DSL myths

As DSL deployment grows, the biggest challenge for service providers will be to improve the customer installation

ECALL 1980, WHEN YOU were one of the first people to have a mobile telephone installed in the trunk of your car. The radio installation took all day, and the unwieldy antenna was an eyesore.You had to wait for the mobile service because wireless channels were scarce. The "experts" from Bell Labs had fore-cast minimal demand for mobile telephony, but because of unexpected demand, radio channels would be at a premium until cellular frequency reuse was deployed later in the 1980s.

Fast forward to 1990. Although you were a much happier user of the new cell phone, there still were major drawbacks to wireless. You had a choice of only two expensive wireless service providers per market. Your cell phone was expensive and clunky. The providers gouged you on roaming costs. Also, your calls constantly got dropped because handoffs between cells were hit and miss, and cell coverage was still sparse.

Now fast forward to 2000. Your wireless service works great. You can choose among five or more wireless providers per market. The service is cheap and gets more ubiquitous all the time. Handoffs work well (most of the time), and they have abolished roaming charges. Now, even your kids have cell phones.

The old saying is that it is the early adopter pioneers who get all the arrows.Why did people stick with wireless through 20 years of deployment trials and tribulations? For one simple reason: They wanted the service; it added value to their lives.

New technology rollouts, especially those that require major infrastructure, require three key components to prevail: capital, declining cost curves and a major leap in user utility (and, as a result, accelerating demand for the service). Broadband meets all three criteria.

Another time-tested adage is that people underestimate the time and the pain required for a new technology to roll out and then also underestimate the impact of the new technology once it is deployed. So shall it come to pass with broadband. The fact is that customers love broadband.

During the past six months, the media have had a field day publishing stories about the terrible user experiences customers encountered in getting broadband services, including DSL. User horror stories have appeared in publications such as the Wall Street Journal and USA Today, describing the pain of installing DSL. There is even a new catch-phrase - "DSL hell" - that describes the delays and hassles of getting DSL installed in homes and businesses.

To add injury to insult, broadband service providers' stock prices have dropped by 60% to 70% since April. Investors have become skittish about capital-intensive telecom infrastructure businesses. Broadband rollouts have slowed as a result of capital drying up. What changed in the last six months? Has demand for broadband evaporated? Have emerging broadband applications gone away for good? If not, and if the forecasts of 50% Internet usage via broadband in five years are correct, how will the coming insatiable demand for broadband be satisfied?

The fact is that despite the negative press for broadband services - and DSL in particular - customer demand for broadband continues to far outstrip supply. Why? Just like wireless phones in 1985, high-speed, always-on Internet access is a service that people cannot get enough of, and demand for broadband Internet access will only continue to accelerate.

The negative attention surrounding DSL in particular has been doubly painful because most of these criticisms have been right on target. Successful DSL installation often has required complex coordination among multiple - often competing - players. Too often, users have been informed that they "don't qualify" for DSL because of loop length or telco line configurations.

However, there are several myths surrounding the current broadband rollout - particularly DSL.

Broadband changes it Always-on high-speed Internet access at speeds of 300 kb/s or more fundamentally changes the Internet user experience and revamps many Internet business models (Figure 1). The always-on feature means instant availability to the resources of the Internet. Studies show that with always-on service, users often move the PC out of the study or office and put it in the kitchen or family room and begin to use it far more.

High-speed Internet access lets service providers begin deploying thousands of new broadband Internet business models. Everything from streaming video to Napster and beyond is enabled by broadband.

With broadband, small and medium-sized businesses now can afford high-speed connections to tap into business software administered by application service providers (ASPs), a service that previously was only available to large enterprise customers. The twin Holy Grails of voice and videoconferencing become more attainable with broadband.

Some 10,000 new broadband business models already are sprouting. Many of these new broadband experiments will fizzle. Some will fundamentally change the way we work and live. But all of these new broadband applications need high-speed pipes.

Supply and demand Remember in the '50s when IBM scientists forecast that demand for computers would exceed no more that a few hundred Fortune 500 firms worldwide? Insatiable customer demand for computing power and mobile communications made these forecasts look remarkably timid. Customers continually have shown that they will pay for speed and performance. This is why Intel can sell millions of customers a brand new microprocessor every 12 to 18 months.

During the next five years, new local broadband communications services delivered over cable modems, DSL and wireless will alter the local service landscape just as the Internet already has irrevocably altered the business landscape in just five years. By the year 2003, Internet usage will grow from 40% to 50% to 60% to 70% of homes and businesses. According to Goldman Sachs, 50% of these Internet customers will have access via broadband, and 50% of broadband usage will be via DSL. This means that two out of three homes and businesses will be on the Internet, and one out of three homes and businesses will use broadband.

DSL deployments to date have met with demand that far exceeds available DSL supply.

