Thursday, August 3, 2017

80/20 Rule for Fixed Network Assets in the 5G Era

The function of the fixed network has been changing for decades, and should change still more in the 5G era. Originally, the fixed network was the mechanism for supplying retail services to consumer and business customers.

In the 5G era, more of the value of the fixed network will come from its trunking capabilities, to support fixed wireless, small cell, macrocell and business customers, and less from retail fixed network services for consumers.

That seems to provide yet one more example of the 80/20 rule, or Pareto distribution, in business and life. Even if 80 percent of the value of the fixed network eventually comes from 20 percent of the use cases, an access provider likely must continue to operate the full network for all potential use caes.

Several years ago it was possible to argue that three major “rural telecom services providers” had made a clear strategy switch to focus on business customer segments. CenturyLink provides the foremost example. After the acquisition of Level 3 closes, CenturyLink will earn a massive 88 percent of total revenue from business customers.

Windstream and Frontier Communications also had made moves in the direction of business segment revenue. As recently as 2015, all three former rural carriers earned a majority of revenue from business customers. Windstream earned 78 percent of its revenue from business customers, for example.

Then Frontier Communications made a major shift by acquiring Verizon accounts in California, Texas and Florida, representing about 3.7 million voice accounts, mostly of the consumer variety. That backed business revenue down from a high of about 46 percent to 40 percent.


For Windstream and CenturyLink, then, a big issue is that their legacy mass market networks are generating a relatively small portion of total revenue, but likely driving a disproportionate share of costs.

In its second quarter of 2017, CenturyLink earned about $2.2 billion from its enterprise segment and about $1.4 billion from its consumer segment.

But traditional rules of thumb suggest that about 80 percent of total costs come from the access network, and that relatively little of the access network cost actually supports business customers. The other problem is stranded assets.

CenturyLink like has less than half its access lines serving the consumer segment earning revenue, as cable TV competitors tend to have more than 50-percent share in the consumer market. So the cost of the asset base is supported by an account base that is about half what it once was.

That is why CenturyLink  says its operating expense per access line increased by more than 50 percent from 2007 to 2015, from approximately $650 to nearly $1,000.

Enterprise costs per data circuit also increased from $18,831 to $20,832, from 2011 to 2015. But the rate of increase was less than in the consumer services segment, and revenue from enterprise services also is much higher.

It never has been clear that any tier one service provider’s mass market access network could be “spun off,” retaining only the core assets required to support its enterprise customers.

The big challenge is that the mass market network contributes some revenue, but not the growth, for a tier one fixed network such as CenturyLink operates. Verizon’s customer base is different, so consumer arguably makes a bigger growth contribution.

This might be one of those instances where the 80/20 rule operates. At a high level, 80 percent of the growth comes from 20 percent of the network, but no operator can afford to dispense with those assets.

It remains striking that the rather-extensive access networks operated by the former rural carriers generate so much of their total revenue from a relatively small portion of the asset base. That is less true for the cable networks, which were legacy consumer market suppliers, but the portion of cable operator business revenue will continue to grow.

No comments:

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...