In this issue:

The Kerton Group presents
Kerton Call
9 June 2008


The Difference Between Unlimited Voice & Unlimited Data

While the trend in fixed and cellular voice markets is towards flat rates, the trend in broadband is the opposite

When mobile phone services were first launched, customers were billed on a per-minute basis, where those who called more paid more. Around the turn of the century, so-called “buckets” of minutes appeared that would allow use ‘up to’ a certain number of minutes for a set fee. Over time, these buckets have become more and more generous. Marketing schemes like “Rollover minutes”, In-network free calling, “Fave five” plans and such have made the effective price per minute even cheaper. The culmination of this trend is the “All in” plans that the mobile carriers announced this spring, with unlimited voice calling for $100 a month.

On the fixed telephone side, North America has always enjoyed unlimited local calling, but our long distance (LD) calls have typically been billed by the minute. We all know that prices have steadily dropped since the AT&T break-up, and now there is ample competition for unlimited local and LD calling, with VoIP providers offering it for as little as $20/mo, and the incumbents for slightly more.

So the trend in voice is clear. Once the network infrastructure is in place, and voice signals are digitized, the marginal cost of flinging bits of voice calls has gone to near zero. As economics suggests, the price in a competitive market will follow the marginal cost. Thus, price per minute has gone to zero.

On the data side, originally we used dial-up modems, and paid a per minute charge to connect to the Internet. AOL’s famous mailer disks boasted “20-hours of free access!” but that 20 hours steadily increased over time. Finally, unlimited dial-up services rocked the metered world and became the standard, seemingly following the same trend as voice.

When broadband DSL and cable finally arrived, they were marketed from the start as unlimited, in order to compete directly with the unlimited offer of dial-up service. But now, there is talk from the Comcast camp about offering capped services, and possibly multiple tiers of service with different throughput allowances per month. Specifically, Comcast is publicly floating the idea of a 250GB cap, with $15 for each additional 10GB.

So why this seemingly regressive proposal? Do throughput caps for data services make sense?

I say they do.

The reality is that the only reason voice has become free is that it was digitized, compressed, and sent as IP packets. Thus it no longer ties up a dedicated pair of copper wires for each conversation, and now only uses a negligible portion of fixed resources for each call. But voice traffic also has a natural upper limit: There is only so much traffic one can send over a voice circuit in a given month. Let’s calculate what that is:


  • CDMA uses EVRC codec at < 9.6Kbps
  • GSM uses GSM/AMR at < 12.2Kbps
  • Fixed phones use G723.1 at < 6.3Kbps
  • So let’s say that the average bitrate of a compressed call is about 8Kbps.

Voice monthly throughput

  • There are 2,592,000 seconds in a month (60 x 60 x 24 x 30)
  • The most a phone call will consume, if left active for the entire month, is:
  • 20,736,000Kilobits = 2.47GB (z / 8 / 1024 / 1024)

So, a voice call, left connected for an entire month non-stop, would only create network traffic of 2.47GB. Thus, even “unlimited voice” is capped! And that is a figure that neglects a good network trick of not transmitting silent moments. From a network operator’s perspective, fixed or 3G mobile, that is simply NOT a huge amount of monthly traffic. With 2.47GB as the upper limit, it’s easy to see how a company like Verizon Wireless can offer an unlimited voice package for $100/mo. Especially when we contrast this to Verizon’s EV-DO data package, which costs $60/mo and is capped at 5GB. Hmmm, let’s see, the data package costs $12/GB while the voice package costs $40/GB.

If anyone is complaining that data caps are a bad deal, they’re off target. It’s the voice that is the relatively bad deal (this is a simplification, because handling voice has other costs, too: termination fees, 911 costs, number allocation fees, etc.)

So why are data throughput caps needed? Because, without them, the risk of bandwidth hogging is massive. It is almost inevitable that a few bad apples will suck up the bandwidth. With voice, there is little risk that any one user can have a significant impact on the network capacity. The most they can throw around each month is 2.47GB! But with data networks, and just a 3G wireless network speed of 600Kbps, data users could easily consume 1.5 Terabits of traffic per month. That’s a lot of data, and as networks speeds increase, so does the maximum consumable. None of us want to pay the rent required to build a network that could support that much traffic from each user, and it’s not fair to expect the moderate users to support those that hog bandwidth. Thus, throughput caps on individual users help the telco put a limit on this liability – a cap that exists naturally with voice.

In my experience as an officer of a global association of wireless broadband ISPs, I’ve seen the charts showing how the top 5% of bandwidth users can suck up 80% of the network’s capacity. It seems to me fair and logical that those users should be offered the ability to continue doing so…so long as they are willing to shoulder the cost of upgrading system capacity.

My colleague and friend Mike Masnick argues that bandwidth caps are a bad thing, because putting a limit on what people can do serves to suppress innovation. He correctly notes the boom of great new Internet apps and services that emerged when dial-up rates went unlimited. A totally unlimited high-speed Internet will promote future innovations. While I agree with Mike on that, I simply do not agree that the ISPs have an obligation to support and fund future innovations of the greater web community.  Why should they pay for the real estate on which others might build? I hate when incumbents impede innovation, but neither is it their responsibility to make it happen.

Yes, bandwidth caps will impede Internet innovation. It will do so in the same way a price for oranges impedes innovation in citrus concoctions. But a price for oranges also reduces the wasteful use of oranges. In fact, wasteful use is an equal or greater threat to innovation as caps. You see, if bandwidth is free (or flat rate), then wasteful users will soak up a great deal of capacity, and innovation will be squeezed into the margin. But when users are required to pay for throughput, economics suggests that a market allocation will occur, where productive innovation will be more willing to spend on the available bandwidth than non-productive waste, and a proper resource allocation occurs.

Optimally, ISPs should sell their services with clear and transparent pricing. They should offer a few tiers of service that meet the needs of target demographics. For the consumer category, this might mean a tier for light users, a tier for the mass market of consumers, and a tier for heavy users. Users should be notified mid-month if they are over-consuming, and they should be alerted if they are approaching their caps. Once at the cap, the user should be scaled back in speed, and offered an up-sell for additional throughput. Fair, simple, informed, transparent. This is the so-called smart-pipe.

by Derek Kerton
Principal Analyst, The Kerton Group

Stage Notes

For those of you interested in hearing about the vision of a broadband future from the CTO of KT and the CTO of NTT, we will be moderating the Telecom Council's KT-NTT Venture Forum Kickoff on July 1 and may be able to swing some extra seats for you.

About Us

The Kerton Group provides insightful, expert consulting in wireless telecommunications to companies across the entire value chain and around the world. Our analysts' deep, vertical knowledge of the wireless industry delivers a complete understanding of the technology, market, and business issues our clients face today and in the future. The Kerton Group's expertise is internationally recognized in widely distributed industry reports, articles, newsletters, and frequent quotes in national press and industry publications. Our services include strategy, business development, research, marketing, product development, due diligence and technology transfer services. We also run the SPIF, a private investment forum for telecom investors. Our projects, clients and publications are detailed online:

This Newsletter

We encourage you to share, use, and reproduce this newsletter freely if you cite your source as "The Kerton Group." The Kerton Group publishes The Kerton Call, a newsletter that discusses topics affecting the wireless and telecom industries that the trade press isn't covering.

To subscribe or unsubscribe to The Kerton Call, reply to this email.