Why Monthly Broadband Usage Caps Won't Really Work (But Usage-Based Billing Will)
You’re a broadband provider. Do you want to portray yourself as (a) an unabashedly greedy monopolist/duopolist or (b) a clueless monopolist/duopolist whose lame network can’t meet demand?
Or what the heck, take them both, and maybe even throw in the idea that cable and telco TV providers are hoping to inhibit usage of Internet alternatives to their multichannel video packages.
These are the explanations offered by various bloggers for why Internet service providers are looking at consumption-based billing, as indicated in recent posts by GigaOm’s Stacey Higginbotham and BroadbandReports’ Karl Bode. (Oh, and who could forget the StoptheCap! guy, who’s making a career directing his bloggravation at The Man.) I don’t consider myself an edgy blogger type because, as everyone knows, I am The Man.
The idea of charging broadband customers based on what they use is still in play, even after Time Warner Cable’s famous P.R. disaster earlier this year that practically resulted in congressional hearings (see Time Warner Cable: Three Mistakes on Usage Pricing).
More recently, Verizon has revealed that it’s philosophically on the same page as TWC. Verizon’s Richard Lynch in September said that “a flat-rate infinitely expandable service is unachievable” (see Verizon CTO: Metered Bandwidth Is Inevitable). Charter CEO Neil Smit told Bloomberg yesterday that the MSO is going to be looking at consumption-based billing, too.
To Higginbotham, Verizon’s acknowledgement that some kind of usage-based pricing is on its way “was kind of like watching your favorite indie rocker sell out. Why would Verizon, which is building out a fiber-to-the-home network, plan to eventually move to some sort of consumption model?” (Which is actually pretty funny. Now I’ll be thinking of the FiOS guys as über-nerd rockers We Are Scientists.)
Her answer: Verizon just wants to make more money by charging more for the same service. What an outrage! It’s not like the company spent billions and billions to build out their network and needs to recoup that investment (nota bene: sarcasm).
Anyway, my point (talk about burying the lead) is that consumption-based billing models are inevitable mainly because Internet demand is shooting through the roof. Internet bandwidth usage is expected to grow at a 40% compound annual growth rate from 2008 to 2013, according to Cisco’s VNI forecast — quintupling over that time period.
Today’s broadband networks — not even FiOS — are not constructed to deliver peak theoretical demand and adding more capacity to the home or farther upstream will require investment.
Moreover, monthly usage caps alone — e.g., Comcast’s 250-Gigabyte per month limit — will ultimately not do anything to address the congestion introduced by higher and higher per-user bandwidth consumption.
Why? Because, according to recent analysis of Internet usage patterns, there’s a “primetime” for broadband consumption, which a huge monthly cap won’t do anything to address. A monthly 250 GB cap only filters out the bit junkies who are literally sharing terabytes of stuff; granted, they unequivocally use more than their fair share but capping them doesn’t solve long-term congestion issues.
Network-management equipment vendor Sandvine says that between 7 and 10 p.m. in any given region around the world, the usage profile among all users was roughly equivalent, mainly thanks to the explosive popularity of Internet TV and online video (see Video-On-Demand Now 27% Of Internet Traffic: Study). That means in that primetime window, you and I use the same amount of bandwidth as the overall heaviest users (”bandwidth hogs”) who use their connections 24 hours per day.
“From 7 to 10 p.m., we’re all consumption kings,” Sandvine CEO David Caputo told me. “Bandwidth caps don’t do anything for you.” The implication of this finding is that “the Internet is really becoming like the electrical grid in the sense that it’s only peak that matters,” he added.
One thing is clear: broadband providers will need to invest in more capacity to meet the ever-growing peak demand in the primetime hours, even if they try to shift usage that’s not time-sensitive (i.e. peer-to-peer file sharing) to off-hours.
Then we’re back to the original part of this discussion.
Is it fairer to recover that necessary investment in additional capacity from the heaviest users, who are driving the most demand? Perhaps ISPs will introduce a time-of-day surcharge, so from, say, 7-10 p.m. you’d pay a premium for higher-speed access.
Or, do you think spreading the cost across all subscribers, thereby raising the flat-rate pricing for everyone, is the better option? Note that Comcast did this to an extent when it raised the monthly lease fee for cable modems by $2 (to $5), citing costs associated with its DOCSIS 3.0 buildout.
If, on the other hand, you want to pretend that all-you-can-eat plans are sustainable at today’s price tiers, you’d be kind of clueless.
Blog Type commented:
Todd,
Bottom line: if you give a monopoly a chance to "rewrite" their pricing model, they are going to come out ahead. These are smart people and they have all the data (something notably missing from your article). Why do you think they'll play nice and give all this increased revenue back to users in the form of upgrades? BTW, I have a nice bridge to sell you (it'll definitely increase traffic flow).
