Battle for the backbone: Can the incumbent telco empire strike back at the attackers threatening the financial health of the transport business?
? HIMESH BHISE, RUSSELL G. BUNDSCHUH, JEFF HAWN, KEVIN LALANDE,
AND NAGENDRA L. RAO
The McKinsey Quarterly, 1999 Number 2, pp. 77?87
Back when Ma Bell and its progeny ? the incumbent local and long-distance telephone
companies in the United States ? were the masters of all they surveyed,
communications networks were a scarce and expensive resource. But Ma Bell?s
telephone system is now under attack by new fiber-optic networks that are faster, more
economical, and better suited to carrying the data traffic that makes up an increasing
part of the communications business.
The new networks came into existence as a result of an ambitious rush by attacker
telecommunications companies to exploit the capacity constraints of the early 1990s
and to cash in on the data traffic boom set off by the rise of the Internet. To date, such
upstarts as Qwest, Level 3, and Williams Communications have announced plans to
bury some four million miles of optical fiber in the United States. Nearly half of this fiber
has already been placed in the ground.
These developments will almost certainly create an oversupply of raw, or dark, fiber
capacity and could make a commodity business out of important parts of the
communications industry. Such companies as Arbinet, Band-X, and Global Clearing
House have already started to create commodity exchanges where digital
communications capacity will soon be bought and sold like orange juice, heating oil, and
soybeans.
At stake is much of the $25 billion in revenue generated by the data communications
component of the US telecommunications industry, as well as a small but growing
component of voice traffic. In particular, the long-haul ?backbone? routes that voice and
data calls use to travel between big cities are at risk. Backbone transmission has been
a mainstay of the incumbent long-distance carriers. But now, as the challengers?
network build-outs move forward, these incumbents face a significant challenge on their
turf.
How should they respond? Nearly all long-distance incumbents feel pressure to upgrade
their aging networks across the board. Some, including Sprint and MCI WorldCom, have
already announced plans to build or upgrade their own networks to exploit the
capabilities of the latest advances in fiber and electronics. Meanwhile, the ?Baby Bells?
(the successors of the old AT&T regional operating companies) may be tempted to build
their own long-distance networks as soon as US regulators permit them to enter the
long-distance voice and data businesses ? something that could happen as early as
this year.
But attempts by the local and long-distance incumbents to compete directly with the
attackers in the long-haul transport business are not only unnecessary but also
potentially disastrous. With a few exceptions, incumbents should concede the ?transport
only? business to the attackers, which have cost and capacity advantages, and focus
instead on strengthening their retail interface with business and residential customers
and on providing a more complete set of solutions that begin (but do not end) with raw
transport. As backbone transport increasingly becomes a commodity, business
customers will care less about which companies own the ?pipes? and more about which
provide the best solutions to their problems. That is where the money will be.
WHY GET OUT OF TRANSPORT?
Long-distance and local incumbents have at least three good reasons to retreat from the
long-haul transport business and to abandon all thought of building their own long-haul
networks, respectively (Exhibit 1). First, the attacker data networks ? built with
state-of-the-art fiber-optic cable that can support the latest advances in optical and
electronic switching hardware, such as higher Sonet speeds and dense wave division
multiplexing ? have strong technological advantages over their older rivals. Embedded
network management software allows customers to monitor and maintain their own
pieces of these attacker networks, which can carry data as well as voice thanks to the
use of Internet Protocol and other packet-switching technologies. For some incumbents,
upgrading older fiber networks to match these technological advantages would be hugely
expensive.
The second reason for long-distance and local incumbents to eschew the backbone
business is that the latest technologies create astounding amounts of new capacity. If
all the fiber announced by US telcos were fully utilized ? and there is no technological
reason not to do so ? the backbone capacity of the United States could increase as
much as 200 times during the next three to five years (Exhibit 2). Even with the most
optimistic forecasts of demand, backbone transport capacity is not likely to be
constrained for the next seven to ten years and will probably be in significant oversupply.
Digital communications will probably sell on the cheap.
The third and most important consideration is economics. As customer acquisition
costs and churn rates soar because of the fierce competition for retail business and
residential customers, it will become increasingly important for telcos to capture a
bigger share of the customer?s total telecommunications spending. Only in this way can
they create the economies of scope that would provide an attractive return for high
customer acquisition costs. Incumbent long-distance and local carriers should therefore
focus on locking up their customer relationships ? before the attackers do.
Here at least the incumbents have a head start. Newcomers like Qwest and Level 3 have
created or are creating vast fiber-optic capacity across the United States and Europe.
