There was a question on LinkedIn yesterday: "Is the Broadband ISP world moving away from all-you-can-eat pricing?" I went to answer, then LinkedIn crashed, and now the question is closed. Never mind, here's my answer anyway.
So first of all, there's a number of misconceptions about current broadband usage and issues that need clarifying before going further:
Misconception number one: why is there a problem?
First of all, flat rate broadband tariffs which are currently the norm in many Western countries emerged in order to drive penetration and acquisition. They were designed as an acquisition tool, to reassure users that they would know exactly how much broadband would cost them at the end of the month no matter how they used it. Since there is theoretically no per usage costs in an IP network architecture (although see caveat below), this was perceived as safe both for the users and the ISPs.
So the issue ISPs currently have is not with flat-rate tariffs per se, it's with what flat-rate tariff was supposed to get you. And what it was supposed to get you was unlimited acces at an advertised bandwidth for you to use as you saw fit. Except that in order to cut costs, ISPs used a couple of slightly devious techniques in the early days that are coming back to haunt them now. Namely that
- they advertised theoretical maximum speeds for DSL based on equipment capacity, but certainly not on real capacity and/or customer specifics (like distance from the MDF)
- they used contention ratios for backhaul and off-net traffic of upwards of 50:1 which meant that if in the theoretical instance that all broadband users would try to access the internet at the same time, they would all only get 1/50th of the advertised bandwidth at best.
Of course, as long as customers were clueless and had simple needs, this was easy to keep up. But strangely enough, users ended up doing what ISPs had been enticing them to do for years: actually use their available bandwidth (and not always legally, but strangely enough that didn't seem to bother ISPs when their pipes were still empty and they were acquiring customers by the bucketload with the promise of free music and movies...)
Misconception number two: who is the problem?
There's this internet meme going around that a skewed Pareto law is the reason for all the ISPs woes, namely that 5% of the users account for 80% of the bandwidth hog (or 10/80, or 5/60...) I haven't read many studies that looked at that phenomenon in detail, but the one I have studied (by Cho in Japan) showed conclusively that while this meme was indeed true, it was also misinterpreted: it is true that a small percentage of users at any given time account for a large amount of usage, but these users are not identical at two different points in time. What this means, is that while the Comcasts of this world may believe that punishing the heavy hitters will solve the problem, what it will do instead is alienate a large proportion of their customer base who at one time or another are heavy hitters.
In other words, everyone is the problem.
Misconception number three: why is this an issue in an IP world ?
The answer to that of course is that the customer-facing world may be IP, but the wholesale world that undelies it is not necessarily. What happens when customers suddenly start asking for the bandwidth they think they have purchased, is that if you want to serve them, you need more bandwidth at the aggregation level, and more bandwidth off-net (or you accept to degrade quality for all, with likely churn as a consequence). The latter is usually a non-issue - at least between Europe, Asia and the US where it's cheap and widely available. The issue for ISPs is with aggregation, also known as backhaul. Now if an ISP controls its aggregation network (either owning the dark fiber or renting it) that connects its DSLAMs, providing more bandwidth may be costly and slow (it means equipment upgrades, mostly) but it's an investment and it's feasible.
The real issue is if your backhaul is a wholesale offer from another player (usually an incumbent) especially if that offer is metered and/or if you pay by the Gb/s. In the latter case, if you want to improve customer experience you have to pay more. In the former, you're doubly screwed because not only do you pay as your users use, but you can't prevent them from using (collectively) up to the cap that these offers usually feature. In other words, you're selling a flat-rate retail offer over a metered wholesale offer, which frankly was a bad idea in the first place. That's kind of what's happening in the UK right now.
So that being said, how is broadband pricing going to evolve.