After 18 months, despite price wars and installation nightmares, Covad Communications, NorthPoint Communications and Rhythms NetConnections continue to experience high demand, and all are turning EBITDA-positive in the same metro markets. Enterprise customers desperately want ubiquitous national broad-band coverage for services such as work-at-home. National retailers and ISPs are even now preparing for "Broadband X-mas 2000."

Broadband penetration is now less than 2%. Where will broadband supply come from? For the next five years, DSL, cable and wireless providers will be hard pressed to keep up with customer demand for broadband services (Figure 2).

Realities of broadband rollouts At present, service providers have three economical ways of delivering broadband services over the last mile to homes and businesses. These include cable modems, wireless and DSL.

Cable modems carve out two-way communication on upgraded cable plant. Cable modems are limited primarily to residential areas, but they are cheap and easy to install. The drawbacks to cable modems are that they require expensive cable system upgrades, and they may have speed and security issues stemming from their inherent shared backbone architecture. For example, cable modems slow down as more users are added, and service providers don't like to give up more channels for additional cable modems.

Today, high-speed wireless access is limited to line-of-sight deployment. This means providers must deal with issues such as roof rights, foliage and weather. Wireless deployment is pricey per user; therefore, providers such as Winstar Communications focus primarily on more dense metropolitan areas.

Mobile wireless probably will augment rather than replace wireline broadband. After 15 years, the cellular user base has grown dramatically, but wireless still constitutes only 7% of U.S. voice traffic. Surfing the Internet requires higher speeds and higher quality of service. It's difficult to see users giving up the broadband wireline solutions that will be in place in five years except for data applications tailored for wireless.

DSL uses fancy digital signaling over existing telephone wires. DSL is particularly well-suited for the bandwidth demands of small to medium-sized businesses and small office/home office markets because it covers the business and residential districts. Because DSL rides the improvement curve for microprocessors (otherwise known as Moore's Law), DSL already is touted as a solution that will go further and faster. For example, very high bit-rate DSL offers speeds of 40 to 50 Mb/s over telephone wires.

Of course, the greatest strength and the greatest drawback for DSL is that it uses the existing copper wires owned by incumbent telephone companies. Regulated telephone companies can be difficult.

DSL is the most capital-efficient telecom infrastructure deployment to date. When compared with a Class 5 telephone switch at $5 million to $6 million, the capital expense associated with DSL rollouts is relatively modest at and expense for DSL has traditionally been in the backroom - the messy labor and systems-intensive business of working with the telephone companies in provisioning and billing service.

The other potential candidates for broad-band delivery include dial up, satellite and direct fiber. But dial up has physical limitations because of its analog signaling technologies. Satellite is making noises about broadband delivery, but this solution still has uplink, capacity and cost issues. And despite the incredible improvements in optical communications, this solution will probably always be expensive. It remains very labor-intensive and disruptive to dig up the streets or even string fiber on poles at the edge of the network (Table 1).

For the next five years, DSL, cable modem and wireless solutions probably will remain the most viable solutions for the mass-market rollout of broadband, unless a truly disruptive access technology appears. Barring this, there will be tremendous pressure to fix these solutions.

The DSL naysayers DSL technology was invented by Bell Labs and has actually been around for 20 years. But because customers had no alternative local providers, high-speed private line solutions such as DS-1 and DS-3 still were lucrative for incumbent players.

It is a little known fact that telcos actually have been using a DSL technology solution called high bit-rate DSL to inexpensively provision T-1 lines for business customers for 10 years. Until faced by competitive pressures, they chose not to pass their cost savings on to business customers. Even today these telcos still target DSL primarily for consumers rather than for businesses because they want to maintain high-margin private-line businesses and milk them as long as they can.

Five years ago, many reasons were given by naysayers as to why competitive DSL would fail. Let's call these the "old myths." Most old myths centered around the belief that the Telecom Act of 1996 specified that incumbent telephone companies would have to unbundle the network elements and lease them to competitors; investors believed that the telcos would make it near impossible to do this in practice. The thinking was that competitors would never get the space to co-locate their equipment in incumbent COs, and even if they did, the prices for network elements such as local loops never would be cost-effective.

During the past four years there has been a ferocious running battle between incumbents and new competitive insurgents. Fought in courts and in front of state and federal regulators, the battle has centered on arcane topics such as physical, shared and virtual co-location and obscure pricing models for network elements.And in general, competition has prevailed. The result is that there are now more than 8000 competitive co-locations and more than 350,000 competitive DSL loops in service.

Incumbents have only recently started to wake up to the threat of cable modems and DSL. They now are aggressively rolling out consumer DSL services. This is a direct result of inroads by competitive DSL providers and cable TV players. The legal and regulatory battles now have moved to equally arcane but important disputes such as sub loop unbundling and jurisdictional pricing. As long as incumbents know they can leverage their legal and political lobbying strengths to slow competition and gain time, these battles will continue.

