Predicting Financial Success

I’m often asked to provide a rule-of-thumb metric to predict the financial success of a broadband business plan. The two most commonly requested metrics are customer density (how many households are needed per mile of road) or the percentage of customers needed (penetration rate) to make a fiber business plan work.

After having done hundreds of feasibility studies I’ve stopped boiling success down to any simple metric – it’s never that simple. The reality is that there are a number of important variables that have a major impact on operating a successful broadband business plan. Every ISP and every market is different, and one or two variables can have a huge positive or negative impact on a given business plan.

Following are the major variables that can make a difference when building fiber. A similar list can be made for deploying fixed wireless or other technologies.

  • Customer penetration rate. An area with low density might still have great financial results if the penetration rates are high enough. I’ve seen expected penetration rates vary from 40% in some large markets to over 90% in markets with no existing broadband. Changing the expected penetration just a few percentage points can have a big impact on cash flow. This is why we think it’s mandatory to do a survey to understand customer interest in fiber broadband.
  • Labor rates. The cost of staffing varies widely across the country and between companies. and there are places where labor costs twice as much as in other parts of the country. Labor costs also include taxes and benefits which vary widely by state and between ISPs. The staffing structure of the ISP also comes into play since companies vary between lean and staff heavy.
  • Borrowing costs. The interest rates and the term of a loan (15-years versus 25-years) can have a huge impact on a fiber project since the size of the borrowing is usually significant. Things that mitigate borrowing costs such using some equity, getting grants, etc. can have a big positive impact.
  • Prices. Broadband prices can have a big impact. We know that most customers will buy the lowest priced broadband that has a reasonable speed. There is a big difference if this primary product is at $50 versus $60.
  • Cost of the Network. The metric I’m often asked about is the minimum number of households needed per road mile. While customer density is an important factor, there are many other issues that can have a big impact. The cost of building fiber varies widely across the country due to some of the following:
    • The mix of aerial and buried fiber has a giant impact.
    • For buried fiber, the type of soil matters, because the presence of rock adds big costs.
    • Again labor rates, meaning the cost of construction crews. We’ve also seen projects that took federal money that had to pay prevailing wages for rural construction that killed the project.
    • The condition of the poles and the effort and cost needed for make-ready can be a huge factor.
    • The difference between building in the power space versus the communications space on poles can be significant.
    • Choosing PON versus active Ethernet can have a difference, with Active E having larger fiber bundles needing more splicing.
    • One of the biggest impacts is the cost of fiber drops – the two important factors are 1) average distance customers are from the road, and 2) who builds the drops (we’ve seen the labor costs for drops vary by several hundred percent).
    • Building in phases versus building as quickly as possible can sometimes make a big difference.
    • Customer density is important, but the above factors can matter a lot more. Density can also be a tricky number. Consider two examples of companies that would have the same average density but significantly different costs: Company A has no towns and the rural areas average 10 households per road mile. Company B includes one decent sized town but is surrounded by big farms but still averages 10 households per road mile.

Clients always want me to predict the outcome of a business plan before we undertake the needed business models. I’ve learned to not predict. I’ve worked on projects that look to be far more profitable than I would have expected and looked at others that don’t look feasible for some reason. As an example, I recently finished a business plan model where it turns out that the existing poles in the new market were nearly unusable and the alternative of going underground was impractical because of rock. This one factor made it hard to justify building fiber in a market that otherwise would have passed the sniff test using high-level metrics.

Where’s the Top of the Broadband Market?

Last week I looked at the performance of the cable TV industry and today I’m taking a comparative look at broadband customers for all of the large ISPs in the country. Following are the comparative results comparing the end of 2Q 2017 to 2Q 2016.

2017 2016 Change
Comcast 25,306,000 23,987,000 1,319,000 5.5%
Charter 23,318,000 21,815,000 1,503,000 6.9%
AT&T 15,686,000 15,641,000 45,000 0.3%
Verizon 6,988,000 7,014,000 (26,000) -0.4%
CenturyLink 5,868,000 5,990,000 (122,000) -2.0%
Cox 4,845,000 4,745,000 100,000 2.1%
Frontier 4,063,000 4,552,000 (489,000) -10.7%
Altice 4,004,000 4,105,000 (101,000) -2.5%
Mediacom 1,185,000 1,128,000 57,000 5.1%
Windstream 1,025,800 1,075,800 (50,000) -4.6%
WOW 727,600 725,700 1,900 0.3%
Cable ONE 521,724 508,317 13,407 2.6%
Fairpoint 307,100 311,440 (4,340) -1.4%
Cincinnati Bell 304,193 296,700 7,493 2.5%
94,149,417 91,894,957 2,254,460 2.5%

All of these figures come from reports published each quarter by Leichtman Research Group. Just like with cable subscribers, these large companies control over 95% of the broadband market in the country – so looking at them provides a good picture of all broadband. Not included in these numbers are the broadband customers of the smaller ISPs, the subscribers of WISPs (wireless ISPs) and customers of the various satellite services. It’s always been fuzzy about how MDUs are included in these numbers. The MDUs served by the major ISPs above are probably counted fairly well. But today there are numerous MDU owners who are buying a large broadband pipe from a fiber provider and then giving broadband to tenants. These customers are a growing demographic and are likely not included accurately in these numbers.

