For Every Public Purpose, There is a Corresponding Private Interest

Who pays for Internet access? Consumers and businesses. 

But advertisers or sponsors might pay on behalf of users of their services. Internet service providers might sponsor use of some applications. 

In some cases, application providers pay, on behalf of their customers. 

But mostly, it is end users who pay all the costs.

And though it is true that there are genuine policy issues surrounding a seemingly-endless list of "network neutrality" instances, there also are important commercial interests.

For access providers, the issue is whether apps that impose disproportionate network costs should help defray the direct costs they impose. 

For application providers, the issue is avoiding such costs, as they would directly affect app provider business models. 

And as the Internet has fragmented, there now are different kinds of Internet domains. The sort people generally are familiar with are Internet service providers who provide mobile or fixed network access. Those "eyeball" networks aggregate end users. 

Content domains are different, especially domains that supply video entertainment or video content. Such domains represent the majority of all demand on access networks. 

To the extent that ISP eyeball networks have to supply additional capacity to support such apps, the costs now are borne exclusively by end users, in the form of higher access fees. 

The issue is whether dual revenue streams will develop that resemble the way much print, TV and audio content is subsidized by advertisers. 

That notion is contentious as a matter of public policy. But the differences also reflect very real business models, and revenue and cost winners and losers in the internet ecosystem. 

As always is the case, for every public purpose there is a corresponding private interest. Proponents never directly say so. But it always is there. 

Which Analogy for ISP Interconnection: Retransmission or Carrier Interconnection?

Netflix and major U.S. ISPs use different metaphors to describe the process of interconnecting Internet domains.



Verizon, AT&T and Comcast, for example, use the analogy of carrier interconnection, where the amount of traffic exchanged determines whether any particular bilateral interconnection is settlement free (roughly equal amounts of traffic exchanged) or requires payment by the network delivering much more traffic than that network is accepting.



Netflix uses a different analogy, that of broadcast TV "retransmission fees," the fees paid by video subscription services to TV broadcasters for the rights to retransmit off-air signals as part of a video subscription.



Whatever one thinks of the reasonablenes of those analogies, there now is a huge traffic imbalance between "eyeball networks" that terminate Internet traffic for consumers, and "content networks" that deliver traffic to eyeball networks, but accept only modest traffic from the eyeball networks back to the content networks.



The reason is simple enough: content networks send video and other content to end users, but generally do not need to accept much upstream traffic from consumers, whose operations are generally confined to ordering a movie to watch or updating a play list. 

"Opportunity Cost" Might be the Biggest Downside to "Upgrading" Existing Product Lines

Marginal cost has proven to be a key concept for products sold in most mass markets, and telecommunications is not exempt from that trend.

Any number of retail prices and packages are set basically to reflect the marginal cost of adding the next incremental customer. In fact, over time, economists might argue, current retail costs for goods and services tend to move towards marginal cost.

The concept has lots of relevance in telecommunications, where large amounts of capital investment are “sunk,” and incremental usage actually imposes little additional operating cost.

The traditional way of illustrating the principle is to answer the question “how much does adding one more minute of use of the voice network cost?” In practice, the cost is almost solely limited to the cost of adding one more transaction on a billing system.

In the Internet era, we are accustomed to the notion that the next increment of usage of any consumer app actually costs the suppliers almost nothing.

In most cases, that next increment of usage costs the end user almost nothing, as well. You of course know the business problem thus created. When costs are nearly zero, retail price will tend to move towards zero as well.

As useful as that is for consumers, it is a huge problem for communications suppliers. Very low marginal cost explains why VoIP and messaging providers are able to offer their services literally for free, or nearly for free.

Low marginal cost is the foundation for the business model known as “freemium.”

Low marginal cost explains why suppliers of long distance calling, facing declining margins, try to compensate by encouraging additional usage volume.

And low marginal cost will be the reason why gigabit access services costing only $70 to $80 a month are possible, in large part.

Observers offer any number of suggestions to service providers about how to sustain their businesses under conditions where retail pricing for many products drops to marginal cost, and when marginal cost is quite low.

Many of those suggestions, though sound enough, have problems. Solutions that call for adding more value to existing products assume users will value the incremental new value enough to pay incrementally more revenue.

Strategies that call for creating new lines of business face execution risk (can telcos really do it?) as well as scale risk (will the new revenue streams be big enough to justify the effort?).

Assuming that creating new lines of business is both essential and realistic, the subsidiary issue then is how much to continue investing in legacy businesses that are declining in absolute value.

Two fundamental approaches can be taken: harvest or invest. The former essentially admits a business is mature, and will decline. The objective then is to preserve the magnitude of the revenue stream as long as possible, at the highest level possible.

The latter calls for spending more money to upgrade products, adding enough value that prices and usage can be sustained, or hopefully that usage and prices can even raised.

Low marginal cost might suggest harvesting is the more realistic strategy for most service providers. Adding more value might be capital and human capital intensive enough that the net result is a negative number.

Some of us would argue that if a given product line is powerfully affected by low marginal cost, the wiser choice is harvesting. The upside from big, or even significant investments, might not be large enough to justify the cost, time and human effort.

“Opportunity cost,” effort that might have gone elsewhere, also is a concern. No matter how large, organizations only have so much ability to invest in brand-new lines of business and revenue streams.

