Open Letter to Senator John F. Kerry — Price Tiers Benefit Consumers

April 17, 2009

Senator John F. Kerry
218 Russell Senate Office Building
Washington DC 20510

Dear Chairman Kerry:

You and your colleagues will very likely soon receive a petition from members of Free Press urging you to launch “an investigation of Time Warner Cable’s plans to impose an unfair penalty against Internet users.” The correspondence will be motivated in substantial part by a recent solicitation to its members containing misstatement of facts, by Free Press, regarding causes and effects of “metered” billing trials.

The occasion for the controversy follows: Time Warner Cable recently announced its intention to expand its (Beaumont, Texas, metered billing trial) to four more markets later this year. The company will restructure its rates into tiers each of which reflects a different usage level by a subscriber. There will be seven rate/service tiers (1, 10, 20, 30, 40, 60 and 100 gigabits with different speeds). The slowest, least bandwidth intensive tier (one gigabit per month with 768 kbps upstream and 128 kbps downstream) is likely to apply to about a third of TWC users and will be priced at $15 per month. The top end tier for “power users” (100 gigabits per month at 10 mbps downstream and one mbps upstream) would be priced at $75 per month. The incremental usage charge for all tiers would be $1 per gigabit over the tier limit.

The proposal and the reaction of Free Press and others have motivated substantial controversy, one result of which was the announcement by Time Warner Cable to postpone the trials pending continuation of its customer education process designed to provide better information to consumers about the details and implications of the plan. According to a company release yesterday, it is developing measurement tools that will permit consumers to monitor their use, thereby eliminating concerns that consumers will be billed for bandwidth usage of which they were unaware.

We support this delay to make certain that consumers are fully informed and equipped to make rational choices about the level and quality of services they buy. More generally, we believe that the form of rate differentiation in the TWC plan serves consumers’ interests by providing them the opportunity to choose a plan that best fits their individual needs. Rather than forcing every consumer to pay an equal amount to cover network costs, differentiation means consumers pay for what they use – no more and no less – as with most other products and services in our economy.

Dramatic differences in usage pattern by Internet customers – enabled by changes in technology – are leading to circumstances in which flat fee plans mean that occasional, modest or average users of Internet services will pay higher rates in order to subsidize a relatively small group of subscribers who customarily use hundreds or thousands of times as much scarce bandwidth capacity as the average Internet user. Contrary to Free Press contentions, charging users who demand extraordinary bandwidth for the costs they impose on the system is the best way to spare the rest of us from paying additional costs, in effect a tax, for the extraordinary appetites for bandwidth of a relatively small group of users.

The basis for our sharply contrasting view of what is good for consumers follows from three facts. First, different uses of the many applications availed by an Internet connection require dramatically different amounts of bandwidth. Voice, audio, regular video, HD video and other information “bundles” require widely differing amounts of capacity. Thus, a minute of downloading of high-definition television uses about 10,000 times as much bandwidth as an 800-word e-mail message. A minute of ordinary video streaming uses about 2,500 as much as that same email. Audio streaming, gaming, assorted multimedia web interactions, and related uses each consumer large multiples of the bandwidth required by simple voice or messaging or search applications.

Secondly, different users have different usage patterns. Some of us use email primarily and surf for information to download. Others focus on large data files, graphics, photos or music. Still others use their connections intensively for video file transfers. A variety of usage studies indicate that the top ten percent of bandwidth users account for more than half the total, while the bottom 40 percent of users generate less than five percent of the traffic and demand for capacity. We can think of no sound public policy basis for requiring the 40 percent who consume little to subsidize the habits of the five percent who use the lion’s share of available bandwidth.

Third, network costs vary with capacity and usage. Higher usage tends to impose higher costs. There is no such thing as free bandwidth. Capacity is scarce, capital intensive and, according to securities analysts who advise potential investors, investment is risky.

These three facts are essential to any accurate assessment of the interest of all consumers as a group. The TWC trials and those of other ISPs involve tiers of service defined by monthly rates of use and priced in ways that enable users to pay in accordance with their own usage, which is consistent with how most services are priced. Free Press argues that heavy users are “penalized;” but the more accurate description is to say they are paying for the amount of service they choose to consume.