One of the biggest stories here is that the overall market is still growing at a significant rate of almost 2.5% per year. A little over half of the growth is coming from sales of broadband to new housing units. In the last year, with a good economy the country added almost 1.5 million new living units. But there are obviously still other homes buying broadband for the first time.

There has been a debate for years in the country about where the broadband market will top out. Those that don’t have broadband today can be put into four basic categories: 1) those that can’t afford broadband, 2) those that don’t want it 3) those that are happy with a substitute like cellular broadband, and 4) those who have zero broadband available, such as much of rural America.

It’s obvious that cable companies are outperforming telcos and Comcast, Charter and Mediacom gained more than 5% new broadband customers over the last year. But compared to more recent years the telcos have largely held their own, except for Frontier – which had numerous problems during the year including a botched transition for customers purchased from Verizon.

There are a number of industry trends that will be affecting broadband customers over the next few years:

  • We should start seeing rural customers getting broadband for the first time due to the FCC’s CAF II program. We are now in the third year of that program. The number of customers could be significant and CenturyLink estimates it will get at least a 60% penetration where it is expanding its DSL. I have seen reports from all over the country of fixed cellular wireless customers being connected by AT&T and Verizon.
  • The introduction of ‘unlimited’ cellular plans ought to make cellular broadband more attractive, at least to some demographics. While not really unlimited, the data caps of 20 GB or more per month are a huge increase over data caps from prior years.
  • There are almost a dozen companies that have filed requests with the FCC to launch new broadband satellites. The first major such launch was done recently by ViaSat which will use the new satellite to beef up its Excede product. There’s no telling how many of the other FCC filings represent real satellites or just vaporware, but there should be more competition from satellites, particular those that launch in low orbits to reduce the latency issue. The really big unknown is if Elon Musk will be able to launch the massive satellite network he has promised.
  • Lifeline programs. Companies like Comcast and AT&T have quietly launched low-price broadband options for low-income homes. The companies don’t advertise the plans broadly, but there are communities where significant numbers of customers have been added to these programs.

Broadband Adoption and Income

eyeballThe Brookings Institute just released a report, Broadband Adoption Rates and Gaps in U.S. Metropolitan Areas, that looks at metropolitan broadband rates around the country. The report uses a broad definition of broadband that includes cable modem, DSL, fiber, cellular data, satellite, and fixed wireless service.

The report acknowledges that broadband is still growing and the country saw 2.6 million households add broadband from 2013 to 2014, bringing the overall national penetration rate to 75.1%. But Brookings found that there is a lot of variance in the penetration rates in different parts of the country.

There are metropolitan areas like San Jose where the broadband penetration rate is greater than 88%. The top ten metro markets for broadband has Washington DC in tenth place at 84.7%. But there are a number of other cities that lag behind these national statistics. At the bottom is Laredo, TX at 56.2%, joined at the bottom of the list with places like McAllen, TX, Visalia, CA, Dothan, AL, and Beaumont, TX.

Brookings looked at a number of different factors that affect broadband usage for households. It’s not surprising that household income is a factor. Households with an annual income greater than $50,000 have an 88.8% broadband penetration rate while those with less than $20,000 income are only at 46.8%. Education also seems to be an influence and 91.5% of households with somebody with a bachelor’s degree have broadband while households where nobody finished high school are at 54.1%.

The report did not find a big correlation between race and broadband adoption. While there were cities where blacks or Hispanics have low broadband adoption rates, there were others where they did not. The report concludes that the determining factor is household income and not race.

The report also found some correlation with age and households that have a family member under 18 had a penetration rate of 81.9% while those with everybody over 65 were at 64.5%. But the correlation with age did not hold across all markets and the places where the elderly have lower broadband acceptance seems to be where their income is the lowest. So again, income seems like a more important factor than age.

The report found a few correlations that make a lot of sense. For instance, it found that almost all homes that include a telecommuter have broadband.

Overall the report concludes that metropolitan areas with the highest incomes, with the highest percentage of tech workers, and with the highest average education also have the highest broadband penetration rates.

The report observes that households widely value broadband and that the rate of broadband subscriptions continues to climb. But they conclude that we cannot make the transition to an all-digital society until broadband penetration rate is as ubiquitous as the rates for water and electricity. They conclude that it is going to take targeted assistance programs to get broadband into more homes. While they point to the federal Lifeline and the newly named ConnectHome programs as being a needed part of the solution, they don’t see these kind of programs closing the digital divide. They recommend many more local initiatives, including programs by carriers, to try to get broadband into more households.