First of all, metered for broadband is dead as a dodo. People simply do not want the uncertainty of metered, and there's been too many horror stories in the news about people faced with thousands of dollars of bills because of absurd metered schemes (there's one right now about an £11.000 bill for downloading four episodes of Friends on a mobile...) If they have an alternative, they will use it. And that's true of the semi-flat capped offers too: it will only take one competitor to go fully flat to shift the market in an instant, and I suspect the only reason some countries like Australia have all capped offers is because the wholesale offer that they are all using for backhaul is itself metered.
However, and that's my second point, that does not mean that there's no way to make heavy users pay more than light users. First of all, things are different now than they used to be. Heavy users who would have happily reverted to being light users a few years ago would no longer be willing to do so. So there's demand. Furthermore, you can offer more services alongside more bandwidth and/or upload-download capacity for these users to sweeten the deal. The key is not to make it variable. It has to be a fixed rate. My suggestion would be tiered pricing, with each higher tier opening up other options, like the possibility of lower latency (for gamers), more guaranteed QoS, or various other "technical" options that would appeal to more savvy users. Say you keep it simple and have:
- for 40 EUR/month your bog-standard unlimited access with a high cap on download or upload and the option to either lower bandwidth considerable (say down to 512k) if you hit the cap or pay one shot to go above
- for 50 EUR/month you have an unlimited access with a higher cap and a number of technical options, tweakings that the user can access on his service, partner offers, whayever.
- for 60 EUR/month you have the premium "you're a mean hacker" offer that is truly unlimited with all options accessible, etc.
The thing is the differentiation is not on bandwidth, it's on caps and services. I'm pretty sure such a model would work provided you have a disruption on the market that allows you to change the way things are done (you don't want to be the one to make heavy users on your network suddenly pay more unless you provide them with extremely hot additional services...) FttH could be such a disruption though... just saying.
The third and last point is slightly more complex to implement from a marketing point of view but it's an interesting one. And to give Caesar his due, this came up in an exchange with Brough Turner in January in Boston, and I'm only expressing here something he suggested to me. If P2P is the culprit, as it seems to be, there's one thing you can count on it's that any pricing scheme you implement that penalises certain uses of P2P will be taken into account in the next releases of whatever P2P client is hot that month. In other words, whether they know it or not, your customers will adapt to your pricing schemes.
Therefore, if you reflect some of your costs in your prices, customers who use these clients will try and avoid the "high-cost" areas, which is in your interest. Pricing scheme as a deterrent rather than an additional revenue generator. What Brough suggested was a pricing scheme that was unlimited for on-net and metered for off-net traffic.
Say you download Piaf to see what the fuss about Marion Cotillard is all about using Bit Torrent. With current pricing schemes, your Bit Torrent client will indiscriminately find users on or off-net to download bits of it. If an ISP launched a differentiated pricing scheme where on and off-net were treated differently (and off-net cost you more, one way or another) you can be pretty sure that a Bit Torrent client that allowed you to favour on-net as opposed to off-net would see the light a day in a matter of weeks.
The impact for the ISP would be immediately more on-net traffic (that costs nothing if his network architecture is sound) and less off-net which costs money.
And now you're going to say "but you said metered was dead". You're right. That's slightly paradoxical. I'm not saying it should necessarily be an on-net=flat and off-net=metered, but I am saying they should be differentiated, and off-net should be more penalising to the user if you want the desired effect. And I do acknowledge that it raises marketing issues, as in "how the hell do you explain that to Mrs. Average Customer".
I'll let you sort that one out. It's late here and I've rambled on for too long. And as for fiber pricing, it's a more complex issue that latches onto this one, but I'll keep that for another day!
Before I jump in bed though, when I broached this last pricing idea to wonderfully-hatted-diamond-subscriber Magnus last week, he said that he'd heard of broadband offers in Iceland priced using similar principals, mainly because from Iceland, any traffic to elsewhere is bloody expensive. If any of my readers can point me to any details on said Icelandic offers (hopefully in a language not too reminiscent of XIIth century sagas) they will instantly gain the status of platinum-subscribers with all associated privileges!
