Why broadband over power line (BPL) can't stand alone as a high-speed Internet offering.
You could almost feel former FCC Chairman Michael Powell's enthusiasm for broadband over power line (BPL) technology when he called it "the most important third way" to provide broadband to markets across the United States.
Despite advances by DSL and cable modems, the United States ranks 10th in penetration of broadband and data rate of the broadband services offered (among Organization for Economic Cooperations and Development countries)-a fall from the number three spot in 2000.
Although investment in broadband facilities to serve enterprise and large business customers is well under way, the FCC recognizes that improving the U.S. position requires addressing the mass market-and it is here that BPL holds promise. But is BPL financially attractive enough to the nation's electric utilities to warrant investment and management attention?
The Story of Broadband Demand
During the 1990s, entrepreneurs built excruciatingly detailed cost models of telecom networks, but the projections for demand, pricing, and market size were remarkably simplistic. The resulting impact on investors was widespread and devastating. Today's investors are much smarter, and they pay closer attention to demand.
BPL is best suited for mass-market applications, and BPL initiatives right now are targeted at this segment-particularly residential customers, the majority of whom already have two broadband options, DSL and cable modem-while other options are being piloted or deployed.1
Broadband market analysts typically measure the size of the residential segment in terms of households. Analysts expect that by year end 2005, overall Internet penetration will reach around 64 percent of the 115 million households in the United States. Assuming current levels of price and income elasticities, 62 percent of households using the Internet will subscribe to broadband service, if available in their area.2 By the end of 2005, approximately 46 million U.S. households will subscribe to broadband service (or would if it were available), and approximately 89 percent (or about 102 million U.S. households) will have access to DSL or cable modem. The remaining 11 percent (or approximately 13 million households) are too difficult or costly to serve via cable modem or DSL services.
This snapshot of the general landscape in which BPL will compete confirms a common phenomenon faced by developers of telecom infrastructure. Geographic markets that are viewed as highly attractive because of high density levels have already attracted rivals, while areas that are less attractive because of low density levels are wide open to new entrants. New entrants into markets where incumbents already are operating typically face an uphill battle. In these markets, BPL has to compete with DSL and cable modem in terms of price, service quality,3 and brand. For the market currently unaddressed by DSL and cable modem, BPL has an advantage, but the lack of density in these markets makes achieving scale and favorable economics a challenge.
The Broadband Financial Model: The Devil in the Data
The economics of broadband markets are determined by several key factors, notably: prices and average revenue per user (ARPU), capital expenditures (cap-ex),4 operating costs (op-ex), customer acquisition costs, and churn. The values for each of these variables depend on the specific geographic areas under study, with rural areas tending to have higher cap-ex and op-ex per line but lower levels of competition (and thus lower customer acquisition costs and lower levels of churn). Urban and suburban areas tend to have lower cap-ex and op-ex per line but higher levels of competition.
We developed a simple model of a retail BPL provider.5 We use a per-customer discounted-cash-flow model to calculate value per customer, measured as internal rate of return (IRR) per customer.6 We developed scenarios of different combinations of ARPU, monthly churn and incremental capital expenditure per line to reflect different geographic and market conditions. Higher levels of IRR are achieved under conditions in which cap-ex is low and ARPU is high but, as we discussed above, this combination of factors occurs infrequently.
The contours in Figure 1 show the per-customer IRR for representative combinations of ARPU, monthly churn, and cap-ex, and demonstrate the sensitivity of IRR to changes in these variables. We highlight three specific scenarios that illustrate markets in which utilities may operate.7
- Area A represents a rural market with a low level of broadband competition. This market is characterized by ARPU between $38 and $42 per month (which is on the high side of the national average for DSL), cap-ex of $700 to $1,000 per line, and a churn rate of 1.3 percent to 1.5 percent. The resulting IRR is between 0 and 10 percent. Applying a 10 percent cost of capital yields a negative net present value (NPV) on a per-customer basis.