A Free Press spokesperson has erroneously concluded that metered pricing of the sort being tried by Time Warner is inconsistent with efforts to boost the economy by expanding Internet coverage and usage. We agree with Free Press that investment and capacity expansion are critical to making the Internet faster and more widely available as means for stimulating the economy. But, investment and expansion of Internet capacity depend, in part, on the ability of service providers to price services in ways sufficient to generate the enormous cash required to underwrite the necessary capital expenditures. Those costs are most appropriately imposed on users responsible for creating the demand.

TWC has publicly justified the rate restructuring and increases for extraordinary usage on grounds that doing so is required to maintain and upgrade its network and to cover rising traffic-sensitive costs. A quick look at the financials of TWC indicates that over the past three years the company has expended over 70% of cash flow generated by its operations in the form of capital expenditures, a ratio that is in line with investment behavior of Internet service providers in general and well above investment rates of other firms in the Internet value cluster (applications, software, content, etc.).

Monopoly profiteering? TWC shareholders would be delighted if that were the case, but again the data indicate otherwise. The profit margin for TWC for the past twelve months was minus 43%; its return on equity was minus 35%; while its return on total assets for past year was 3.7%. The return on investment by TWC is almost identical to recent rates on 20 year US Treasuries that are widely regarded as risk free. The point of course is that investors are reluctant to underwrite the risks of investing in broadband networks on the basis of returns that do not reflect such risk adequately.

Consumers have quite diverse interests in Internet services pricing. Policy should focus on the interests of all users, not just a few. This is particularly the case here where needs and preferences of senior, low income, rural users contrast sharply with the habits of younger, better off, highly educated, urbanized consumers. Surveys by Pew and others indicate that many Americans do not use broadband services in part because they are too expensive. New users are likely to be use meager bandwidth and should not be burdened with costs imposed by veteran, heavy users. There is no basis for making “Aunt Minnie” and other low volume, occasional users pay for the capacity required by extraordinarily intensive use by a small minority of heavy users.

Dr. Larry F. Darby
American Consumer Institute
Center for Citizen Research
1701 Pennsylvania Ave, NW, Suite 300
Washington, DC 20006

cc:
Daniel K. Inouye
Byron L. Dorgan
Bill Nelson
Maria Cantwell
Frank R. Lautenberg
Mark Pryor
Claire McCaskill
Amy Klobuchar
Tom Udall
Mark Warner
Mark Begich
John Ensign
Olympia J. Snowe
Jim DeMint
John Thune
Roger Wicker
Johnny Isakson
David Vitter
Sam Brownback
Mel Martinez
Mike Johanns

FacebooktwitterredditlinkedinFacebooktwitterredditlinkedin

5 thoughts on “Open Letter to Senator John F. Kerry — Price Tiers Benefit Consumers

  1. The problem with the proposed Time Warner bandwidth caps is the timing and the size of the caps strongly suggests that the real motivating factor is to slow the growth of disruptive technologies like streaming video services and VoIP. Sanford C. Bernstein, a research firm, puts conservative estimates of potential costs for users who view streaming online video for roughly 7.25 hours a week at up to $200 in overage charges a month under Time Warner’s proposed service. Bernstein’s estimates, mind you, are based on streaming video services alone and don’t include other media transfers, such as purchases from iTunes.

    The solution is probably going to be legislation that will further open cable networks to competitive providers, and in doing so create an opportunity for other business models.

  2. Mr. Darby,

    I appreciate your concerns and agree with several of your points in your letter, however I am concerned that you may be taking Time Warner Cable’s word from their press releases as solid fact including their own confusion about bandwidth capacity versus Internet data consumption.

    Please allow me to offer a simple comparison to help illustrate the difference between capacity and usage.

    Webster’s dictionary gives one definition of capacity as: “the maximum amount or number that can be contained or accommodated .”[1] Using this definition as it relates to the Internet and data, we say that the capacity of a network connection is equal to it’s bandwidth, for example 10 megabits per second (Mbps). Likewise Webster defines bandwidth as: “the capacity for data transfer of an electronic communications system.”[2] Please note, I take exception to their example as it is contrary to the definition.

    Other things we are familiar with that have capacities are rooms (maximum capacity), highways (maximum number of cars at a time), and pipes (maximum amount of fluid that may pass per second). All of these capacities refer to simultaneous usage, not total usage.

    The capacity of a room is not diminished when one person enters and subsequently leaves. It is only diminished as long as they remain in the room. Likewise a highway’s capacity for traffic is not reduced as cars enter AND leave, rather only when cars enter and remain.

