What’s Holding Netflix Back?

Broadband usage fees and limits are back in the news again: the New York Times editorialized against them recently, and the Washington Post’s Cecilia Kang reported that “experts” have concerns about them.  We covered this issue in a previous post on AT&T’s usage limits back in March, but it’s worth revisiting the issue to examine the latest round of complaints.

Broadband usage caps or service tiers follow from the principle that the more of a resource you use, the more you pay. This principle applies to practically every commodity, service, or utility that we use and isn’t the least bit controversial in the abstract. Broadband has traditionally been a flat-rate, unmetered service in the United States, but not so in the rest of the world. This was possible because of infrastructure competition in the U. S. and also because patterns of broadband usage have been very uniform. As Internet applications become more diverse, the patterns are changing and usage-based pricing is a response. It makes no more sense for light consumers of bandwidth to subsidize heavy users than it would to make those who eat salad for lunch subsidize the consumption of Porterhouse steaks by others.

What’s the Beef with Caps?

Critics fear that usage caps are aimed at stifling the growth of Internet TV in general and Netflix in particular. The Times says that usage caps are unfairly applied to Internet streaming but not to garden-variety television viewing over cable and the cable-like telco services such as Verizon’s FiOS and AT&T’s U-verse:

Caps can be used anticompetitively — to discourage the use of services that rival an Internet service provider’s in-house offerings. For instance, AT&T points out that Netflix hogs 30 percent of peak-hour Internet traffic in North America. Netflix also competes with television offerings on AT&T’s U-verse network. Watching TV on U-verse does not count against the data cap. Streaming Netflix does.

And Kang raises a similar point, albeit less stridently:

Why would a user watch HD feature films on Hulu or Netflix if they could blow past their monthly data limits? And is it fair if Comcast or AT&T doesn’t charge by the bit for their own television services?         

Both writers presuppose a fact that’s not really a fact. Whether it’s fair to exempt particular services from overall usage limits depends in part on the nature of the service. Many different kinds of traffic transit modern broadband networks, with many different costs of delivery and potential for congestion. Live TV in particular does not stress a cable network the same way that Netflix does. Live TV is broadcast throughout the cable system on dedicated channels that can be watched by millions of viewers at a time without any effect on network load. If every TiVo in America is tuned to the live broadcast of American Idol, the cable network is no more stressed than if only one person is watching.

Netflix doesn’t work this way. If everyone of Netflix’ customers uses their system at the same time, the Internet would simply back up in a colossal fashion. This is because Netflix is a personalized service in which every user gets a unique copy of the movie or television program they’re watching, regardless of how many other people may be viewing the same content. So the comparison of live TV to streaming isn’t an apples-to-apples comparison.  The pertinent comparison would come between cable partner services such as HBO Go or Hulu and Netflix, and here the complainers have no case: These services, which load cable networks the same way that Netflix does, are not exempt from usage caps. This is the fact that the writers get wrong.

To Peak or Not to Peak

The Times also points out that usage caps are a blunt instrument that apply to both peak load and non-peak load hours. This is a fair point that we’ve made ourselves: An exemption of traffic from usage caps that doesn’t contribute to congestion has much to recommend it.

The Times fails to appreciate that relaxed off-peak usage limits aren’t going to help Netflix in any substantial way because streaming is not a time-shiftable service. Unlike P2P piracy or Amazon digital video rental, Netflix doesn’t transfer a file to the end user’s computer that can be brought out of a local archive and played at will. Netflix is a real-time, peak hour experience that can only be transferred at the same time that it’s viewed. While the theory of exempting non-peak usage is sound, the use of this argument in the context of a plea to protect Netflix from predatory usage limits falls well far of the mark.

Faulty Arithmetic

The Times makes some other observations about the cost of bandwidth that are faulty as well, asserting that upgrades to current network systems are extremely inexpensive and a better means of addressing the broadband capacity crunch than usage caps. These arguments circulated the network neutrality debate with great vigor, but the passion with which they’re delivered doesn’t make them any more sound today than they were five years ago. Broadband networks are shared resource systems in which a very small number of users can sap capacity faster than the service provider can add it. Bandwidth is consumed by software, which has little or no cost, but it’s provided by hardware that has substantial unit costs that correlate with distance.

The current capacity crunch was created by video streaming systems that move packets a few feet before handing them off to the far-flung broadband networks that then must move them hundreds of miles.  The cost imbalance between the two sides of the broadband equation motivates usage caps, and it goes unmentioned in the Times’ editorial and Kang’s interviews with policy experts.

In our previous analysis, we showed that the usage limits currently employed by Comcast and AT&T don’t impair the ability of American consumers to use the Netflix system:

The accounting is straightforward: According to Netlfix, they stream videos at an average rate of 1.5 Mbps into the AT&T network, which works out to 675 Kilobytes per hour. That comes out to 222 hours of streaming TV per month, or 7 hours a day plus change, for ADSL users. For U-Verse users, the cap allows 370 hours of streaming per month, which is 12 hours a day…

Netflix streams at a higher rate into the Comcast network, but the end result is well over seven hours a day of video programming before the user encounters the limit.

Barriers to Netflix’s Growth

Netflix is in the news today for two legitimate reasons: They recently announced a 60% price hike that enraged its customers; the reaction was termed a “social media firestorm” by multiple news organizations. Netflix has up to now given streaming video users a free ride on the charges paid for DVD rental, but they would like to grow the high-margin side of their business, streaming, while shrinking the low margin side of the business, DVD rental.  Netflix also reported slower growth in new subscribers in the most recent quarter, adding about as many new subscribers as analysts expect them to lose when their price rises at the end of the summer. It’s likely that the costs of the content Netflix streams will rise substantially over the course of the next year as well. Over the last five days, Netflix stock has fallen from $280 to $269 a share.

The only conclusion that should be drawn about Netflix and usage caps is that caps are a non-issue. There are many things that might prevent Netflix from growing as astronomically as it has in the past two years, but usage caps are not high on the list. Netflix has traditionally been a very customer-friendly organization with a well deserved reputation for honesty. It would be shame if they were to tarnish that image by attempting to shift the blame for customer anger at an outsized price increase onto the ISPs that have enabled their service to flourish so far.

Print Friendly

About the author

Richard Bennett is an ITIF Senior Research Fellow specializing in broadband networking and Internet policy. He has a 30 year background in network engineering and standards. He was vice-chair of the IEEE 802.3 task group that devised the original Ethernet over Twisted Pair standard, and has contributed to Wi-Fi standards for fifteen years. He was active in OSI, the instigator of RFC 1001, and founder, along with Bob Metcalfe, of the Open Token Foundation, the first network industry alliance to operate an interoperability lab. He has worked for leading applied research labs, where portions of his work were underwritten by DARPA. Richard is also the inventor of four networking patents and a member of the BITAG Technical Working Group.