On Fiber or Copper Access Connections, Heavy Users are Just Heavy Users

Given that a small single-digit percentage of consumers consume perhaps 80 percent of all ISP bandwidth, and that most consumers use more bandwidth when they use faster access services, it might seem axiomatic that customers on optical fiber connections would use more data than consumers on copper-based connections.

That is the case, data from Ofcom, the U.K. communications regulator, suggests. What might be more surprising is that heavy users, on either optical or copper access connections, do not consume vastly-different amounts of data.

Ofcom data suggests heavy users on optical access networks might consume about 1 Gbyte a day more than heavy users on copper access connections. 

But the 90 percent of other users have consumption patterns that are much less disparate. To wit, 90 percent of customers on optical connections use 665 Mb a day, while 90 percent of customers on copper connections consume 465 Mb a day, a difference of just about 100 Mb a day. 

The point is that heavy users behave pretty much the same way on optical or copper access networks: they consume 80 percent of total data used by all customers. 

Increase Access Speed 1 Mbps, Consumption Grows by 1 Gbyte

Ironically, as Internet service providers boost access speeds on their networks, they virtually automatically increase the amount of bandwidth users consume, as this data provided by Ofcom , the United Kingdom communications regulator, clearly shows. 

Note that the relationship is linear. As a rule of thumb, every increase of 1 Mbps in access speed leads to more than 1 Gbyte of data consumption by an average user.

Designers of highway systems around major urban areas are familiar enough with the process. Highways are built to alleviate congestion. But the existence of highways generates more traffic. So congestion never really improves. 

Anybody getting the picture of a gerbil on a treadmill?

Ethernet Delivers Most of the Bandwidth, Special Access Makes More Money

With the caveat that “bandwidth” is not the same as “lines” or “revenue,” business Ethernet bandwidth passed the volume of legacy TDM bandwidth in 2012, says Vertical Systems Group.

“Propelled by purchases of 1 Gbps and 10+ Gbps services, Carrier Ethernet will contribute more than 75 percent of total global business bandwidth by 2017,” says Rosemary Cochran, Vertical Systems Group principal.

By some estimates, the U.S. special access market (T1 connections, for example) generates $12 billion to $18 billion in annual revenue for service providers, with the overwhelming amount earned by AT&T and Verizon Communications.

As more incumbent and competitive service providers alike expand their respective wide area and metro Ethernet service portfolios, Insight Research argues the market will grow from $4 billion in annual revenues in 2011 to nearly $11.1 billion by 2016.

In other words, as often is the case for IP-based services, bandwidth consumed generates less revenue per bit than legacy services.

New Report Confirms: Investment or Competition is a Real Issue for Access Networks

The latest Ofcom report on U.K. broadband infrastructure illustrates the inherent tension between promoting investment in next generation networks and fostering robust competition between suppliers of such services.

Put simply, there often, if not always, is a tension between policies that promote competition and policies that create incentives for investment.

In addition, different nations have historical differences on the supply side that create different outcomes on the demand side.

For reasons having to do with widespread facilities-based cable TV networks being deployed, competition in the United States, for example, is inter-platform (between cable and telco networks), where in Europe, because of less-prevalent cable TV deployment, competition has been intra-platform, based on competitor wholesale access to the telephone network.

Ofcom notes that infrastructure-based competition between local incumbent telcos and local cable companies lead to early investment in faster access networks.

By way of contrast, in Europe fixed-infrastructure competition is less common, so regulators have required that incumbent telcos provide wholesale access to their networks, fostering competition but also creating negative incentives for investment in faster networks.

Ofcom also argues that retail pricing policies have differing impact on demand, though some might attribute much of the difference to supply constraints, in particular the longer average loop length in the United States, compared to Europe.

Ofcom argues that speed-tiered pricing in the U.S. market accounts for the much-larger proportion of fixed broadband connections are at speeds of 2Mbps or less, compared to Europe. In other words, supply shapes demand. Because higher speed tiers cost more, people buy more of the slower tiers of service.

Others might argue demand is limited by supply, in the sense that longer loops limit maximum speed, so speed tiers do not limit demand, physical limitations of the network limit uptake of services faster than 2 Mbps.

IDATE, for example, says 21 percent of fixed broadband connections in the United States at the end of 2012 were at a headline speed of 2Mbps or less, compared to just eight percent of connections across the UK, France, Italy, Germany and Spain.

Generally speaking, 2 Mbps is a hard limit on speed at about 5.1 kilofeet. The "average" loop length in the United States is about 4.25 kfeet, meaning half are longer than 4.25 kfeet. Lots of long loops mean lots of lines that are not capable of speeds much faster than a few megabits per second, on all-copper loops.

