In two previous posts (here and here), we discussed bandwidth cost estimates, especially in the form of data caps, and their overall relation to the CRTC’s mandate to protect the public interest. The story continues, with a few more plot twists…
This is nominally the third segment in what may turn out to be four, though new gems keep popping out of the woodwork. (Btw, I’ve decided to go with “data cap” rather than “bit cap” since it’s clearer to most people and more accurate). My main aim is to show further evidence of the bankruptcy of the CRTC’s economic ITMP strategy, through an analysis of Rogers’ Internet rate card, along with data from the OECD and some press commentary.
Let’s start with my favorite linguistic hobby horse. One of the greatest obstacles to consumer protection is bafflegab, be it the product of sinister plots to pull the wool over our eyes or just plain stupidity. If you have no clue what you’re buying, then you’re buying a pig in a poke and you’re an ideal customer for an incumbent service provider, which will ARPU you till all the farm animals come home. In a survey it fielded earlier this year, the FCC found 80% of its respondents were unable to say how fast the broadband connection is in their own homes.
Here’s an item from the current Rogers Internet FAQ that illustrates my point about helping customers understand, or not, how much they’re paying and why (accessed Oct 7). Rogers charges for additional bandwidth usage on all six of its access tiers. In the screen grab below, the subscriber is offered an explanation of how these extra charges are calculated:
I’m not making this up.
In my 4th-year seminar this week, I put the first part of the answer up on the screen, leaving off the last 3 lines. I asked my 25 students to tackle the calculation and confirm the amount to be charged. After 7 or 8 minutes, no one had the answer. One complaint was having to do long division without a calculator; mostly the problem had to do with the perennial confusion over the Greek prefixes used in computer science and engineering – mega-, giga-, tera-, etc. Especially since the Rogers FAQ has its own idea of how to apply these terms.
Did you notice that to convert megs to gigs, the FAQ tells you to divide by 1,024, not 1,000? I wonder if Rogers gets many customers phoning in to ask why they’re using the binary notation (multiples of 1024, i.e. 2^10) instead of decimal notation (multiples of 1000, i.e. 10^3). The engineering dept doesn’t know the IEEE, ISO and other standards-setting bodies have expressly deprecated the binary interpretation of “megabyte,” “gigabyte,” etc? The binary usage is cool if, say, you’re designing memory chips. But not if you’re measuring throughput in a network.
O.K., boys, what is it – 1,000 megs, 1,024 megs or some approximation? And why is this woman smiling? It’s not her online behavior that’s adding up, it’s her ISP’s prices and data caps. ISP service tier configurations with caps don’t reflect onliner behavior, quite the contrary. Canadian onliners have been retro-fitted into a very uncomfortable Procrustean bed (more on this below).
Things could be worse, and they are
The civilized world has rejected data caps. Let’s review the bidding.
In its 2010 broadband research, the OECD reports Canada and Australia are the only two member countries of the 30 involved (members now total 33) that have caps on 100% of the broadband services it surveyed. More than half (17) of the member countries were found to be 100% data-cap free as of October 2009 (see OECD worksheet 5j).
Of the 13 countries found to have caps on the services observed, Australia and Canada had by far and away the highest price per megabyte for usage over the cap. As indicated by the OECD chart reproduced below (from worksheet 4g), Australia is the undisputed leader (at 0.036 USD PPP), with Canada solidly in second place at 0.019 – both way ahead of the average cost for the 13 countries of 0.007. In this same chart, there seems to be some good news, namely that Canadian ISPs have the largest average caps – 76.7 GB, as indicated by the first vertical bar on the left (which also shows the red line with our 2nd-only-to-Australia data cap average price).
Average size, price of caps in 13 OECD countries (source OECD)
But large caps, indeed cap sizes in general, are meaningless without reference to the particular speed tier they’ve been put in place to protect. (My misgivings about caps haven’t changed much since 2002, when I wrote a Web memo entitled “How to make friends with a bandwidth hog,” which I reposted here last May).
The ISP strategy has been to blame their price-gouging activities (known officially in Canada as “economic ITMPs”) on the bandwidth hogs, who share a compulsion for heavy use with their close cousins the copyright pirates (wink, wink). One big problem with the segmentation of onliners into heavy users (bad!) and regular users (okay!) is we’re all becoming hogs thanks to video (now including Internet video, video-to-TV, VOD, P2P, video communications, telepresence, ambient video, 3D, HD, etc). Cisco estimates North American bandwidth usage on the consumer Internet (all applications) is growing at a CAGR of 33%. And as if by some self-fulfilling prophesy, this week Cisco introduced a consumer telepresence product known as Cisco ūmi. If you want the 1080p-grade reception, you’ll need a broadband connection of 3.5 Mbps… but both up and down. So people are going to use that uplink for something.
On average, if you're consuming e.g. 60 GB per month right now, you'll be consuming 80 GB per month a year from now. That means if you're on Rogers' 10-meg tier (Express), whose overage cost is $2/GB and whose cap is 60 GB, you'll suddenly start paying another $40 a month. And since the tier already costs you $46.99/month, your cost will be nearly doubling.
Welcome to one of the world’s most retrograde Internet policy frameworks.
This math doesn’t make you a bandwidth hog, though it may make you feel like one. It means you’re just another average user who will get punished by Rogers and the CRTC for nothing more than organic growth in your usage. That, however, could be the least of your worries. as we’ll see from this rate card in our next instalment (I tacked on the figures in the right-hand column to show how much each Rogers tier costs in CDN$ per megabit per second).
Oops, got a little carried away, we have some burn rate data to cover first…