The TeleChoice Take
Driving massive deployment and adoption of broadband
services without addressing the fundamental pricing flaws in the
IP services that those broadband services primarily feed is likely
to represent a "fatal cure" for the most "successful" IP carriers.
The primary driver for the adoption of broadband access services is
the additional bandwidth available for IP applications. In other
words, broadband access represents a faster "on-ramp" for IP
services; however, IP services are generally not profitable. We
believe this lack of profitability is not an issue of economies of scale
but fundamental flaws in how IP services are priced and packaged
to consumer and business customers.
Although IP services have been a high-growth area for most
service providers, IP service businesses are almost universally
operated at a loss. The problem has been compounded with
increasing broadband adoption, with broadband users
representing exponential increases in IP traffic load without
representing corresponding increases in revenue.
The broadband value proposition to customers has been compelling,
but the business model for providers has proven shaky at
best. Consumer and small business customers have been lured
to upgrade from 56k dialup at $20/month to 512k DSL at
$50/month -- nearly 10x the bandwidth for 2.5x the cost. Mid-sized
to large business customers have more recently been lured to
upgrade from 1.5Mbps T1 connections at $500/month to
100Mbps Ethernet connections at $1,000/month -- 67x the bandwidth
for 2x the cost. Although the technology-driven cost advantages to
DSL compared to dialup and Ethernet compared to T1, combined
with economies of scale can and, in some cases, have made
the broadband access business profitable, the IP service
businesses being fed by those broadband pipes are
increasingly financially challenged.
The real issue is that the entire IP services pricing structure was
defined at a time and by players that were not well positioned to
create a sustainable business model. Over time, continued
competitive intensity has only amplified an initially flawed
Very simply, the problem with IP services is in network
utilization. Although peak-utilization will be 100%, at least for a
fraction of a second, average utilization on IP networks is much
lower, typically 10% to 30%.
IP applications tend to be very bursty -- the faster the access speed,
the more bursty the traffic. Virtually all significant IP applications
adhere to a file-transfer model (email, the web, even P2P
Napster-like applications), meaning the traffic pattern is one of a
burst of traffic (as the file is transferred) followed by an extended
period of relative inactivity. Faster connections simply translate
to shorter bursts (at full line rate) and a greater ratio of
idle-to-burst time. Although aggregation of multiple
customer connections results in some level of statistical multiplexing,
IP traffic at any point in the network remains bursty and
Since a userís perception of network performance is driven by
his experiences at the peak ("Iím paying for a high-speed line, why
isnít this file downloading faster?"), networks must be designed
to perform reasonably well at that peak instant. The net result is
that service providers must design their networks with
enough resources (switch/router capacity, network capacity) to
handle the peak load reasonably well. Of course, as we all
know, providers donít build their networks with enough capacity
to handle all the peak load, so some packets get dropped resulting
in the "best-effort" quality of service for which IP is notorious.
So What? The real impact is a double whammy on the IP
services business model. On one hand, the providers must spend
way too much money (CapEx and OpEx) building and operating
a network big enough to operate "OK" during peak instants, but end
up with a network that doesnít operate well enough to
carry performance-critical traffic (at premium price
No Single New Technology Solution Fixing this problem doesnít require
a newfangled technology solution. In fact, a handful of
different available technologies are already available, some of
which have been around nearly a decade that provide the
technical ability to address this problem. Frame Relay, ATM, IP
Diffserv, and MPLS are the most well known standards-
based approaches to smoothing out traffic flows. Each
technology provides mechanisms for prioritizing packets (or frames
or cells) and identifying which ones must get through with
high performance and which can afford to be buffered for a short
period of time until the peak instant in the network passes and
capacity is available. The result is a lower traffic peak, requiring
less network capacity.
The real problem isnít a technology problem; itís a
service definition/pricing problem. Although the technology
capability has been available for a long time, we are not aware of
any service provider effectively using it to define its customer offer
to incent the kinds of behavior necessary for profitable services.
Think about it. Voice traffic is much more predictable and much
less bursty than IP traffic. Yet long ago, voice carriers learned it
made business sense to incent customers to move some traffic to
off-peak times. Remember the days when a daytime call cost 25
cents, but a call after 10pm only cost a dime? The net result had at
least two positive impacts on the service business model. Some
traffic that used to happen during the day was moved off-peak,
allowing carriers to build less capacity into their networks.
Additionally, new traffic (and new revenues) actually happened late
at night that wouldnít happen at the daytime rates. Since these
off-peak calls could use capacity that otherwise would be sitting
idle, much of that incremental revenue flowed directly to the bottom
Data services and, specifically, IP have failed to adopt
similar approaches. One challenge is that IP traffic will shift from
peak to off-peak very rapidly, so defining standard peak and
off-peak times of the day doesnít really make sense. However,
by prioritizing packets in the network, the same effect can be
achieved on a much shorter cycle.
If service providers offer a meaningful discount for each packet
marked low priority, they can reduce their network costs and
perhaps even gain some new price-sensitive traffic. However,
perhaps more importantly, by truly providing a premium quality
of service, providers stand a chance of gaining customer confidence
in carrying premium traffic at a premium price. We have done
some initial calculations of the impact under different scenarios and
the financial impact on IP service profitability can be dramatic
(see the "Super Broadband Deployment Initiatives" whitepaper,
We believe IP services are headed for a crisis situation. Solving
this crisis will be critical for the survival of todayís IP carriers.
Most equipment in the IP service networks already supports
traffic prioritization, so this problem may be solvable with little or
no additional investment in the network. However,
significant implications around existing customer agreements
and software systems changes need to be addressed.
What Can You Do?
We strongly encourage you to stop and consider the implications of
this situation on your business. Specifically, consider these key issues:
- If youíre a service provider, assess what it would take for you
to restructure your IP offers along the lines identified here. Feel free
to call us if youíd like to better understand what we have in mind.
Start talking to your customers about their desire for/openness to
tiered pricing on a packet-by-packet basis. Will customers increase
low-priority traffic if discounts are available? Will they use your
IP service for performance-demanding applications if premium
services are available? Are they willing to restructure their
current service agreements to accommodate these changes?
- If youíre an IP equipment vendor, understand how this
redefined service offer would be implemented by your customers.
Can you provide assistance to existing and new customers in
making this transition? How can you simplify the implementation
of features that exist in equipment already installed in your
customersí networks? Can you help your customers to model
the economic benefits of this change using your equipment?
Remember the greater the financial success your customers achieve
in operating services using your equipment, the greater their
appetite will be for additional gear in the future.
- For software vendors, what role can you play in fixing this
critical problem? Does your product represent part of the solution?
By making an incremental software investment, can service
providers unlock latent profitability in the huge investments
theyíve already made building out their networks?
Through challenging times such as these, strategic focus is
critical. Making wise product/service decisions and
effectively communicating the value of these changes to your
customers will be critical to your success. Let us know if we can
For Further Reading
If this topic interests you, we recommend you download the
following white paper for free from the TeleChoice Website
( http://www.telechoice.com/inprint_whitepapers.htm ):
TC Perspectives: Super Broadband Deployment Initiatives
( http://www.telechoice.com/inprint_RegMailer3.htm )
Need Some Help?
TeleChoice helps companies everyday better position their firms
and products for success, whether re-examining fundamental
business strategy or clearly communicating unique position and value
in todayís tough marketplace. Contact us at firstname.lastname@example.org.