I was just perusing the call charges for international calling and roaming on Orange. Here’s a small and abbreviated extract:
| Ireland | Western Europe | |
| Calling from the UK | 15p | 20p |
| Receiving calls abroad | 20p | 30p |
| Calling from abroad | 40p | 70p |
At first sight this looks like the usual telco price-gouging and service-based price discrimination between users. This works by driving business people who are price-sensitive to international call charges to business plans (with higher up-front MRC commitments); likewise, it discourages their paying these casual use charges for occasional international roaming. Your mum on holiday gets a briefly robbed by the phone companies, but not enough to do anything about it.
Yet there is something deeper going on, even in these circuit-switched price plans. Whilst you’re forced to buy service and connectivity together with circuit voice, that doesn’t mean we can’t reason about them separately. After all, there are already distinct components of the architecture that reflect these economic functions: the provisioning system, and the call detail record generator.
The network’s provisioning system detects whether you are entitled to access the network, authenticates your device, and determines on what overall basis you will be charged. The call detail records track how much service you have used, and the service addresses you have accessed. The billing system blends the connectivity and service portions together to determine the actual money charges.
Orange, along with the other carriers, is allocating all their variable charges into the “service” bucket. The monthly plan fee is fixed; they don’t charge you extra to provision (and remain attached). Yet by attaching to a “foreign” chunk of network connectivity, your call charges are doubled or tripled. But an international phone call isn’t a different product — it’s the same product as a local one. You dial a string of digits and talk either way. It happens to run on a differently provisioned piece of connectivity. This difference between service and connectivity is subtle but important and is obscured by traditional voice pricing.
With a circuit-based network, you’re not really provisioned to have any connectivity at all until you dial. The dial tone is an illusion of connectivity, a psychological fraud. (Why is there no such thing as the “Internet dialtone”? Because you’re already connected to everyone, everywhere, always, without asking!) All circuit telephony connectivity is rented on a temporary basis. You dial, and you’re allocated a single point-to-point conduit. You end the call and the connection dies. Once you understand the hidden connectivity element bundled in traditional voice, the dynamics of telecom become a lot clearer.
As an aside, the prices for incoming roaming calls are lower than outgoing ones in the table above. Why? Probably a whole bunch of reasons. Maybe cultural/historical (like the difference between caller pays in the US and callee pays in Europe). Maybe the varying existence of substitute products. You can call forward a home landline for incoming calls, which isn’t a big deal — or just pick up landline voicemail. For outgoing calls you can buy a pre-paid cell phone or go to a call box, both of which could be expensive or inconvenient. Hence incoming is cheaper than outgoing.
So what happens when user-controlled VoIP comes along and eliminates service-based charging? Let’s project ourselves forward a few years. Some mobile IP-friendly technology like Flarion or WiMax has become pervasive. Open WiFi access points and community networks remain too sparse to rely on. So you’re still buying connectivity from an Evil Telco™. Thus the price discrimination simply switches from service to connectivity, no different to GPRS roaming extortion today.
What this also tells us is that the mobile network operators have a fundamentally different economic structure than fixed network operators. You can use the location of the connectivity as a variable in your pricing. Fixed operators can’t do that so effectively. Remember, the paradox of the best network only refers to the loss of ability to price discriminate based on service. The stupid network is service-agnostic. Yet “service” isn’t the only axis you can map to. To what extent can location-based price discrimination substitute for service-based pricing? And if they are substitutable, it makes you wonder: how truly quizzical is the paradox for mobile network operators?
With any access network, you can price discriminate based on two things: the service type, and the connectivity characteristics. The connectivity could be described in terms of its destination, time, volume and bandwidth. Assume service-based pricing dies with IP networking; it’s too easy to bypass. Score zero out of ten for benefit to price-discriminating carrier.
The asymmetry element of incoming/outgoing connections probably dies too with it. I’m not expecting any carrier’s marketing department to get away with charging by the SYN packet. This is especially true given the troubles of educating the public and gaining acceptance for the much simpler concept of metered megabyte-based pricing. What else does that leave us with?