Must Telcos Cut Dividends and Structurally Separate Networks?

At least some small telcos that once paid rather sizable dividends have had to cut back or end such payments. 



Typically, when that happens, the company risks disruption of its owner base, as equity owners formerly buying a dividend payer are replaced by equity owners that seek growth. 



On the other hand, such moves free up capital to invest in the network. 



How well that strategy works long term remains to be seen. `But such decisions are anything but casual. 



Forrester Research analyst Dan Bieler argues that so important is the "connectivity services" function that  telcos should consider two courses of action they have traditionally opposed, namely separating network operations from retail operations, and paring back dividends.



Separating network operations from retail operations has been proposed before, and never to any widespread agreement on the part of telco executives. In part, that reflects a concern about commoditizing the access function.



Also, though, such structural separation also tends to come with greater pressure to sell transport and access services to third parties. 



Potential wholesale customers tend to argue that makes rivals "customers."  Facilities owners tend to argue structural separation eliminates a key perceived source of business advantage. 



Australia's National Broadband Network is among the more-prominent examples of the structural separation approach. Of course, that move was fought vigorously by Telstra, which arguably had the most to lose. 



Reducing dividend payments might not have direct competitive implications, but to the extent a stock price is a currency that can be used to acquire other firms, there is a negative strategic impact. 



Calls to slice dividends and separate network operations from the retail sales organization are not new. Neither are the business repercussions. 



Structural separation, and its related "mandatory wholesale" policies, arguably have proven to help competitors more than such policies help the owners of the facilities. 



The exceptions have been scenarios where the facilities owner traded vertical integration for regulator permission to enter new markets. SingTel did so. So did Rochester Telephone. 



But most telco executives continue to see more downside than upside, no matter how much advice they get to the contrary.

Could Google be Regulated Like a Common Carrier in Germany?

Google could be regulated as a common carrier or utility in Germany, if German regulators conclude Google has gained too much power and influence, a report by the German Federal Cartel Office suggests.



Such regulation could include price regulations governing prices for advertising, if regulators think Google's market position allows practices that violate antitrust rules. 

SoftBank, Deutsche Telekom Reach Fundamental Agreement on T-Mobile US Buy

With the caveat that antitrust review and clearance from the Federal Communications Commission is required, and by no means certain, SoftBank and Deutsche Telekom apparently have reached agreement on the broad outlines of a Softbank deal to buy the assets of T-Mobile US.



That would merge Sprint with T-Mobile US, the number three and number four national U.S. carriers.  



Under the plan, Softbank will buy more than 50 percent of T-Mobile US shares through Sprint, directly from Deutsche Telekom, which owns 67 percent of T-Mobile US. The deal is valued at about $16 billion.



The chances of regulatory approval are highly uncertain at the moment, though. Some might rate the odds of success as high as 70 percent. 



Others rate odds of success at about 55 percent. And some think the  odds of success are no better than 10 percent. 

A Second Wave of MVNO Growth Coming?

In January 2014, China issued 11 MVNO licences. In March 2014, Virgin Mobile acquired a MVNO licence from Saudi Arabian regulator. Those are potentially-important agreements because in many markets in the Global South, it has not been possible to lawfully operate as a mobile virtual network operator.

It is impossible to say how well such new MVNOs might fare. But past experience suggests that, in most markets, growth much beyond 12 percent customer share could be a challenge.

In 2009, for example, the market share held by MVNOs in Western Europe and North America was about nine percent, according to TeleGeography.

Globally, MVNO market share was about one percent. So MVNOs were a significant market presence only in two regions.

But logic would suggest that growth will occur in most markets globally where MVNOs either are not legal or commercially viable because carrier interest in supporting wholesale is low.

Asia is one of the regions where MVNO market share could be poised to grow from 50 percent to 100 percent annually, albeit from very-low installed customer bases.

The long-term issue is how much market share such new MVNOs might gain.

At least so far, it has proven difficult for MVNOs to break out from about 12 percent market share, perhaps because many MVNOs have focused on the “value” segment of the market, and “standard” offers from the facilities-based carriers have grown in that segment as well.

Whether that also will be the case in MVNO markets in the Global South is the issue. One might note that retail offers already are fairly aggressive in those markets. That suggests niche market strategies will be important, as the lure of “lower price” will be harder to create, than has been the case in Western Europe and North America.

By 2011, a separate analysis by Pyramid Research suggested MVNO adoption had reached about 12 percent, while in other regions MVNOs had less than three percent market share, most less than one percent market share.

That lead Analysys Mason to suggest that MVNO market share had essentially reached its limit, in Western Europe. To the extent that growth remained, it would come from wholesale-based offers adopted by device suppliers and service providers targeting market niches.

But in 2011, Scandinavian MVNOs faced falling market share. Denmark’s MVNO penetration fell to 4.2 percent from a peak of 10.9 percent.

In Finland, MVNO share fell to two percent from a peak of 11.7 percent. In Sweden, MVNO share dropped to 0.4 percent from a peak of three percent.

In 2010, there were abut 60 U.S. MVNOs in operation, according to the Federal Communications Commission. Most were quite small, but Tracfone had about five percent market share, earned by focusing on prepaid customers with low average revenue per user.