- Area C illustrates an urban or suburban market with a high level of broadband competition. This market is characterized by ARPU between $26 and $32 per month, cap-ex per line between $500 and $700, and a churn rate of 1.9 percent to 2.1 percent. The resulting IRR is between -10 percent and 0. Applying a cost of capital of 10 percent yields a negative NPV on a per-customer basis.
- Area B is an example of a niche market. This may be a densely clustered community in an otherwise rural area, with a low level of broadband competition. This market is characterized by ARPU between $38 and $44 per month, cap-ex per line between $300 and $500, and a churn rate of 1.2 percent to 1.5 percent. The resulting IRR is between 10 percent and 20 percent. Applying a cost of capital of 10 percent yields an NPV from $300 to $500 on a per-customer basis.
These low and sometimes negative IRR and NPVs call into question whether entry into broadband markets is financially attractive enough to generate sufficient enthusiasm for utilities to continue to pursue this opportunity. To this, we also add the question of whether or not the BPL opportunity is big enough to make a difference to utilities.
By Enterprise Value: How Attractive Is BPL?
To answer the question of financial attractiveness, we compare the enterprise value per customer of a hypothetical BPL player with a panel of telecommunications and cable television companies.8
We include the BPL player operating in Area B, the scenario which had the highest IRR per customer, for this comparison. (Below we discuss the BPL players operating in Areas A and C.) As shown in Figure 2, the BPL provider depicted here has an estimated enterprise value per line that is well below that of the panel of telecommunications and cable television companies. Cox and Comcast have the highest enterprise value per customer. These cable television companies offer cable television, high-speed Internet access over cable modem, and digital cable telephony over their hybrid coaxial cable and fiber-optic networks. The major telephone companies (SBC, Verizon, BellSouth, and Qwest) have lower enterprise values per customer than Cox and Comcast. They have been losing access lines but gaining DSL subscribers, and they plan on offering video services by improving capacity on their networks (through the deployment of fiber optics). Telephone companies CenturyTel, Citizens, and AllTel operate primarily in rural areas and face lower levels of competition than the regional Bells.
Three primary factors contribute to the comparatively low enterprise values for BPL. First and foremost is the difference in ARPU realized by the BPL provider compared with others in the panel. The other companies provide multiple services to customers over a common infrastructure. Cable television companies provide a "triple play" of voice, data, and video, and telephone companies hope to soon. The monthly bill for a residential customer subscribing to the entire triple play might be as high as $120 to $150.9 Since all customers will not take the triple play, ARPU will not rise to this level. It may increase to about $70 per month, though. If the ARPU for BPL could rise to this level, enterprise value would increase by up to $1,000 per customer.
Second, churn, or the degree to which customers switch to other providers, has a pronounced impact on value. Incumbents typically enjoy lower levels of churn than newer entrants. We assumed a relatively low churn level of 1.5 percent for the BPL provider shown here. If the BPL provider could lower this assumed churn level as a result of brand recognition, consumer recognition of high-quality service, or bundling options, churn could be driven down to levels closer to the incumbents. This could boost value another $300 or $400 per customer.
Third, a portion of the difference in value may be attributed to the capital expenditures that are included in the calculation of enterprise value. The telephone companies and cable companies in our panel have deployed their networks over decades. Much of their investment already has been recouped. Utilities currently have a network in place but will have to incur cap-ex to realize BPL functionality. If the BPL provider could recoup its investment, value could increase by about $400 per customer. However, there is little that the utility can do to change this circumstance other than strive to keep cap-ex per line as low as possible or to receive a subsidy.
Combining the three factors—higher ARPU, lower churn, and lower cap-ex—could increase BPL enterprise value by about $1,800. This would place this BPL player's enterprise value in the mainstream of our panel. The economics of the BPL players operating in Areas A and C could also improve with these factors, although they would not compare as favorably to the panel as does the BPL player in Area B. The BPL market for the U.S. market as a whole likely would be a blend of these three areas. This means that ARPU and enterprise value per customer likely would be lower industry-wide than for the player operating in Area B.
