DSL Prime: Universal Coverage is Possible

It's a neglected story in most of the broadband trade press, but small, independent ILECs are showing that complete broadband availability is possible.

by Dave Burstein of DSL Prime and Future of TV [November 17, 2004]

100 percent DSL coverage from 30 rural telcos
Who says it can't be done?
28 percent of those responding to a survey of NECA members deliver DSL to 100 percent of their customer base. Surewest's Dave Kamp wrote me a good while back "Two facts that we are most proud of are that in our LEC 83-square mile territory we are capable of providing DSL to 100 percent of the residences...and the fact that we have just hit 20 percent residential DSL penetration." I recently visited Vermont Tel, also close to 100 percent.

There's no technical or economic reason not to meet George Bush's "100 percent" target in three years. Ed Whitacre of SBC says he already serves everyone in his third of the country, using repeaters if necessary, although SBC in practice isn't meeting his target. I hope to get a chance in Nashville to ask Whitacre, "What will it take to make your goal real throughout SBC territory."

DSL Discontinuity Costs plunge 2001-2003
Costs dropped 40-70 percent, making universal service economical
Universal DSL service, or at least 95 percent+ availability, is now sensible and economical, DSL Prime has often reported. So I was surprised to see that DSL cost figures from the FCC TAC, developed three years ago, are noticeably higher than the ones I work with. Much of the difference was due to my working with marginal costs, appropriate if the networks are already built and you're figuring what to price additional users. The TAC data was closer to average costs. I assumed telco costs, existing offices, fiber, and loops, were in place for telephony; TAC numbers apparently include significant overhead I would attribute to the telcos side as sunk. I note, however, that DSL costs dropped dramatically between 2001 and 2003. I think this discontinuity explains much of the difference.

• SBC turned the operational corner in early 2001, with operating standards, a staff trained to follow them, and customer self-install. SBC reported a one-third drop in support costs, then the following year reported another one-third drop. Other telcos are perhaps behind, but nearly all are operationally much better. Bids for bell support contracts in 2003 were just over $2. Support is still less than acceptable, but no longer an unbelievably expensive operational disaster.
  
  
• Equipment prices went down, with the Chunghwa public tender for 800,000 units of DSLAM + modem going for $60 for excellent gear. Most large carriers are paying somewhat more.
  
  
• Remote equipment became dramatically smaller, cheaper, and reliable. 24 ports, field hardened, are "pizza box" size. Ericsson's 8-port unit is the size of a large paperback. They no longer require a full power connection, rather drawing juice from a share phone line. This allows them to be easily installed in field connects or on the pole.
  
  
• OSS systems are now in place and a sunk cost
  
  
• Volumes are now twice as high, which besides spreading fixed costs over a larger base hence spreading fixed costs over far more subscribers.
  
  
• Churn, still high, is down.

Result: Between 2001 and 2003, the price of delivering DSL went down dramatically. Randall Stephenson reported 30-40 percent margins "where we have volume, like California" in 2003, and confirmed costs have come down so much margins are similar in his 2004 $30 and $35 prices. Yahoo BB claims they are profitable at $22 for 7 Mbps service with telephony, except for the cost of customer acquisition. Free.fr charges 30 euros for 5 Mbps DSL, unlimited calls across France, and 60 TV channels. It reported finances attractive enough to IPO this year. I suspect Yahoo BB and Free.fr are stretching their numbers, but clearly the average cost of delivering DSL in telco volume is below $25 with reasonable overhead and marketing, and could be as low as $15. The direct cost of serving another subscriber I calculate at $8 to $12 for a large telco. Competitors who do not have COs in place with fiber connections and must pay the telcos to share lines need to add that to their costs. That requires them to "run lean and mean," to stay in the game, unless the incumbent is greedy (DT) or incompetent.

Copyright 2004 Dave Burstein.
The DSL Prime Newsletter is reprinted with permission.

"The power of the printing press belongs solely to those who own the presses"
—A.J. Leibling

The Internet is the cheapest printing press ever invented.