    So-called bandwidth caps as proposed by Time Warner Cable, AT&T, Comcast and others are actually not related to bandwidth at all. They do not limit the speed of your network connection – the maximum simultaneous data transfer – they limit the total amount of data.

    That limit is like specifying that a room has a maximum capacity of 1,000 people and after 1,000 people have entered (and left) the room no more may do so. Or in the case of Time Warner, any additional people to enter the room will pay an additional fee.

    This extra fee does not do anything to alleviate the problem of having too many people in the room at once. It only discourages people from going into the room in the first place for fear of being the one that goes over the 1,000th person limit.

    I hope this illustrates for you how “bandwidth” caps as proposed are not a solution to a capacity problem.

    I agree that perhaps an emergency tier should be created for highly time-critical information such as medical records to be delivered, but I strongly disagree with the notion that the access you pay for is somehow subsidizing another’s usage.

    I also request you publish the raw data used in the studies to which you refer so they may be reviewed publicly as a scientific paper would be reviewed by others in the field before it is accepted.

    Thank you,
    Brion Swanson

    [1] http://www.merriam-webster.com/dictionary/capacity%5B1%5D
    [2] http://www.merriam-webster.com/dictionary/bandwidth

  3. you know what would benifit consumers more why not put a tax on the cable companies who implment caps say 5 dollors per customer this would be good for end users and id rather my money go to congress or al quida then the greedy demon known as time warner cable

  4. Some good points. I would not be surprised if a form of tiered service would actually stimulate widespread availability of low-level internet service provision. Though Brian Swanson makes a good point with the capacity vs. data consumption argument, it remains that, provided bandwidth windows are staggered appropriately, the tiered pricing schedule will have the desired effect of bringing bandwidth consumption more in line with fees accrued.

    However, it is undeniable that the bandwidth caps are absurd in scale. Even moderate bandwidth users like myself would exceed the ‘power user’ rating in the first week of service. The measurement of overall consumption in megabits is very misleading. 100 gigabits of transferred information over the network is roughly equivalent to 10 gigabytes of actual content, possibly substantially lower depending on how packet overhead is accounted for. Internet radio at a decent quality will generally stream at a content rate of about 100 kilobits/s (roughly 10 kilobytes/s); at this rate listening for 8 hours a day (a not-unreasonable expectation for the near future of internet usage, considering that many work all day with the radio on) would consume the entire allocation. Streaming video bandwidth consumption would begin at around 5x this number, and this does not account for any additional usage (such as email, web browsing, and miscellaneous file transfers). Additionally, peer-to-peer technologies (such as Skype, Grooveshark, BitTorrent, and several online television streaming sources), in which an end user’s computer may be fetching, processing and uploading information which is not directly related to that user’s personal use, could increase the amount of bandwidth consumed by a typical user hugely, likely without their knowledge.

    In fact, the knowledge problem is even more serious than that; any user in the 1 gigabit tier could, at any time, through the discovery of any single online service, easily exceed their limit within a single day. As the internet is growing in breadth and technology constantly, anyone could catapult from a low-usage pattern to a high-usage pattern at any time. In this way the internet differs radically from other potentially consumption-based services such as telephone or cable television.

    A parallel situation might be if a cable carrier were to mix high-definition and low-definition content into their cable service, and charge users by their bandwidth consumption. This could result in an unsavvy consumer purchasing a low-tier service level receiving a bandwidth limit warning on May 3rd because they started watching a new nightly television show–and in order to keep their bandwidth usage within their limit they would need to forego watching television for the rest of the month (perhaps not a bad idea from time to time, but problematic nevertheless)?

    A much more reasonable tiered pricing model would scale the data allocation exponentially while still increasing cost linearly, and where the customer is simply charged for the next tier if they exceed their data allocation:

    allocation price
    1 gigabit (Gb) $15
    10 Gb $25
    100 Gb $35
    1 terabit (Tb) $45
    10 Tb $55

    …and so forth. This is more reasonable in the context of the technology in question, because having links in the first place costs significantly, but having larger links (larger fiber bundles, or multiple bundles following the same path) is only slightly more expensive. It is also more reasonable in the context of information technology changes, which cause the average user’s bandwidth consumption to rise exponentially (through moving from text to images, images to audio, audio to video) more often than it rises linearly (through the actual amount of time spent in front of the computer).

  5. you know what would benifit consumers more why not put a tax on the cable companies who implment caps say 5 dollors per customer this would be good for end users and id rather my money go to congress or al quida then the greedy demon known as time warner cable

Comments are closed.