So here’s a question for you. I’d like you to have a wee look at this video and see what you think of the concepts demonstrated.
I came across it in the ‘Demo Lab’ area at MACH Insights 2011. It’s showing off a MACH product by the name of the Data Roaming Engine which enables a massively flexible raft of policies and controls allowing operators to get rather inventive with their roaming deals.
The concept is this. You arrive on to a gorgeous desert island and you immediately put your phone into ‘no data roaming’ mode.
However you then get a text message from your operator encouraging you to browse (for free) the possible roaming options. One of the items in the shopping cart is (for example), the ability to browse Facebook free for a whole day for just €2. When you select and confirm, you can then use Facebook ’til your heart is content, for that hour/day/week/period.
There’s a few other options shown in the video.
This is the kind of inventive thing I’d like to see operators play with, especially given the fact that 42% of travellers deliberately do not use roaming (MACH’s figure) and the vast majority of the rest of the users are ultra careful about what they spend (my viewpoint).
Here’s the video:
And another point I’d like to raise is flexibility. If operators get themselves a decent roaming control engine, then they can easily test and experiment with different options without having to pay crazy amounts of money to change their billing systems for every little marketing test.
Ewan
Sorry – but application-specific roaming is even more unworkable than application-specific domestic data plans, especially when the visited operator isn’t part of the same company group. The main problem is that despite rhetoric from companies like MACH or assorted DPI vendors, the network *cannot* reliably detect applications.
The problem is that what the network thinks is an application, is different from what a user thinks is an app. Take Facebook. If I share a web or video link with you on your Facebook wall, and you click on it, it appears *inside* the Facebook app (with the blue banner at the top). It might use its own browser “engine” to display it, or it might use the phone’s native browser to render it. But to you, it’s still “Facebook”, because it gets mashed-up in the phone and presented to you inside the app. But to the network, it looks like ordinary web traffic.
And who knows what will go into future versions of the Facebook app? Maybe they copy Opera Mini and route the compressed web pages through your Facebook page? Maybe they put a video player in the app? Doesn’t matter: to the *user* it’s still Facebook.
So the operator risks “false positives” selling the entire Internet, funnelled through Facebook, for €2 a day, while the user risks “false negatives” by having certain “Facebook” traffic counted & charged separately.
Now add in two operators (visited & domestic) plus possibly a roaming hub provider, all with different network capabilities, DPIs and policies, with complex wholesale / retail structures for mobile data, trying to get this right. Not possible.
(not to mention encrypted traffic, notifications, proxy servers, CDNs and a ton of other gotchas)
Dean
A very comprehensive overview, Dean.
I was giving that some thought last night because I actually interact with Facebook in multiple, multiple ways.
E.g. When I take a photo with instagr.am, I upload it to Facebook. Strictly speaking *they* upload it on my behalf as I transmit the photo to their servers through their app. A little tenuous but I think many consumers wouldn’t differentiate that.
I get SMS messages from Facebook.
I use their app.
I use their mobile website.
Yeah. I see your point.
I shall see if I can find a MACH chap to answer it!
Ewan, thought you would have realised by now – don’t float any hypothesis involving mobile billing, data, traffic, roaming or interconnect without asking Dean first 😉
Hi Dean,
The attitude expressed below is exactly why Ewan’s excellent moderation brought out the fact that of the four people panel none of them had roaming switched on while working in the industry and even working for operators. So the structure is now so complex that noone buys the product and then let’s rather keep it this way.
The – “not possible” – I do not buy that – IOTs are at a point where you can easily do package bundles for data. It is a matter of doing you homework – traditionally operator always take the domestic ideas to roaming – and so in you current market – is there any 1 GB data a month package ?
Maybe the pack is not 1Gb for roaming – but aligned with consumption – and then if you will see the normal distribution of 10% being network hogs and the rest underconsuming … hence no IOT costs incurred … increased profitability.
In case the EU commision even decides to make the purchase of roaming separate to domestic – operators would need these packages anyway.
On the capability side you place the solution on GRX side (Gn) to avoid the complications of various networks. If ever in the neighbourhood will be happy to do a demo of the solution for you – and lets see if the “false positives” destroy your user experience.
And then consider if the app instead of messing up the direction of traffic might assist the application to be more network aware.
Oh very interesting points Rasmus
Rasmus
I have no problems with the idea of volume-based tariffs & packages, either domestic or roaming. I use the £2 a day Vodafone plan myself, which at 25MB makes a fair amount of sense, especially when combined with Onavo to compress/optimise certain data.
Similarly, device-specific, location-specific, time-of-day and so on are all valid parameters. Mechanisms to hook in to WiFi for offload via roaming or other systems also makes a lot of sense.
What I have a problem with is application-specific or content-specific tiering (with the exception of the zero-rating use case that Mike suggests). The nature of applications is changing too fast for the network industry to really keep up – for example, everything changes again with HTML5. Also think what happens with M&A – if an operator had sold an €2/day MSN/Hotmail package on a 2yr contract, and suddenly it was updated to include Skype integration, would they be very happy?
Separately – the “network hog” argument is largely obsolete and woefully simplistic. For many networks it is signalling (both at radio & IP layers) that is the problem, not MB/GB of “tonnage”.
Dean
The network hog was more in the context of the roaming network more than the bandwith consumption – when you calculate the business case you want to have some headroom to make the case profitable.The operators will be banking on not all using their bought capacity to improve the earnings on the offer.
I accept you point on application specific being diffuciult to manage – but again I believe operators need to embrace this evolution because similarly to your Skype example the reverse could be (in case of non-application-regulated) breaking the bank for the operators if you are on an Unlimited Day Pack of data. So in this case the issue is the same even without the issue of M/A. Take iCloud – for roaming it would be disasterous to sync your movie library from the Maldives – so here maybe not a pack that is a positive opt. in but maybe more a warning of what people are about to do.
So i think there are business models to evolve around this topic as i.e. Facebook also functions as advertizing space and i.e. picture uploads from holidays delivers a lot of targeted views for selling holidays (and/or roaming packs) to that destination.
Let’s see – I guess more of these thoughts will arise in case the EU decides to look at separating domestic and international mobile – i.e. you local voice from VF and roaming from EE … then you need your bundling strategy in place.
I mean if you take the Onavo idea – the operator can even use these type of applications to channel the traffic so that it is easier for the network to comprehend … and in Onavo’s case even save consumers money and operators bandwidth.
Lokdeep thank you for taking the time to contribute 🙂