Time Warner rationale for bandwidth caps doesn’t add up
Time Warner Cable’s bandwidth caps are such a hit that the company is rolling them out in more cities this year. The caps start at 5GB per month, but Time Warner Cable’s CEO says that they are needed to cover its infrastructure costs. We’re (deeply) skeptical.
By Nate Anderson | Last updated April 1, 2009 8:16 PM CT
Time Warner Cable’s bandwidth cap test in Texas has apparently gone so well that the company plans to deploy the system in Austin, San Antonio, Rochester (NY), and Greensboro (NC) this year. The highest current cap? 40GB per month. The lowest? A mere five.
CEO Glenn Britt tells BusinessWeek, “We need a viable model to be able to support the infrastructure of the broadband business. We made a mistake early on by not defining our business based on the consumption dimension.”
To do that, TWC has been testing plans with 5, 10, 20, or 40GB of data transfer per month, with prices ranging from $30 to $55 a month. Compared to other ISPs, this seems like an abysmal deal, since AT&T’s DSL has no official cap and Comcast offers 250GB per month, generous by comparison. As it expands the trial this year, TWC is adding a 100GB/month level as well, though pricing is not yet available.
Britt’s rationale for the change?infrastructure is expensive?is tough to understand. Cable’s physical plant has been in the ground for years; even hybrid fiber-coax systems have been widely deployed for some time. Internet access simply runs across the existing network, and one of cable’s big advantages over DSL is that speeds can be upgraded cheaply by swapping in new DOCSIS headend gear, with DOCSIS 3.0 the current standard. Compared to what Verizon is doing with fiber and AT&T with its quasi-fiber U-Verse, cable Internet is a bargain (well, for the operators).
But perhaps consumers are insatiable bandwidth hounds who are simply overloading TWC’s system?or perhaps not. The BusinessWeek article notes that only 14 percent of users in TWC’s trial city of Beaumont, Texas even exceeded their caps at all. My own recent conversations with other major ISPs suggest that the average broadband user only pulls down 2-6GB of data per month as it is.
One the one hand, this suggests that caps don’t really bother most people; on the other, it indicates that low cap levels aren’t needed to keep traffic “reasonable” since it’s actually quite low to begin with. The shift toward metered pricing can’t help but benefit TWC’s cable business (unless consumers truly revolt), since low caps will ensure that Internet users think twice before doing data-intensive activities on the Internet such as streaming files from Hulu or downloading HD movies and TV episodes from Amazon and iTunes. TWC’s own cable video service may prove more compelling than the online alternatives because it has no “caps” or additional fees for watching too much TV.
It’s certainly understandable that ISPs would want to curtail abusive users. One ISP told me that a single subscriber last year used 4TB of data transfer?in a single month. This can be a special problem with cable, which uses a shared neighborhood architecture, and one such user can cause problems for others, especially with uploads.
But Time Warner’s low caps make it hard to believe that this is all about infrastructure costs. Internet backbone bandwidth is cheap or free (and plummeting in cost each year), DOCSIS upgrades are a bargain, and competitors in the same business can offer multiple times higher than TWC’s maximums. TWC’s lowest-priced cap, at $30, is cheaper than Comcast’s offerings and more in line with AT&T’s DSL pricing, so perhaps TWC hopes to compete on price for value-conscious customers.
Will customers just switch en masse to uncapped AT&T DSL then? Probably not. Not only is switching a hassle, but AT&T has decided that trialing bandwidth caps in Texas is such a good idea, it should do the same thing. In the same town.