Just over a year ago, Time Warner Cable rolled out an experiment in several cities: monthly data limits for Internet usage that ranged from 5GB to 40GB. Data costs money, and consumers would need to start paying their fair share; the experiment seemed to promise an end to the all-you-can-eat Internet buffet at which contented consumers had stuffed themselves for a decade. Food analogies were embraced by the company, with COO Landel Hobbs saying at the time, “When you go to lunch with a friend, do you split the bill in half if he gets the steak and you have a salad?”
In the middle of the controversy, TWC boss Glenn Britt told BusinessWeek something similar, though with less edible imagery. “We need a viable model to be able to support the infrastructure of the broadband business,” he said. “We made a mistake early on by not defining our business based on the consumption dimension.”
This basic argument has a compelling logic—pay for what you consume—and it came with a side order of “implied apocalypse.” Unless a major shift in pricing happens in the near future, TWC’s Internet business won’t be “viable” and the infrastructure won’t keep pace with demand.
This key assertion underlies numerous industry experiments with consumption pricing (AT&T just wrapped up a trial of its own tight data caps in a few test markets, and other ISPs have mooted the idea for years). Few consumers are in a position to judge such claims; maybe the sky is falling. Maybe home Internet use is unsustainable without far more caps or far less data. Maybe those Netflix and Hulu users really are pigs at the broadband trough.
But there’s reason to doubt. Big ISPs usually rely on peered connections to other major ISPs, connections which incur no per-bit cost. As for the cables in the ground, they’ve been there for years. The equipment back at the headend must be installed once, after which it runs for years. Cable node splits and DOCSIS hardware upgrades are relatively cheap. Requesting one additional bit does not necessarily incur any additional charge to the ISP.
If most Internet costs are fixed (and the National Broadband Plan agrees that they are), and if bandwidth is dirt cheap, what “charges” are heavy Internet users ringing up for ISPs like Time Warner? As a New York Times writer summed it up in the middle of last year’s debate:
I tried to explore the marginal costs with Mr. Hobbs. When someone decides to spend a day doing nothing but downloading every Jerry Lewis movie from BitTorrent, Time Warner doesn’t have to write a bigger check to anyone. Rather, as best as I can figure it, the costs are all about building the network equipment and buying long-haul bandwidth for peak capacity.
If that is true, the question of what is “fair” is somewhat more abstract than just saying someone who uses more should pay more. After all, people who watch more hours of cable television don’t pay more than those who don’t.
Mr. Hobbs declined to react to my hypothesis about how costs are almost all fixed costs.
To get some answers, we dug into TWC’s financial statements, then spoke to the company and to its critics. One thing quickly became clear: it’s good to be an ISP. In fact, it’s better than being a cable operator, since there are no multibillion-dollar payments to content creators. As TWC said in a recent filing, “Once again, High Speed Data was our best performing Primary Service Unit category.”
A very good year
TWC’s revenues from Internet access have soared in the last few years, surging from $2.7 billion in 2006 to $4.5 billion in 2009. Customer numbers have grown, too, from 7.6 million in 2007 to 8.9 million in 2009.
But this growth doesn’t translate into higher bandwidth costs for the company; in fact, bandwidth costs have dropped. TWC spent $164 million on data contracts in 2007, but only $132 million in 2009.
What about investing in its infrastructure? That’s down too as a percentage of revenue. TWC does spend billions each year building and improving its network ($3.2 billion in 2009), but the raw number alone is meaningless; what matters is relative investment, and it has declined even as subscribers increased and revenues surged. “Total CapEx [capital expenses] as a percentage of revenues for the year [2009] was 18.1 percent versus 20.5 percent in 2008,” said the company a few months ago.
In fact, CapEx has declined for the industry as a whole. As the National Broadband Plan noted, the big ISPs invested $48 billion in their networks in 2008 and $40 billion in 2009. (About half of this money can be chalked up to broadband; the rest of the improvements were done to aid cable or phone service.)
To recap: subscribers up, revenues up, bandwidth costs down, infrastructure costs down. This might seem like a textbook case of “viability”; what were execs like Britt and Hobbs talking about last year when data caps were held up as a necessary safeguard against doom?
It’s about bandwidth labor
Several months ago, while on a business trip to Manhattan, I entered a nondescript building near the Flatiron building and rode the elevator to the top. Inside was one of TWC’s main New York operations centers, hosting an astonishing array of cable and Internet gear. But the real showpiece was the monitoring room, a darkened room with control hardware, computers, and a wall of TVs showing every cable channel currently running out over TWC’s network.
It looked brand new and obscenely expensive. Engineers slipped in and out in silence. A huge pile of boxes on the floor held a new set of replacement TVs. When I make my career shift from ink-stained wretch to Evil Genius, this is exactly the sort of room I will build in order to plot my world domination.