Destination-based pricing is too hard to sell to the masses, who don’t want to have to know which continent each IP address and domain name get routed to. (Heavens, if you though micropayments for content are bad, imagine if they existed for IP connectivity!) Score one out of ten. Time-based pricing was receding along with dial-up, but is re-appearing with paid WiFi and internet cafes, where you pay by the hour. Maybe give it three out of ten, with time-metered broadband a tragic reality for my parents-in-law. Volume pricing is highly prevalent, with different service caps at different price levels. A similar story applies to speed. Once IPTV becomes popular, the idea of megabit DSL and gigabyte caps will be laughably stingy. Someone will get rich offering connectivity that reflects the true value of video content. This is much the same as today’s broadband revenues that rest on the value of, ahem, music content. Score eight out of ten; you lose the ability to do fine-grained pricing, but you can tier the data-hungry applications.
Thus fixed networks retain two strong forms of price discrimination (volume and bandwidth), whereas mobile networks have three by adding location. Furthermore, these factors are likely to be more variable over time on mobile networks. That tips users towards fixed all-you-can-eat bundles that potentially cost more than buying components separately when there is high variability in usage and difficulty in comparing prices.
So the summary? Whilst the fixed-line telcos remain a terrible repository for your savings, there is a little more hope for mobile networks because there are more possibilities for effective price discrimination. Furthermore, the disappearance of service-based pricing doesn’t automatically mean the end of attempts at price discrimination. Someone might lay a fibre to your home, but that doesn’t mean they’ll offer your gigabit nirvana for without consideration to the value you get from it.
If you don’t believe me, just check this one out:
One of the most innovative players in the broadband business, BellSouth is testing the on-demand concept in up to 750 homes in southeast Florida. As part of the scheduled three-month trial, which started in the Miami/Fort Lauderdale area in August, the regional Bell is offering customers an always-on, 56 kilobits per second (kbps) DSL connection with the ability to upgrade to as high as 3 Megabits per second (Mbps) speed at a moment’s notice.
On-demand subscribers need only click on a “Turbo Button” icon in the corner of their screen to make use of this special “Turbozone” feature. Subscribers also enjoy the higher speeds when BellSouth or a content partner wants to provide faster downloads from its Web site. Similar to the phone company’s dial-up service, the on-demand service costs trial customers anywhere from $5 to $15 a month, depending upon their bundle.
I couldn’t make this sort of stuff up. I just look forward to the 56kbit fiber optic system. And you thought telcos weren’t innovative?
Posted by Martin Geddes at 11:22 PMTrackBack URL for this entry:
http://www.telepocalypse.net/cgi-sys/cgiwrap/mgeddes/MT/mt-tb.cgi/304.
When a mobile phone subscriber roams abroad, the largest portion of the call price (either for incoming or outgoing calls) is made up of interconnect fees. These are the payments made to the foreign networks and international carriers by the subscriber's home network.
Using your example of an Orange customer: If the Orange customer is making and receiving calls, say, on the Vodafone network in Hungary, then Orange have to pay Vodafone Hungary (and any other telcos involved in helping to route the call between the UK and Hungary) to use their network. And from Vodafone Hungary's perspective, they have no motivation to be terribly competitive on price, as the person making the calls isn't their customer, so they charge as much as they think they can get away with.
When an international call translates into the VoIP arena, you immediately cut out a whole load of middle-men. You certainly don't need to pay an international carrier. You may not even have to use a national carrier. If you can keep a call on the internet, all the way up to the called-party's local exchange, all you'll have to pay for is the equivalent of a local call - a price that, these days, is pretty-much negligable. Hence why VoIP telecoms services manage to deliver calls almost anywhere in the world for the price of only a couple of pence.
It would be nice to think that a Mobile roaming calls might be able to be routed over the internet as well, so that roaming subscribers can achieve the same cost savings. However, GSM roaming regulations don't readily allow for that kind of thing.
Posted by: at November 16, 2004 04:45 PM