Concern One: Has anyone considered that the usage rates of many (most?) average consumer users is pretty light compared to most ISP’s AUP or Contract? Email and a bit of browsing isn’t going to suck up gig after gig. Current plans assume a fair usage (you can see that most ISP’s have a pretty generous transfer policy, which I’m guessing most people don’t soak up or there would be a lot more pressure to meter & charge for overages). So, gievn margins on broadband today and assumptions about ISP costs, what if that meant that revenue would actually fall for a wide base of consumers and rise (though not nearly in compensating amounts) for power users based on a “competitive-to-cost usage model? Of course I don’t have the data to back this up (but I have a fair idea working in the industry for long enough), but it’s a pretty big what-if. That would also suggest then that the new rates would have to be based on a sustainable business case that would see per-gig rates on consumer access that would average monthly bills to what they currently are or better for standard users, and unless there was a favourable sliding scale, heavy users would be unduly punished for consumption in a market whose prices were essentially artificially inflated to meet revenue targets. I’m not sure I’d want to be a power-user in those sorts of circumstances. Alternatively, with a model like Dan suggests which incorporates a fixed cost plus usage, the real question is how ISP’s will translate their current margin into the fixed cost (well and whether they can even split out their costs effectively enough to generate a vaguely realistic per-gig price), but then you know how this is going to go – just like the System Access Fee, the fixed cost will be the black hole where providers will charge for the bandwidth revenue they’re not making up (or just the margin they want to be making…?). Also, how will economies of scale in bandwidth delivery be realized by the consumer – will I be paying for capital infrastructre the ISP put in three years ago and is trying to pay off at a rate which doesn’t at all recognize the new infrastructure they have added since then which operates at a consideraby more favourable rate? And how will that particular issue stack up in differing geographic regions where, let’s face it, costs aren’t all equal? And finally, the big caveat: most metered industries are *government regulated*, with the government approving rate scales, and operaters/distributors having to apply to the appropriate government body to charge rates to customers. What does that look like in the bandwidth world – do you trust the CRTC to stay on top of this? To audit all those ISP’s out there? To standardize costs and rates across Canadian markets where it may or may not make sense? That’s a tough one, I suspect.

Concern Two:

If we assume (worst case) today that most consumers under-utilize the potential limits/AUP of their broadband contract, and that ISP’s will have to stretch to maintain revenue in a usage-based model or incorporate fixed charges (real or imagined), I don’t see much (/enough) additional incentive for them to sink a lot of capital into infrastructure – if they’re not already – and I don’t see them making enough incremental revenue right off the line to be able to just divert directly into infrastructure. More of what we’re seeing is providers developing infrastructure to meet higher bandwidth plans (and subsequent demand), which does create a predictable revenue and cost model, as opposed to having to base infrastructure build-outs on the relatively unpredictable sale of metered bandwidth (though if the power companies can sort out a formula, I’m sure it’s only a matter of time before ISP’s can).

Of course it’s Monday, so I could be completely off on all of this.
Maybe we should just go back to charging by hours of usage, like in the dial-up days? ;)

In many respects I’m happier with my non-metered connection right now than I would be with something obscenely fast that I’m going to pay more for and probably not utilize to its potential (not for the cost anyhow). It’s currently fast enough for me to do all the things I need to do personally. Granted it’s a bit slow in terms of serving content (I have dedicated servers, and pay an extra monthly fee for address space, but not for bandwidth) to users, but since I have negligable revenue back on that content I’m putting out, I’m much happier not having to worry about people driving up my monthly bill. Because of course if you up my bandwidth, and that of people whom I’m serving content to, and charging all parties for the increase, the only real winner is the ISP. And frankly, they seem to be in decent enough health at the moment.

I’m not even going to touch wireless – but I agree with Karthik, given the state of radio networks today, and of the serving technology, metered rates and the resulting capital upgrades they fund make sense – but of course save Rogers’ (ahem) “unlimited” plan, that seems to be the case today. And a distinct lack of competition or innovation driving down rates/costs likely seems to make this the status quo for the time being.

Now, once we see some LTE deployments or another player or two through the spectrum auctions, this wireless data discussion should get a little more interesting…

Charge a fixed cost to cover fixed cost * 1.xx reasonable markup + a per-gig fee * 1.xx reasonable markup

That’s how electricity works, and hyper-competitive industries that would get crushed if their prices didn’t agree with their actual costs (express mail service, car operation, taxicabs, Costco/Price Club, web hosting, virtually anything that can be sold in quantity).

Paying per gigabyte enables the operator to finance further enhancements to the network and invest in overcoming the limitations – or atleast that is the hope.