But as yet, they have few retail customers. Many local and long-distance incumbents in
the United States face the opposite problem: they have plenty of retail customers but
are hampered by their legacy technology. Since value will ultimately migrate to
companies that control customer relationships, incumbents face the challenge of
maintaining them as voice migrates to data networks and as data traffic, traditionally
carried on high-cost private networks, migrates to public networks. The best way will be
to focus on the high value-added data services for which customers are most willing to
pay.
LIGHTING THE DARK
To succeed, incumbents will have to move fast. Demand for bandwidth is now increasing
more quickly than the cost/performance ratio of microprocessors, a ratio that according
to Moore?s law doubles every 18 months. Technology will continue to increase the
capacity of the networks being built today and is also going to unleash an astonishing
array of applications using this new capacity. Moore?s law has made many Internet and
other technology companies both paranoid and fast moving. Yet many telecom
incumbents find it hard to abandon the leisurely pace that prevailed in the days of
extensive regulation.
What makes the predicament of the incumbents still more urgent is their inability to do
much to forestall the transformation of the transport part of the industry into a
commodity business. Whether and how quickly the impending oversupply of fiber
creates wholesale price competition will depend almost entirely on the collective
discipline of the attackers.
Typically, those attackers each spend from $2 billion to $3 billion constructing a national
backbone network with 48 to 96 fibers connecting about 125 cities. Some $150 million
to $300 million of the total goes to ?light? only 2 to 4 fiber pairs with the optical and
electronic equipment needed to carry data traffic. The rest of the fiber remains ?dark,?
ready to be sold to wholesale customers or lit to accommodate additional demand.
Lighting the whole network at once would be extremely expensive, requiring an
additional $3 billion to $6 billion.
Attackers, however, have a good deal of control ? down to the city-by-city, fiber-by-fiber
level ? over the way they light new fiber. If they act with discipline and light it only to
meet incremental but substantial demand above the constrained capacity of today?s
incumbents, both sides may be able to earn a reasonable return on capital. But if one or
all of the attackers, with their cost advantages, should act irrationally and light new fiber
ahead of demand, damaging price wars will probably ensue as they struggle desperately
to fill their networks to cover the billions of dollars invested in construction.
The impact of this evolution on long-distance and local incumbents will probably be
considerable. Voice traffic, currently owned by the incumbents, would be targeted for
migration onto these new networks because of the high premium it now commands over
data. This traffic, and the customer relationships that come with it, can be used as a
platform to expand beyond mere transport into the higher value-added and more lucrative
data services dominated until now by incumbents. Qwest?s recently announced entry
into Web hosting, managed network services, virtual private networks, and other data-
and voice-related services shows that this expansion has already started.
WHAT TO DO
As transport evolves into a commodity business, value will come to lie not in the
long-haul backbone business but in ?integrated solutions? for customers: transport,
access,1 and professional services (see text panel, on the next page). Different kinds of
incumbents should follow different strategies for the transport element of these
integrated solutions.
Age gracefully
Long-distance incumbents like Sprint, MCI WorldCom, and AT&T already have
nationwide backbone networks. With the possible exception of Sprint?s relatively new
Integrated On-demand Network (ION), which gives that company the strongest position
of all the incumbents, much of this network infrastructure suffers from technological and
economic disadvantages (Exhibit 3, on the previous page). These incumbent networks
are heavily depreciated, however, so they look inexpensive from a financial standpoint.
Since integrated product offerings to customers necessarily have a transport element,
long-distance incumbents should manage down their own transport assets while
simultaneously partnering with attacker transport providers for any extra capacity they
need. They should first evaluate their aging networks on a link-by-link basis and upgrade
by installing new electronics (or even new fiber links) where demand justifies breaking
capacity constraints. Or like AT&T with its acquisition of IBM?s Global Network, the
incumbents could buy several routes at once when the price is right.2
Nonetheless, with so much backbone capacity coming on line ? and with prices thus
likely to drop ? purchasing and assembling capacity from such attackers as Williams
should be the centerpiece of the incumbents? long-haul transport strategies. Williams
has stated its intention of acting as a ?carrier?s carrier?; other attackers may soon follow
suit. Leasing arrangements for backbone capacity will likely benefit both sides:
attackers will get the volume they desperately need to fill their new networks;
incumbents will get cheap backbone capacity while retaining control of the customer
interface.
These partnerships will take some work on the incumbents? part. Much effort will be
required to devise billing system interfaces, metering mechanisms that measure and
record network usage, and network management and provisioning systems that will let
incumbents monitor the leased lines. Given the technical and contractual complexity of
these arrangements, deals between incumbents and attackers will necessarily look
more like long-term strategic partnerships than like mere outsourcing contracts.