That will start to fade, as a limitation, as more networks are reinforced with optical fiber. Newer versions of digital subscriber line technology really do help, but loop length has to be controlled.

Telekom Austria Wins Half of LTE Spectrum

Once again, it seems, next generation mobile spectrum auctions have a mobile service provider caught between a rock and a hard place.

Telekom Austria won licenses to half  of the 4G spectrum sold recently by the Austrian Telecom Control Commission. That's the good news. The bad news is the debt load the purchase will impose.

Executives with long memories will remember what happened when many mobile service providers overbid for 3G spectrum. Several firms faced bankruptcy when revenues did not grow enough to justify the payments.

The Austrian Telecom Control Commission said the EUR2.01 billion ($2.75 billion) sale had Telekom Austria AG paying about EUR1.03 billion for 14 frequency blocks, while Deutsche Telekom AG's Austrian subsidiary T-Mobile paid around EUR654 million for nine frequency blocks. Hutchison paid about EUR330 million for five blocks of spectrum.

The good news for Telekom Austria is the strategic spectrum advantage it will have.

Compare the 3 Austria holding of about 18 percent of  Long Term Evolution spectrum auctioned, or T Mobile’s 32 percent holding.

It is possible 3 Austria will face capacity constraints, eventually.

But the costs of acquiring so much spectrum will be costly for Telekom Austria. The acquired spectrum costs will increase its net debt to 3.3 times earnings before interest, tax, depreciation and amortization, risking a credit downgrade.

Tablets, U-verse Drive AT&T 3Q 2013 Results

“Connected devices” (tablets, principally) drove net mobile additions at AT&T during the third quarter of 2013, as did U-verse broadband services in the fixed network segment.

AT&T added nearly one million net mobile subscribers, including 63,000 mobile postpaid accounts and 192,000 prepaid accounts.

But connected device net adds were 719,000, or 73 percent of net additions.

In the fixed network segment, overall consumer revenue grew 2.4 percent, year over year, principally because of U-verse services.

About 70 percent of AT&T U-verse TV subscribers take three or four services from AT&T, driving average revenue per user  for U-verse triple-play customers to more than $170 a month.

Also, In the third quarter, more than 90 percent of new U-verse TV customers also signed up for U-verse high speed Internet services, showing the growing importance of the broadband double play (high speed Internet access plus video entertainment).

U-verse TV penetration of customer locations continues to grow and was at 20.8 percent at the end of the third quarter, AT&T said.

U-verse revenues grew 28.1 percent, on the strength of 655,000 high speed Internet subscribers and 265,000 U-verse TV subscribers. AT&T also added 97,000 U-verse business customers.

For the quarter ended September 30, 2013, AT&T's consolidated revenues totaled $32.2 billion, up 2.2 percent year over year.  Total third-quarter fixed network revenues were $14.7 billion, down one percent, year over year.

AT&T’s fixed network operating income totaled $1.5 billion, down 13.7 percent versus the third quarter of 2012.

Third-quarter fixed network operating income margin was 10.6 percent, compared to 12.1 percent in the year-earlier quarter, down primarily due to customer growth-related costs and costs incurred as part of Project VIP, AT&T’s network upgrade program.

Mobile segment revenues grew 5.1 percent, with service revenues up 3.7 percent, year over year.

Mobile data revenues grew 17.6 percent compared to the year-earlier period, and mobile operating income margin was 26.4 percent, while mobile EBITDA service margin was 42.0 percent.

One might draw several conclusions from those figures. First, the sharply-reduced rate of net additions confirms that the mobile phone market is saturated. In past years the U.S. mobile industry added 300,000 to 800,000 net new customers every quarter, most added by just two carriers, Verizon Wireless and AT&T Wireless.

So for AT&T to add just 63,000 postpaid customers is a huge reduction of about an order of magnitude.

Also, prepaid now accounts for an order of magnitude more net new adds than postpaid, showing where growth now is occurring, at least for AT&T.

The new development is the commanding role of tablet connections, even if the average revenue per device is going to be far lower than for phones.

The third quarter results also showed continued movement towards usage-based plans. About 72 percent, or 36.4 million, of all AT&T postpaid smart phone subscribers are on usage-based data plans, up from 64 percent, or 28.5 million, a year ago.

The newer trend is a shift of connections from the new “Mobile Share” plans, which also likely accounts for the surge of tablet connections. More than 16 million connections, or more than 22 percent of postpaid subscribers, have moved to Mobile Share plans.

The number of Mobile Share accounts reached 5.3 million in the third quarter with an average of about three devices per account.

Take rates on the higher-data plans continue to be strong with 30 percent of Mobile Share accounts choosing 10 gigabyte or higher plans. About 15 percent of Mobile Share subscribers came from unlimited plans.

