Why Overbuilding is So Rare
Both for-profit and non-profit entities long have looked at whether it is feasible to compete against cable TV and telcos for internet access, video and voice services using their own facilities, with the objective of providing higher-quality service (internet access speeds, for example) and lower cost.
Generally speaking, more success has been found by suppliers who, in fact, do not “overbuild” telco and cable TV networks across an entire metro footprint, but pick niches (apartments and high-rise living units; neighborhoods and suburbs).
That is the same “cherrypicking” strategy used by business service specialists, who only operate in business districts and office parks.
A study of potential lost market share if municipal internet access networks were built in Seattle and Fort Collins, Colo. assumes Comcast, the leading internet service provider in those cities, would lose 20 percent to 30 percent of its current customer base. That is a reasonable assumption, given other experience in the U.S. market.
Assuming 20 percent loss of share and $50 average revenue per account, Comcast might lose $1.38 million a month, or $16.6 million in annual revenue, in Seattle, or $373,500 a month ($4.5 million annual) in Fort Collins.
Seattle has about 340,479 housing units, of which 327,188 are occupied at any particular time. At an average network cost of about $700 per location passed (Seattle is highly urban, so costs to build a new cabled network are likely higher than that), the core network might cost $238 million.
Connecting an actual customer might cost $300 per location, if internet access is the only service offered.
At 20 percent take rates, the new provider might have 65,437 customers. That represents activation costs of about $19.6 million.
Assume the new provider sells internet access at the same average of $50 per account (so there actually is zero savings for each customer), but that delivered bandwidth is a gigabit per second, with annual revenue of $16.6 million.
Assume the new operator has a 40-percent gross margin (revenue minus direct costs), implying annual net cash of about $6.6 million (before taxes, interest, depreciation and amortization).
And, at five percent cost of capital, interest payments on network construction ($238 million) alone are $12 million a year.
Without getting any more detailed, the business case simply does not work.
It is fine to criticize ISPs for not delivering higher-quality services at more affordable prices. But the costs of building internet access infrastructure, using any cabled method, plus reasonable market share expectations, explain why facilities-based competition is so difficult.
Ubiquitous third suppliers might be an unreasonable expectation. Limited builds in some neighborhoods, suburbs or high-rise living units might have sustainable economics. But advocated likely hope in vain for the emergence of full citywide competitors.