Common DSL myths You must be retail to control the customer. Monthly churn for retail ISPs for dial-up service is 5% per month. Churn in long-distance is often 6% to 7% or more. In both these situations, the retail player "acquires, brands and bills" and yet it still does not "own" the customer per se. Dedicated Internet services such as T-1, frame relay and DSL have less than 1% monthly churn. If you combine this with Web hosting, churn is reduced even further. The point is that a physical service connection to a customer makes all the difference in "owning" a customer. And layering a key service such as hosting locks in a customer further.

Wholesale DSL providers such as New Edge Networks and NorthPoint have had great success in partnering with DSL ISPs to sell DSL. Both the DSL wholesalers and the ISP retailers enjoy 40%-plus gross margins. Both the DSL carrier with its control of the physical loop and the retail ISP have customer control. If you doubt this, just ask an ISP retailer how many customers they have switched from one DSL wholesaler to another.

DSL doesn't have sustainable margins; it will be a commodity business. Wholesale gross margins are more than 40%. Because of the capital efficiency of DSL deployment, most COs go EBITDA-positive between six and 18 months after DSL access multiplexer (DSLAM) deployment. It takes about 100 to 200 users to prove out a CO.

Here is what we know will happen based on past histories for technologies such as wireless. Retail consumer DSL prices will drop to $30 to $40 and compete against cable modems as volumes rapidly pick up. Business DSL prices will still range from $100 to $300, depending upon speed.

DSL margins will improve dramatically. Here is why: First, costs are dropping fast. Equipment costs (riding that microprocessor improvement curve) are declining by 15% to 20% a year. Loop costs are coming down as DSL begins to share the customer's primary telephone line rather than require a second line. Backhaul transport costs are dropping by 30% to 40% per year as more fiber goes in and signaling technologies such as dense wave division multiplexing (DWDM) improve. DSL modems are dropping in price and soon will be customer-installable, allowing providers to forgo expensive truck rolls.

There also will be additional revenue streams for DSL requiring minimal incremental capital. Data is just Phase I for DSL service offerings; Phase II includes layered services such as voice, virtual private networks, frame relay and video. These services already are being deployed. And don't forget about the futuristic broadband applications with undefined business models. DSL providers will be able to double or triple revenue per customer with little or know incremental capital investment.

DSL deployments will run out of capital. As previously mentioned, DSL is much less capital-intensive than deploying voice switches at $5 million to $6 million per switch or fiber at $20,000 to $150,000 per mile. It costs only about $150,000 to co-locate a DSLAM in a CO. In the past, investors rewarded growth over cash flow. As a result, almost all DSL providers had grand plans for broadest possible deployment, and large amounts of capital were required.

Of course, if the same DSL players limit their growth and focus on market penetration, they could get EBITDA-positive that much faster. The new mantra is "fully funded." Because investors have become jittery about capital in general, they now want to see a capital plan that can get the DSL company to positive cash flow.

The "running out of capital" argument can also be a self-fulfilling prophecy. If investors won't invest in DSL players, then they must limit their growth or actually run out of capital. In the end it will matter little. We no longer are talking about whether broadband will be deployed. Customers want it and are willing to pay for it. The only questions are who will have the capital, how fast will broad-band be deployed and which investors and service providers will win as a result.

The RBOCs will crush you. This is a common myth. But so far, the RBOCs have missed every DSL target they have set to date. It is a fact that the RBOCs are subject to the very same loop provisioning challenges that any DSL player encounters. They also must fight DSL and other competitive battles with a company culture that has been optimized for a regulated telephone environment for the past 50 years. It is no wonder that large telcos have problems being creative and responsive.

We have heard corporate titans talk about "turning the aircraft carrier." What we have also seen is that it is virtually impossible to turn the aircraft carrier in Internet time. Only Bill Gates and Microsoft accomplished this feat, and the results of their efforts to re-make Microsoft for the Internet still are not conclusive.

The `net' on DSL Here are the basic facts for DSL for the next five years: Demand for broadband and DSL is accelerating rapidly. Broadband will be 50% of Internet access in four years and when total broadband penetration surpasses 10% of Internet users, broadband applications will explode. Because of deployment issues, broadband demand will outstrip broadband supply for the foreseeable future.

For a large percentage of broadband users, DSL using copper lines already in place will deliver the best broadband price performance, and, as a result DSL, will enjoy explosive growth. That said, DSL is a tough business and the backroom challenges are formidable. However, DSL business costs will continue to decline. That includes lower backhaul costs, lower equipment costs, lower installation costs and lower loop costs as demand continues to accelerate.

The DSL addressable market will continue to expand as new digital signaling allows DSL to cover even longer distances at higher speeds. DSL will soon be able to serve customers now served by digital loop carriers. A whole new set of services, including voice and video, will be layered upon today's DSL data infrastructure. The result will be that DSL providers can enjoy better margins and serve ever smaller markets more profitably. The single biggest challenge for DSL providers will be to improve the customer installation experience and the ratio of successful DSL installs as DSL continues to scale. Customers who order DSL must actually get DSL in a timely manner.

Broadband is inevitable because customers truly want it. The only question is who does it, how fast and who wins. Certainly, the customers win.