In the third quarter, postpaid churn was down slightly to 1.07 percent compared to 1.08 percent in the year-ago quarter. Total churn was 1.31 percent versus 1.34 percent in the year-ago quarter and 1.36 percent in the second quarter of 2013.

About 90 percent of postpaid subscribers are on FamilyTalk, Mobile Share or business plans. Churn levels for these subscribers are significantly lower than for other postpaid subscribers, which is one reason service providers such as AT&T like them.

Mobile and Fixed Network ISPs Face Different "Key Challenges"

You likely could get a good argument about whether fixed network or mobile network operators face tougher financial challenges balancing revenues and network costs, over the next decade.

Both types of networks will have to be upgraded to handle expected bandwidth demands, and yet both fixed and mobile network businesses are likely to face significant revenue challenges, at the same time.

You might be tempted to argue mobile service providers are better situated, even if the financial pressures on mobile operators have to “tear apart their network economics,”  along with their network architectures, Maravedis-Rethink.

“Operators will slash costs by leaving only ultra-low cost equipment at the cell site, eventually driving the equipment cost down below $100 by 2020,” Maravedis-Rethink analysts argue.

New mobile network architectures will rely much more extensively on huge numbers of smaller cells, where today the network is characterized by a smaller number of macro cells, says Caroline Gabriel, Maravedis-Rethink research director.

The big story here is the $100 cell site radio infrastructure, part of the deconstruction of the transmission infrastructure to radically reduce costs.

A 2011 study by Juniper Research predicted that global operator-billed revenues will exceed $1 trillion annually by 2016,  but that operator costs would exceed revenue within four years unless changes are made.

Many would argue that service providers are smart enough to revamp cost structures even as they work on generating higher revenues. Some might argue that users already have learned to shift as much as 80 percent of data consumption on their smart phones to Wi-Fi networks.

Small cells are seen as one way to offload even more traffic, in outdoor spaces, from the mobile network.

Fixed network operators arguably face a different sort of problem. The ability to reduce the capital intensity of a fiber to home network argubly is vastly less than what a mobile operator could do.

But the big challenge probably is not “network cost” so much as a destabilizing change in end user expectations about value and price relationships, namely the challenge of a competitor selling 1 Gbps service for $70 a month. To be sure, few ISPs actually face the problem today.

AT&T, which says it is beginning construction of its gigabit network in Austin, Texas, might reach the gigabit standard at some locations by mid-2014.

AT&T says it is starting with “tens of thousands” of customer locations throughout Austin and the surrounding areas at the end of 2013 (300 Mbps), with additional local expansion planned in 2014, though no specific targets were immediately revealed.

But AT&T also faces a problem Google Fiber does not, namely an installed base of customers whose perception of what a reasonable Internet access offer looks like will change, if AT&T starts selling gigabit connections for prices anywhere close to what Google Fiber offers.

The problem is the same as AT&T and other telcos faced when VoIP began to get traction in the market. Telcos had scores of millions of legacy voice accounts sold for healthy premiums compared to many of the upstart VoIP providers.

So the strategic challenges was simple: match VoIP prices and features, or essentially refuse to match those offers and face erosion of accounts. One might argue that is not smart, but the math works.

Instead of taking an across the board revenue hit on all existing lines, telcos mostly decided to lose market share over time, while maintaining prices on a dwindling customer base. To be sure, customers were deserting in any case for mobile alternatives.

AT&T will face the same issue with gigabit networks, as the new value-price relationships will be reset, by its own gigabit offer and Google Fiber and other offers that are hovering around a 1-Gbps for $70 to $80 retail pricing range.

That of course has implications for the value-price relationship for the 300-Mbps service and all other slower speed services, which essentially will be priced in relation to the 1-Gbps service.

Google Fiber, for example, already has done so. Customers can sign up for free 5 Mbps service, guaranteed for seven years.

In other words, the big problem for a growing number of ISPs is different expectations on the part of consumers. Few consumers may want to buy a gigabit service costing $300 a month or more.

Many will pay for a gigabit service costing $70 to $80 a month. One survey suggests 50 percent will buy Google Fiber at rates in that range.

So while the bigger problem for mobile service providers might be new ways to dramatically boost capacity while controlling cost, fixed network ISPs can do only so much on the capital investment front, but face unprecedented pressures on the retail pricing front.

Where today fixed network customers expect to pay $50 for service up to 15 Mbps or 20 Mbps, or perhaps $80 for 50 Mbps access, that obviously will have to change when the ISP is confronted with a rival ISP selling gigabit connections for $70 to $80.

For mobile service providers, the key challenges arguably are related to network cost. For fixed line providers, the key challenge is pricing pressure.