Management: What to Consider?
Assuming that BPL is sufficiently financially attractive to raise the interest of electric utilities, is it material enough an opportunity (, is it big enough) to merit management attention? A definitive estimate of the size of BPL opportunity (, its industry-wide value) is difficult to gauge fully at this time because the key driver—cap-ex per customer—is still in flux. We estimate values for BPL industry-wide based on combinations of market share and cap-ex per line at various levels of service offerings and ARPU.
As shown in Figure 3, the potential market value is greatest with high market shares and high levels of ARPU.
If we take the middle of the ranges shown in this figure, industry-wide value for BPL is around $2.9 billion. This value depends on a set of modestly favorable assumptions: moderate to low cap-ex (about $400 per customer), ARPU that is higher than that for standalone, high-speed Internet access (about $45 per month),10 and a strong market share (about 20 percent).11 To attain a 20 percent market share nationwide, BPL would have to capture at least 10 percent of the market where broadband is already available and nearly all of the addressable market in areas where broadband currently is not available.
This value of $2.9 billion is relatively small when compared with the market capitalization of the electric utility industry. In Figure 4, we show the market capitalizations for five large electric utilities in the United States.
If we assume that each of these utilities entered the BPL business, and we further assumed a uniform allocation of total BPL industry value (, $2.9 billion) based on number of residential customers for these utilities, we see that the contribution of BPL to overall enterprise value would be between 0.2 percent and 1 percent for each utility.12 If we make our assumptions more optimistic—higher market share, higher ARPU, and lower cap-ex per customer—total industry value may increase three-fold, to about $10.4 billion. Still, using the simplifying assumptions made earlier, the impact on the utilities shown above would be less than 4 percent.
In our view, all other things being equal, the size of the opportunity would have to be greater than this level to merit significant management interest in BPL commercialization for many of the nation's electric utilities. Several utilities have expressed their concerns about straying out of their core areas of business competency, especially in light of failed rounds of utility diversification in the past and the wave of telecom bankruptcies following the bursting of the telecom bubble. The economics of BPL and the overall contribution to enterprise value may change considerably, however, if BPL is viewed as an enabling technology for utility operations in addition to its commercial potential.
Each year, the U.S. electric utility industry spends about $3.5 billion on telecommunications equipment and services for its own internal uses.13 Further, many utilities have represented the need to drive enhanced bandwidth deeper into their systems to enable advances in their own internal utility operation. As utilities push intelligence further into their systems to realize "smart grid" functionality,14 greater communication capabilities are required closer to the end-use customer. Allocating a portion of the cost of deploying BPL to the utility improves the economics of the commercial portion of BPL.15 The utility still will have to make the full investment in BPL, but it will realize rewards in terms of improvements in its core electric utility operations, as well as returns from its BPL business. Here, the utility may view BPL's commercial opportunity as a secondary benefit, as opposed to a primary driver.
The Obstacles to Broadband
It goes without saying that BPL's entry into the commercial broadband market, even under the most favorable bundling circumstances, will be challenging. The battle for broadband customers already is highly competitive. Given the difficult time telephone companies have had capturing market share for high-speed Internet access from cable companies, BPL providers likely will face similar, if not greater, challenges. If the ramp-up in BPL market share is too slow, the cost per customer will be too high (due to lack of scale) to yield a strong return on investment. Conversely, if customers are quite willing to switch to a BPL provider (a necessary condition for a quick ramp-up), they also may be quite willing to switch away just as easily. The resulting high churn in such a scenario would hurt per-customer economics. This situation is exacerbated by the dynamic nature of telecommunications and media markets in general. New technologies and applications have had highly disruptive effects upon even the most well thought-out business plans.