Edgy commented:
Todd, you're being disingenuous. You ARE edgy! Don't pretend otherwise. ;)
Todd Spangler commented:
IB -
You mean facts like the assertion that the cost of monthly Internet service will forever be “efficiently recovered through a fixed monthly price”? Where is the evidence to support this statement when Internet bandwidth usage is expected to continue to grow at a 40% compound annual growth rate from 2008 to 2013 (according to Cisco’s VNI study)?
I never said I was “friendly” with ISPs. I’ve spoken to them to get an understanding of their business. I’m not sure some of the commenters have done that.
IB commented:
If you're friendly with the ISPs, why are you writing from the point of view of the unaffiliated? Clearly you're spouting their talking points without really bringing anything new to the table! The "edgy bloggers" at least try to support their statements with facts, not propaganda from ISPs and industry representatives that have a stake in metered billing and network throttling (Sandvine).
It doesn't matter what these other people tell you in the comments section, you're just going to keep coming back with BS responses missing a lot of facts :'(
Todd Spangler commented:
>> under a usage based billing most peoples bill will go down? or up?
Potentially down or up.
Under Time Warner Cable's revised proposal for the usage-based billing tests, there was an entry-level tier for $15 per month - however, with a ridiculously low 1GB cap and max download speed would have been 768 Kbps.
But still, the idea was that people who just wanted e-mail and simple web browsing would pay less.
Jeremy commented:
The idea of metered usage when the users can't totally control it is absurd.
Most people have several programs running on their internet connected device that automatically go out and grab data without asking for permission. Heck, even Windows will download 100+MB updates in the background without telling you first. It's almost predatory to charge consumers per byte when they can't control (currently) each byte that goes in or out.
This makes me truly wonder if the whole net neutrality idea has been a cover for ISPs to distract the public. Chances are there was NEVER any *REAL* intention to handle traffic this way, it was just a way to 'ease the pain' of trying to move to metered usage.
nice hat commented:
where can I buy that baseball cap? looks cool
Todd Spangler commented:
Jane, The capex is really only "fixed" if and when consumption demand levels off. Node splits are relatively cheap but stringing more fiber deeper into the HFC plant is certainly on the horizon.
While progressive-download Web video doesn't need real-time latency/jitter, the video *is* accessed on-demand so the load on the network is the same within that heavy primetime usage period.
Thanks for your thoughtful discussion of this.
lohertz commented:
So, just to be clear, under a usage based billing most peoples bill will go down? or up? Because I would assume that since the peak times indicate when most people use the connection. So lets say I check my email and maybe download a few tunes. Total usage is xGB on average for an average user. So the price is relative to the usage amount. How much is (provider) going to charge for 1GB for data transfer. I would guess right now, average users are less than 5GB/mo. So for 5GB Comcast charge $100/mo for 100/20 MBPS connection. Thats $20/GB!!!!! C'mon. The whole reason its to put more cash in shareholders pockets. Its a way to increase ARPU and show revenue growth in stabilizing market. That's right stabilizing market. There not many new broadband customers coming on-line. And for those making the switch from DSL to FiOS or Cable seem to be the only "new" pieces of revenue, but then again its not really new because the ARPU is higher in DSL than it is in cable of FiOS.
What were talking about is network capacity at peak. And there is nothing to stop everyone in the US from getting on-line at 7PM. No matter how much a person uses, if they're all on at the same time it doesn't matter on the overall usage or pricing.
Jane Addison commented:
Todd,
You raise a good point. Where will the cap ex for upgrades come from, and how is the fair way to raise that cap ex?
First, for cable, the cost of DOCSIS3 upgrades are cheap, with the bulk coming from the new modem cost. Customers who subscribe to a D3 service tier will cover most of that cost. Second, ISPs are free to "price discriminate" by offering higher speed tiers -- those who are heavy users will self-select into the faster Mbps tiers, and will pay more.
In the case of telcos, the upgrade costs are higher. But remember, when Verizon deploys FTTH, they incur a cost to "pass" households, then another to do the drop. These are about $500-$800 each. But VZ has to make this investment in some areas due to their copper being over 60 years old. Also, they are selling triple play services, and recouping their investments with margins on Internet in the 80 percent range, and on TV in the 20-30 percent range. Verizon's FTTH investment is not being driven by bandwidth "hogs", but a need to grow and diversify their business.
But again, from an economic efficiency standpoint, all these upgrade costs (be they cheap or pricey) are fixed costs, which are efficiently recovered through a fixed monthly price.
