The telcos are grumpy about SpaceX’s Starlink satellite broadband service. Through a front organization, the National Rural Broadband Association (promoted hashtag #ruraliscool and I am not making that up), they have magnanimously performed an analysis of Starlink and submitted it to the FCC. They are concerned Starlink may have a “capacity shortfall” in the year 2028. Presumably this civic-minded analytical effort is unrelated to the telcos not wanting to share billions of Federal rural broadband subsidies with the upstart satellite network.
My own submission is that rural telcos might not want to call attention to “capacity shortfalls”.
A Real and Present “Capacity Shortfall”
The performance snapshot on the left is Starlink, while the “capacity shortfall” on the right is DSL service from CenturyLink in rural Washington State (which they actually call “broadband” but inexplicably omit the ironic quotes). The key takeaway, without getting too mathy, is that megabits are a thousand times better than kilobits. Starlink also has significantly less latency, despite bouncing packets through low earth orbit. For everything but video calls, Starlink delivers an experience comparable to urban megabit broadband (the periodic fumbled satellite hand-off, however, plays havoc with Zoom calls).
That blazing 180 kilobit performance is actually CenturyLink on a good day. They bill me for 256 kilobit DSL service (their fastest available!), but suffer major congestion, with actual throughput often closer to zero bits. Text browsing and email are barely possible, never mind using the extant web or even thinking about streaming video. It is inconvenient for me while idling at my rural retreat, but for people living nearby year-round who have been trying to do remote work and school during a pandemic, it is a material “capacity shortfall”.
Telcos are infrastructure businesses (this is where “man with CAPEX hammer” starts hammering). CenturyLink tells the SEC they invested 14-18% of operating revenue in CAPEX over the last decade. They spend over three billion dollars a year on CAPEX (twice what a clown company like Oracle spends, but a pittance compared to the cloud companies). Unlike technology companies, they spend very little on R&D (though here they spew random tech buzzword combinations in an effort to suggest otherwise). The mystery, however, is where this spend is going.
CenturyLink has failed to deliver even its (ridiculously slow) advertised bandwidth since at least 2007, when I first called to complain. You’d think over almost 15 years, they’d have made discernible improvements in the network. In fact, it has gotten worse as congestion has increased (their only apparent mitigation is to stop adding new customers in the area).
At $18-22k per mile, I estimate it would cost roughly $600,000 to lay the necessary fiber back-haul to solve the congestion problem (while also opening the opportunity to sell non-ironic broadband service to thousands of new customers). Some back of the envelope math suggests this is less than 5% of CenturyLink’s total revenue from that isolated community over that period, so well below their expected CAPEX investment.
Stasis as a Service
Instead of fixing network bottlenecks known for decades, CenturyLink has spent that money and much more on multiple corporate rebrands (most recently from CenturyLink to Lumen, which evokes either an apt anatomical orifice or a failed Lululemon ramen brand), and stadium sponsorships.
CenturyLink has also drunk deeply from the trough of rural broadband subsidies (here’s one worth $506 million that, needless to say, is experiencing a “capacity shortfall” with respect to fulfilling their obligations). And I’ll just get crankier if I look up their spending on lobbyists over this time.
Rather than an “OODA loop”, CenturyLink operates on a negative feedback “OOPS loop”:
Underinvesting in the network undercuts customer satisfaction. Instead of improving the network, they rebrand the company in a misguided attempt to duck responsibility and reset the customer relationship. The rebranding diverts still more resources from the network. And repeat.
Jeff Bezos has an evocative description of companies losing their way: “Day Two is stasis. Followed by irrelevance. Followed by excruciating, painful decline. Followed by death.” CenturyLink is the perfect illustration for Day Two in the calendar of corporate torpor.
Mere stagnation would be preferable to a network whose performance is deteriorating in both absolute and relative terms. CenturyLink’s core competency should not be rebranding. As a network company, the network should be – to put on my French chapeau for a moment – their raison d’être. The contrast with companies that obsessively pursue their raison d’être by every possible means is stark. Think of Netflix investing in content. Or Amazon and every bonkers thing they do. Or any Elon Musk endeavor. They are absolutely certain better is better and marshal every possible penny to make it so.
Part of the blame rests with the regulatory apparatus. CenturyLink is a government-regulated monopoly that is failing to deliver on its existing (meager) commitments, much less even contemplating improvements to the status quo. Is their lethargy in spite of their lavish government subsidies and protection, or because of it? The Federal rural broadband subsidy programs of the last 25 years have spent real money (on the order of $100 billion) and inspired many soaring platitudes, but delivered very little rural broadband. These programs are somewhere between deeply flawed and complete failures. “Capacity shortfall” plagues government too.
The inertial gloom suffocating telecom is a cautionary tale for podcast populists and hipster antitrust hobbyists with simplistic regulatory prescriptions for Big Tech (supposition: the easier it is to chant your preferred remedy, the less likely it is to work). Not only do flawed and poorly defined regulations bring unintended or even the opposite of intended effects (AdiEU GDPR! G’day Australia!), but also risk inducing the kind of regulatory capture and innovation-obliterating stasis that pervades telecom.
But you may say, telecommunications networks are a natural monopoly, networks are hard, backhoes don’t follow Moore’s Law, etc., etc., blah, blah. Yes, it is easy to make excuses for inaction as the incumbents so ably demonstrate. Yet Elon somehow finds a way.
I’m not even remotely an Elon fanboy. But his relentlessness to do hard things from scratch is such an utter contrast to the telcos (and so many of our other institutions that also suffer from “capacity shortfall”, but that is a broader rant for another day). Elon certainly has his faults. Like the telcos, he’s playing the subsidy and tax credit game, letting taxpayers subsidize both incumbents and challengers. Schedules and specific deliverables in Elon World are at best well-intentioned. And Starlink may well have scalability issues (team #ruraliscool are not the only party with concerns on this front).
But it is brazenly hypocritical for a gang of do-nothing telcos to carp about a Starlink “capacity shortfall” given their own multi-decade record of indolence.
No More Excuses
I’ve hit “capacity shortfall” in my patience with CenturyLink and the rural broadband industrial complex. If you are familiar with the baroque workings of the Washington Utility and Telecommunications Commission, and/or are fed up with serial re-branders masquerading as network operators, email me at capacityshortfall [at] platformonomics.com, to discuss WHAT IS TO BE DONE…
Thanks to my reviewers who looked at this. Errors remain mine.