Thanks to the relatively small number of major attackers, the first incumbents to strike
such a strategic partnership may reap significant benefits. By signing an exclusive
arrangement with a major backbone attacker rather than waiting for a liquid bandwidth
spot market to develop, for example, an aggressive incumbent could at least temporarily
block the other incumbents? access to low-cost backbone capacity.
Once strategic partnerships are secure, long-distance incumbents should turn their
focus to the high value-added data services3 that will probably be the ?sticky? part of
telecom product bundles. These sticky elements will help attract and retain customers
and reduce costly churn rates.
By investing in value-added services,
the incumbents can strengthen their
customer relationships
Incumbents should take full advantage of the fact that the attackers do not yet have
many large customers. By investing in value-added services, long-distance incumbents
can strengthen customer relationships and move closer to becoming ?infrastructure
integrators,? capable of offering large companies data communications solutions. The
final ingredient would be access to local feeder networks owned by the Baby Bells or to
competing networks now being built by Competitive Local Exchange Companies.
The one obstacle to this vision is the current lack of standards for the interoperability
and quality-of-service dimensions of the business. Without standards, it is more difficult
to structure robust agreements about levels of service ? in the acronym-happy telecom
industry, they are called SLAs, or service-level agreements ? as well as guarantees
about the performance that end users receive. But a flurry of activity by equipment
manufacturers, software providers, and carriers will almost certainly bring huge advances
over the next 12 months.
Many quality-of-service issues can be overcome with current technologies and policy
agreements covering such questions as how to handle top-priority data and the
minimum delay acceptable for voice calls. Both attacker and incumbent carriers already
provide tiered levels of service on their own networks. It is not too great a stretch to
assume that partnerships of carriers could agree on the policies governing the
interchange of data among their networks.
In addition, by working to create uniform industry standards, aggressive incumbents
would help make new backbone bandwidth a commodity, cut prices for new
transmission capacity, and improve their bargaining position as they drive to become
integrators of data communications infrastructure.
Stand by your MAN
The Baby Bell local phone companies of the United States face a somewhat different
situation, since none of them owns a long-haul network. Despite years of wrangling and
legal challenges, the regulators of the US Federal Communications Commission have
restricted them from participating in the long-distance phone business. But this seems
about to change: the Baby Bells will probably overcome the regulatory obstacles ?
perhaps even in 1999. Some may then be tempted to build long-distance networks as a
way of offering integrated service packages. For two reasons, this would be a mistake.
First, it makes no more sense for the Baby Bells than for the long-distance incumbents
to build advanced data networks when so many miles of fiber are already being laid.
Baby Bells that wish to get into the long-distance game should team up with backbone
attackers instead. Second, and perhaps more important, the Baby Bells have a
wholesale arena of their own that needs tending: their feeder networks in major
metropolitan areas. These metropolitan-area networks, or MANs, typically surround
large cities with ring-type structures. After entering such loops much as an automobile
enters a traffic circle, data travel around to the correct ?exit? and jump off to lower-speed
local ?roads.?
In the United States, 70 percent of all fiber route miles have been laid in MANs, with the
rest in long-haul backbones. Our analysis indicates that while long-haul backbone
transport will soon be in oversupply, capacity in MANs is likely to remain constrained.
Prices ? and therefore margins ? will probably hold up better there than in
long-distance transport.
The Baby Bells have several other reasons for paying careful attention to their MANs.
First, these local operating companies have an opportunity to maintain end-to-end
control of much local data traffic and thus to leverage their local brand equity and
existing relationships with regional small and medium-sized businesses. Second, the
Baby Bells will be able to charge other carriers for access. After all, backbone attackers
need to connect with end users and must fill up their new long-haul backbones, while
incumbent long-distance carriers need to connect with the Baby Bells? local feeder
networks to offer customers new high-speed offerings. Owning a superior MAN could
increase the bargaining power of the Baby Bells, especially in negotiating strategic
relationships with the long-haul backbone attackers.
Finally, solidifying the MANs would help the Baby Bells keep their current local
advantage even after local service opens up to competition. It would be prohibitively
expensive for competitors to duplicate all of the critical elements of the Baby Bell MANs
? conduit, fiber, and right-of-way portfolios ? at all quickly.
Yet the Baby Bells? regional advantage is far from guaranteed, especially in view of
mounting competition from Competitive Local Exchange Companies (CLECs), which
compete directly against them by providing local access, voice service, and, in many
cases, data offerings both to business and residential customers. These CLECs enjoy
several regulatory advantages over the Baby Bells, including the ability to buy (at a
discount) and resell unbundled elements of the incumbents? networks and the freedom
to target only a region?s most profitable customers. CLECs are using these strengths,
together with aggressive pricing and intensive customer service, to gain market share
from former Baby Bell monopolies.
