to small cell architectures, compared to traditional macrocell network designs. Under any circumstances (low or high load), small cells boost total capacity as much as four to six times.

And that is before we start adding capacity implications from an order of magnitude or two orders of magnitude more total spectrum; spectrum sharing and spectrum aggregation of all sorts, that increases the efficiency of use of that spectrum.

Business model and market structure changes are virtually inevitable.

Why So Many are Skeptical about "Moving up the Stack"

Generally speaking, as much as service provider executives talk about "moving up the stack" to avoid becoming a "dumb pipe," there is justifiable skepticism about whether that is possible, in many settings. History suggests it is hard to do, hard to do well, and almost never has resulted in significant revenue upside.

There are a few salient exceptions--mobile banking in some markets and video entertainment services in a growing number of markets.

But much depends on where a particular actor sits in the value chain, and what part of the industry that actor occupies.

If the core business is undersea capacity, it is easy to see why skepticism about such moves up the stack is worth attempting. When the core business is "pipe," moves "up the stack" are unlikely to be easy, and might actually be dangerous, as it is so far outside the core competence of the providers.

On the other hand, such efforts already have proven most successful for retail suppliers of communications services who historically come from the "applications" part of the business, and for whom "dumb pipe" has become a reality only in the internet era.

The best example is internet access service. That is the classic dumb pipe service. Traditional voice, messaging, specialized enterprise and business data networks, paging and video services were applications, where the network exists only to support delivery and use of the apps.

“Moving up the stack” might be among the most-difficult challenges a service provider ever faces. Arguably few such attempts ever have succeeded on a massive or even significant scale, where it comes to increasing revenue contribution.

And yet it is hard to argue that the effort must be undertaken. The reason service providers want to move up the stack is that the very structure of modern applications access makes this essential.

Virtually all applications now are delivered over the top, no matter who owns the assets. If you want proof, think about 5G. It now is the second mobile network where the standards were--and are--all about connecting computing devices. Voice support was an overlay for 4G, and so far is not even part of the discussion for 5G.

Think about that. A service that includes, as a core function, the ability to talk and text has, for the last two generations, treated voice as an afterthought. There is a very good reason. Voice--no matter who provides it--now is an OTT application, just like any other.

In other words, OTT is not a choice, it is a basic reality of the way value is delivered in the new communications ecosystem.

That is likely to be the case for internet of things as well. Value will be supplied in all sorts of ways. But the largest amounts of value, quantified as revenue, will happen in the services layer and the platform layer (all the OTT apps), according to John Kjellemo of Yandex.

Others might put the app layer at the top and the platform layer right below, but the net impact is the same. As much as 65 percent of total ecosystem revenue might accrue to the IoT apps enterprises and people want to use, or the platforms and integration services that support such use.

In what is a sobering forecast, Yandex, looking at available forecasts, believes it is possible that in 2020 less than one percent of mobile service provider total revenue will be generated by connectivity revenues, and a similar “less than one percent” of total IoT ecosystem revenue will be contributed by connectivity services.

In other words, even if 5G proves a resoundingly big deal for internet of things revenue, it still might be almost insignificant as a driver of mobile operator revenue, and will have relatively low profit margins as well.

Source: Yandex

M2M Data Networks are REALLY Different

Some idea of the character of machine-to-machine (sensor) applications can be seen by the typical amount of data that has to be transmitted by sensors. It may remind you of paging systems, should you recall them.

Sigfox says that sending GPS location requires about six bytes, temperature just two bytes. Object status might require only a single byte, the same amount of data consumed to report speed.

That shows just how optimized a low power wide area network is for low power consumption. Where one might think kilobytes are required to report at least some types of information, other simple status updates might take even less than that.

Such networks are optimized for uplink communications only, not duplex. So sensor transmitters can be turned off when not actively uploading data. Also, many sensor apps only require that a particular sensor be turned on for a few seconds a day, says Marc Olivier, Sigfox VP.

IoT networks optimized for mobile networks feature much-higher bandwidths. M2M networks on 4G networks tend to support a few megabits per second, while 5G variants might support 10 Mbps.

You might correctly guess that such positioning means mobile M2M sensor networks might have an advantage for higher-bandwidth apps, while purpose-built LPWAN networks might have an advantage for applications where reporting requirements are quite low.


Source: Sigfox


Source: Sigfox

Why Telco OTT Partnerships Rarely Succeed

“Moving up the stack” is about as key a concept in the telecom business as “dumb pipe” likewise is notable. “Up the stack” is a phrase that means an access services company (cable TV, satellite, telco, capacity supplier) moves from being a supplier of “pipe” services (network access and transport) to being an application provider.

In other words, such a firm returns to its roots. Recall that, historically, telcos made their business by providing voice communications. Communications (voice, originally) was the value and the business driver. To do so, telcos had to build networks.

They did not necessarily want to do so, but the product (communications) could not be sold without doing so. The same has been true with most products telcos sell. The product consumers and businesses often buy is some application with value, and the network is required to use those apps.

Broadly speaking this is now true for most consumer and business applications. What matters is the app and the value it provides. To use those apps, internet access is required.

This leads to an idea telco executives often are exhorted to act upon, namely “partner with over the top app providers” to re-establish a position in one or more applications areas with clear value for customers.

There is one big problem with such advice. It almost never works very well, from the telco’s perspective. There is a simple reason. The structure of modern communications, content and computing separates “infrastructure” (infra) from application.

Simply, access to the network is logically separated from application supply. Application providers need to direct business relationship with a network access provider to make their features available to potential users.

Internet access is required, but not the business relationship. Any user requires only some form of internet access, supplied by somebody, to use any internet app.

And that is why “partnering” with over the top app providers rarely works every well. Access providers simply do not offer enough value in an ecosystem where, by definition, any lawful app can be accessed by anyone on any connection, with not business relationship between the app provider and the internet access provider.

When Uber engineers look at how they design software, infra has to be assumed, of course. But Uber does not have to own, manage or operate any internet access infrastructure to create its software solutions.

So even if an infra services provider might like to partner with Uber, what value can it really deliver that Uber might find worth paying money for? All partnerships require value obtained by both or all parties. But no access provider “needs” a business relationship with an infra services supplier to build its business. That is what the internet is all about.

And also why “partnering” is rarely a successful or feasible strategy.

source: Uber

5G Use Cases: Few to None Involving Voice

With the caveat that I find “e-book” formats unpleasant and unnecessary, IHS has produced an “e-book” examining 5G use cases. Actually, it is more akin to a slide deck, so if you want a really quick and high level issue of timelines and concepts, that is what you will find.


One point is worth noting, when looking at 5G: it is all about connecting computing devices. Asked whether any serious work really is being done on “voice,” as part of the 5G standards effort, Reza Arefi, Intel director of spectrum strategy, simply says “no.”


Dean Bubley, founder of Disruptive Analysis, also notes that voice was an afterthought when 4G standards were created.


That speaks volumes about the expected use cases and revenue drivers for 5G.
source: IHS Markit

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...