In our view, multiple product offerings (and of course higher ARPU) are required to make BPL an attractive commercial prospect for investor-owned utilities. As BPL's rivals lower the price for the bundle of triple-play services, ARPU will decline and margins on BPL products will be squeezed. For the BPL player, Voice over Internet Protocol (VoIP) over BPL is at least a partial solution to this problem. The addition of VoIP will serve to increase ARPU levels and increase IRR per customer.16 Failure to expand BPL service offerings from standalone provisions of high-speed Internet access, however, significantly lessens the attractiveness of BPL to both utilities and investors.
The economics of BPL also may be improved through partnering and facility sharing. Partnering with, for example, an Internet service provider may prove attractive because such a relationship can reduce costs through greater scale. The partner also may provide needed technical or marketing expertise. For utilities that view the commercial opportunities of BPL as secondary to enabling improvements to core electric utility operations, partnering also may reduce management distraction.
Without the bundling of multiple services over BPL, high-speed Internet access over BPL as a standalone product may be commercially viable only in certain applications-perhaps from municipal and cooperative utilities, and from some small to mid-sized utilities. Many municipal and cooperative electric utilities view their mission broadly, in terms of providing needed services to their customers, so estimates of low per-customer IRRs may not present a huge hurdle for these utilities, as long as the IRRs are modestly positive and broadband service is needed by customers. This also may facilitate the FCC's vision to provide broadband service in rural areas.
Most investor-owned utilities may not view their charter as broadly, however, so they may be able to justify commercialization of BPL as a standalone product only in limited, niche applications.
- Other broadband technologies that are currently being piloted include fiber-to-the-home (FTTH), loop extended DSL, WiFi and WiMax, 3G wireless, and broadband over satellite.
- We assume that households in areas currently without broadband will switch from dial-up to a new broadband service at the same rate as in areas where broadband is currently available.
- This includes data rate, reliability, and customer care.
- Cap-ex is a major factor affecting the success of the BPL business model. Cap-ex is a function of the prices for BPL equipment, network topology, household density, and market share. Manufacturers have released few prices to date. Overall, cap-ex per customer varies greatly. In our analysis, we use a range of cap-ex values.
- This simple model is based on comprehensive and dynamic models that we regularly develop to analyze broadband markets.
- We show value as an internal rate of return (IRR) per customer over the "customer lifetime" (, the average period of time that the customer generates revenue before disconnecting or churning to a competitor).
- Clearly, a specific BPL deployment may not fit neatly into any of these three scenarios. They are intended as illustrative only.
- We selected a panel of 15 wireline communications companies (i.e., we excluded the wireless communications portions). Enterprise value is the value of business operations without debt or cash. We estimated enterprise value per customer by subtracting cash plus debt from market capitalization and then dividing this value by the number of customers. Enterprise value per customer may be compared to NPV per customer calculated with a discounted cash flow (DCF) model. We used DCF to calculate enterprise value for our illustrative BPL provider.
- For example, this may be composed of $40 for local voice service, $20 for long-distance service, $40 for broadband Internet access, and another $40 to $60 for video or cable television. A discount for subscribing to a bundle of services is then applied.
- Pricing for BPL in most current trials is about $30, the same price as DSL in many markets. To attain a $45 ARPU, a BPL provider would require at least some other revenue stream: perhaps VoIP or Internet value-added services (, security, home management, home area network maintenance, e-mail, Web hosting, etc.).
- This analysis holds churn constant at a relatively low value of 1.5 percent per month.
- Of course, this is a simplifying assumption. Actual value for BPL is a function of a variety of factors, many of which we discussed earlier. The portion of the BPL market realized by the utilities in this panel could be more or less than that represented here.
- According to the United Telecom Council. Uses include coordinating power dispatch and transmission, among other functions.
- These applications include real-time pricing, automatic shut on/off, and monitoring of consumer usage to control and balance loads.
- The economics of BPL also will depend on regulatory treatment of cost allocations.
- It also is likely to reduce churn given the known effect of product bundling on customer "stickiness."