If I go on vacation for a month, Comcast certainly doesn't stop charging me for Internet, TV and Phone service.
Now, on your point about utilization. Yes, the networks today are run with contention ratios at about 60:1 for residential DSL, and about 100:1 for cable Internet. You are implying that changing user behavior (IP video) is changing the ability for multiplexing to keep these utilization ratios this high. But that is simply wrong. On cable, node splits are common, and dirt cheap, and with D3, the MSOs have tons of headroom. On DSL, the line from the CO to premise isn't shared, and the big telcos like verizon have said publicly that they are not seeing traffic spikes that require COs to be upgraded beyond what is considered routine.
Furthermore, there is a myth that Internet video needs to be real-time streaming. Most of it isn't, it's buffered. On Comcast, I have no problem renting an HD movie on my AppleTV during primetime, because the 720p movie runs at 4.5Mbps, and the 16Mbps connection I have builds plenty of buffer.
In short, the economic reasons you offer for usage-based-overage-billing are just not valid. You are not talking about congestion pricing, all you are defending is greed on the part of companies who operate in a duopoly, and are making high margins already.
Love and Kisses,
Jane
Todd Spangler commented:
Jane,
Right back at you.
Are you aware that broadband access networks are *not* currently built to support anywhere near 100% full utilization? Just like the PSTN. The problem is utilization rates on broadband show no sign of slowing. Where will the investment for that additional capacity (capex & opex) come from? The question is, do you spread that across all subscribers, or should the users that consume the most pay more?
Jane Addison commented:
Todd,
Nice try. But you show a stunning lack of knowledge of network economics, the history of pricing in telecommunications, and business.
First, from an economics standpoint, there are 3 segments to a network's cost -- cap ex, op ex maintenance, and op ex transport. Since you cover this beat, I'm surprised you don't know that the FCC's broadband team reported that in urban areas, and ISP's total cost per customer is $7.58. Of this $7.58, only 50 cents is attributed to transport cost. Cap ex is a fixed cost, and most of their op ex is fixed. There is almost no variable cost. From an efficiency standpoint, the proper way to charge for this network is through a fixed cost. It cost's Comcast the same to deploy and operate a line to your house as it does to your P2P-loving neighbor's house.
The latter is an important point. The cable ISPs that support multichannel news are purposely confusing congestion pricing and usage based billing. You seem confused as well. Under the principle of efficiency, they should allow me to download as much as I want, so long as I do it not during peak hours, as the variable cost of my "hogging" is virtually zero (it certainly is for the Tier 1 ISPs).
I'd say more, but why bother. Your biases and lack of a healthy skepticism in this arena is obvious.
Todd Spangler commented:
DM -
Have you ever run an Internet service provider? I haven't but I know people who do.
Also, I said "edgy" bloggers, not "nasty" bloggers.
DM commented:
“If, on the other hand, you want to pretend that all-you-can-eat plans are sustainable at today’s price tiers, you’d be kind of clueless.”
So, you’re suggesting that I am clueless on the subject of consumption-based billing, even though there are plenty of reasons to point out why consumption-based billing is not necessary for ISPs to continue to profit from their current network operations? How about when these “nasty bloggers” point out that the cable industry uses the exact opposite argument for a-la-carte video distribution systems? The truth of the matter is that these ISPs and video distributors (i.e. cable companies) want to have things both ways. They want to continue bundling channels that consumers do not want to subscribe to and want to shift to a consumption-based billing process for internet services without all of the regulatory fanfare that energy and other utility companies have to deal with.
I think that if consumption-based billing is to be executed, then it needs to be fair and under heavy regulatory scrutiny. If I get a bill that says I used X amount of data but feel that I actually used Y amount, and I have my router logs that verify Y amount was used, then how is that problem resolved? If the Time Warner Cable situation from this past spring is any indication of how things will be launched, consumption-based billing will not actually be consumption-based billing. Instead, it will be a system where you still pay a flat rate for a very low amount of data and then are charged ridiculous amounts for data that goes over that limit. Can you explain again how that form of consumption-based billing is good for the consumer?
Consumption-based billing methods can also be challenged by the fact that current speed tiers themselves act as limits to how much data can travel over a network. If I am on a 5 Mbps connection, then the maximum data that I can use cannot be more than the maximum data of someone who is on a 10 Mbps connection. I feel that ISPs have probably over-promised on speed, not having the capacity to handle faster connections but upgrading speed tiers anyway because of marketing reasons.
The truth is in publicly available financial reports. Costs to maintain networks are decreasing. Meanwhile, revenue from broadband products and services is remaining fairly consistent, leading to great profit margins. Keep in mind that broadband has used the same business model for the last decade and that it has worked and been successful for the last decade.