“It’s not a cheap endeavor to run a network like we do,” said TWC’s tweeting VP of Public Relations, Alex Dudley, when I had spoken to him the week before. Here was an obvious reminder of what he meant.
This point is hammered home by most ISPs—the billions of dollars of new investment, the upgrades, the capacity building. But it’s a point only meaningful in the context of revenues. A company’s financials don’t lie, and TWC’s financials showed a declining percentage of revenue spent on infrastructure even as profits soared and bandwidth costs dropped. I pressed Dudley on Glenn Britt’s statements about viability. If these are problems, they’re problems most companies want to have.
Britt is “a long-term-view kind of guy,” Dudley said, and with broadband use surging, “all of the ancillary costs affiliated with broadband are going up.” This didn’t quite compute, since bandwidth and network investment were actually declining as percentages of revenue.
But according to Dudley, those two numbers don’t tell the whole story. TWC’s single biggest expense for Internet access is not network investment or bandwidth. It’s labor.
As Internet use increases, TWC techs, engineers, and executives need to make adjustments such as DOCSIS upgrades at the cable company headend or “node splits” that divide a shared cable loop in two when bandwidth use hits certain metrics. Paying all of these people costs money, and those costs increase as the network is more heavily used.
(This differs from how Landel Hobbs defended the company at the height of the backlash against TWC last year. He quite clearly stated that bandwidth creates real costs for the company and that those need to be covered. “For those who want to use a tremendous amount of bandwidth, there should be a charge, because that costs money,” he told the Times.)
Besides, Dudley said, TWC does invest plenty of money in raw infrastructure. If CapEx spending was down in 2009, chalk it up to the company’s video subscribers, which declined a bit over 2008. One big piece of TWC’s CapEx is buying all those cable set-top boxes (which are then rented on a monthly basis by subscribers), and fewer subscribers mean fewer new boxes to purchase.
The company’s critics couldn’t disagree more with this entire line of argument.
“Greed”
“Hogwash,” says Free Press research director S. Derek Turner. “Their OpEx [Operating Expenses, which includes labor] is not growing; if anything, it’s steady. Their CapEx is decreasing both in overall terms and as a percentage of revenue.”
Turner has little patience for the “woe is me” arguments that ISPs trot out to defend a shift to data caps or per-bit pricing. Free Press, a constant critic of the big ISPs, says it has no philosophical problem with a move to a consumption model for broadband—but such a shift should accurately reflect costs, not serve as an excuse to gouge customers by companies already swimming in cash.
TWC’s data capping trial in 2009 featured “literally ridiculous overage amounts that had no relation to underlying costs,” Turner said. And the danger isn’t just to consumer pocketbooks, it’s to the entire Internet ecosystem. Who will start using the next high-bandwidth YouTube or Netflix when doing so results in big fees? If not done right, consumption pricing “will cripple innovation.”
Turner concedes that networks cost money to build and maintain, but he argues that the costs are wildly overstated. For instance, Comcast is one of the ISPs furthest along with DOCSIS 3.0 upgrades, which do require a labor-intensive card swap at the headend and new modems in people’s homes. But even as it makes this investment, the company’s OpEx and CapEx are declining. As for node splits, many are “virtual” these days and don’t require much labor.
Bandwidth has become dirt cheap; despite the fear-mongering about the “exaflood” and the “zettaflood” and (presumably) the “yottaflood,” bandwidth costs drop significantly every year. As the National Broadband Plan noted earlier this year, international bandwidth has grown by 66 percent each year for the last five years—but the cost of IP transit has dropped 22 percent a year at the same time.
Congestion can happen even on networks with tremendous bandwidth, but consumption pricing doesn’t generally care about congestion (if it did, ISPs could exempt all traffic in the middle of the night, for instance, when congestion is generally absent).
So why the push for consumption pricing? Turner has his own theory.
“This is nothing more than greed,” he says. “The industry may be maturing, and therefore margins aren’t rapidly increasing the way they were.” Consumption pricing could be a way to boost margins. As for ISP complaints that heavy users cost them more money, those are just “excuses that they give.”
Still rare
But low data caps are still not widespread in the US wireline business. That’s due in large part to public resistance to the idea. When TWC expanded its capping trial last year, it took only a couple of weeks for a New York Congressman (the now-disgraced-and-resigned serial tickler of his male staffers, Eric Massa) to pledge a “Broadband Internet Fairness Act” that would “prevent job killing broadband downloading caps.”
Despite a few trials (sorry, Beaumont, Texas), consumption Internet pricing remains unusual. Unless ISPs find a way to make a more compelling case for its necessity—and its fairness—it may remain so.