Focusing on MANs will help the Baby Bells head off the attacks of the CLECs. First,
investing to maintain the advantage conferred by these MANs will make it more
expensive for attackers to differentiate their own build-outs from those of the Baby Bells.
Second, using new technology to improve the price-to-performance level and service of
their offerings will permit the Baby Bells to close the performance and service gaps that
CLECs are now exploiting. Third, if the Baby Bells focus their investment in areas where
they have already built strong relationships and brand images ? meanwhile avoiding
billion-dollar incursions into long-haul transport networks ? they will have a better
chance of establishing significant competitive advantages and a more defensible
competitive position.
THE ONE-STOP BUNDLE
Winning incumbents will probably look quite different after the transformation described
here. Incumbent long-distance companies will exploit strong relationships with local
carriers like CLECs and with long-haul wholesalers like Williams to bring value-adding
sets of applications and offerings to their customers. Winning Baby Bells will partner
extensively to enter the data and long-distance markets quickly and to avoid billion-dollar
investments in infrastructure. By outsourcing commodity-like elements and focusing on
more differentiable, higher value-added service and application components, winning
long-distance and local incumbents will not only be better able to defend existing
customer relationships but also to gain new ones with improved product portfolios, faster
times to market, and cheaper offerings.
These trends have already hit the marketplace: AT&T, for example, has done a deal with
Time Warner to leverage an existing local network (a cable one, though), and SBC has
contracted with Williams for wholesale transport capacity. More such relationships will
emerge over the coming months as players shore up their ability to interface directly
with customers and assemble the necessary bundling elements, such as transport and
access. The recent activity of AT&T exemplifies the effort to lock in customer
relationships; it will get fixed-line access to residential customers through its acquisition
of TCI and its agreement with Time Warner. AT&T has also aggressively priced its One
Rate plan to provide complete bundles of value-added services to customers.
AT&T seems to be acknowledging that a number of forces ? the availability of capital,
ongoing deregulation, the emergence of bandwidth spot markets, the spread of new
technologies, and the new open-standards network paradigm ? are creating
opportunities for very successful focused attackers. Rather oddly, however, they have
given incumbents a way to offer customers truly integrated products, and the
incumbents that do so first stand to gain considerable value. Local and long-distant
incumbents alike should bear this in mind when considering transport strategies.
STICKY SOLUTIONS
Most incumbent telcos are increasingly relying on professional services to tie
together the various components of a product bundle to keep big customers
loyal. Professional services ? that is, consulting, integration, the operation and
support of local-area networks, internetworking, and products and services for
wide-area networks ? are crucial for customers. The importance of professional
services reflects the development of new Internet ProtocolÑ based application
services that customers cannot now create for themselves, the complexity of
products and services, and the increasing cost pressures that are driving the
need for outsourcing.
Professional services help the incumbents by making them better able to build
strong customer relationships based on trust and by ?pulling through? (helping to
sell) other products and services; there are, for instance, economies of scope
between some professional services and transport. The telcos therefore risk
losing business customers to systems integrators and equipment vendors
expanding from their traditional positions of strength: if they succeed, the telcos
will be less able to sell customers transport or access in an integrated, ?bundled?
solution.
Although the benefits of focusing on professional services seem attractive,
pulling off a successful entry strategy will by no means be easy for incumbents,
since they will need new skill sets, new incentive and compensation
mechanisms, and different customer relationships. Incumbents ? both
long-distance and local ? have three broad service strategy options, and they
must evaluate their appetite for risk and reward and their present strengths to
decide which one to pursue (see exhibit). Each could work if executed well.
First, incumbents could build on their expertise in long-haul transport and
access services by offering network integration and management services for
their business customers? infrastructure needs, including the deployment of
intranets and extranets, Web hosting, and security. Many people see the
provision of network integration and management services ? which along with
transport services meet the infrastructure needs of business customers ? as a
natural extension of the transport services many incumbents already provide.
Incumbents could also focus service offerings on attractive horizontal or vertical
segments (customer and industry, respectively) by developing an expert
reputation in them. This approach could allow incumbents to pull through sales
of traditional data transport services.
Finally, incumbents could become pure integrated service providers, maintaining
the independence of their networks by offering transport and access options and
outsourcing low-value network integration when appropriate.
None of these three approaches can be realized unless the incumbents change
their mindsets and skills, but the benefits of success make the effort worthwhile.
The ability to provide integrated solutions for business customers could give the
incumbents a competitive advantage that rivals would find hard to match.
|
