Establishing the mobile television market
Date: Thu, 01/15/2009 - 14:15 Source: Yann Courqueux, Thomson
There is a general feeling that mobile television is potentially an extremely popular concept. In Japan and South Korea it is already a massive success. Yet the rollout in Europe has stalled. What are the issues preventing it becoming a success here?
First, we need to be clear what we are talking about when we discuss mobile television. The first distinction to be made is between unicast and broadcast services.
The 3G mobile system allows you to send video clips on a one to one basis: essentially this is sending video as a multimedia message. This is fine for niche content or personalised services, but not suitable for large audience channels.
The mobile infrastructure cannot support a mass market of 3G video. A single cell can sustain no more than 16 simultaneous mobile television users, even at the poorest video streaming format of 64kb/s, resulting in only 6.25 frames a second – a quarter of the normal frame rate. Neither is it very mobile: the technology for delivering video clips makes for very slow switching between cells means that watching on the move is far from reliable.
Consumers, in turn, find this quality unacceptable. All the data from trial services show that typically a viewer will watch only three minutes of 3G clips a day, whereas those who have access to a true mobile television service watch 50 minutes a day. Asked in a 2007 Tellabs survey if they found the service acceptable, 75% of mobile broadcast users said yes, but for 3G unicast clips the figure was only 18%.
The alternative is to provide a broadcast television service for mobile devices. This uses a separate network to carry the television, freeing the 3G network for high ARPU applications, which will certainly include personalised video content as well as broadband internet access.
The conclusion is that there is an economical threshold above which it is more economical to broadcast and below which it is more economical to unicast. This threshold can be expressed in terms of audience or concurrent rate, and it will vary from one country to the other, but it will always exist.
From this we can see that a hybrid system will always create more value that any single network alone. To take advantage of this added value needs what we might call a bearer-agnostic service platform: one that can deliver broadcast and unicast services using the appropriate platforms. That defines a need for a mobile broadcast network, and here there are two alternative approaches to consider.
The first, and seemingly obvious, option is to use the existing digital television transmission networks, currently being used to carry conventional television broadcasts to our homes from a network of terrestrial transmitters. Handsets could be constructed to decode the existing broadcasts, or mobile services could be added in spare bandwidth using the same transmitter network (hierarchical mode).
Digital television channels are multiplexed together for broadcast. A single transmission channel – which previously would have carried a single analogue television service – has a bandwidth of about 20Mb/s, and typically be used to carry five or six television services.
But there is a fundamental problem in using DVB-T to broadcast to mobile devices: the system was not designed to do it, and there are major technical issues which are largely insurmountable.
DVB-T was designed to give very stable reception on large external roof-top antennae which are not moving, or at best moving relatively slowly (services have been broadcast to city buses, for example). Mobile television, on the contrary, needs to be received by very small antennae built into a device moving randomly inside the building or even below ground level.
Why is it important that the service should reach inside a building? Because that is where mobile television will be consumed. The research shows that 50% of mobile viewing will actually be in the home. Add to that the workplace, transport hubs and shopping malls, and it is clear that good indoor reception is vital. Indeed, it is hard to see how you could even sell a TV-equipped mobile phone if it does not work well in the shop.
Reception of DVB-T inside a building sees an attenuation of about 35dB, or a requirement for 1000 times more power in the transmitter to achieve an acceptable signal. Increasing transmitter power by 1000 times means consuming 1000 times more energy and the use of many repeaters and gap fillers for each of the channels.
The solution lies in transmission standards which have been optimised for reception on hand portable devices. DVB-H, MediaFlo and other standards are designed to meet these constraints.
The downside is that it needs its own transmitters: not necessarily the masts, but certainly transmitter electronics. The fear is that this might make it a much more expensive proposition to roll out. In fact, we have modelled a significant number of cases and have concluded that building a dedicated transmitter network costs less than upgrading an existing network for indoor mobile reception. Without it, the result will be to leave consumers with a poor quality of service, which they will reject.
This brings us to the final decision: how to pay for the capital investment in a new infrastructure. Can a free to air (advertising funded) model be sufficient? Are users willing to pay for mobile access to TV ? The largest deployments of broadcast mobile TV are FTA: Japan, Korea and more recently Italy are following this model.
I would argue that this is the wrong answer, and that the only way to build a popular mobile television service is as a pay TV offering over a dedicated transmission network.
Japan and Korea invested substantial amounts of capital to built their networks. Both are not breaking even despite high service penetration. One reason could be the choice of technology: the former on a technology not originally designed for mobile TV, the other on DMB which is not the most efficient broadcast technology (up to six times more expensive per channel than DVB-H). Another reason may be that advertising funds may not yet be enough to cover network costs.
H3G in Italy seems to have a much better concept. Its pricing strategy has changed and now provides, as part of a premium subscription, a couple of basic services. That is what they call “ree to air”. Additional premium services are charged per channel and/or per view.
The main reason why advertising revenues are less likely to cover the network costs than subscription is the fact that advertisers look for mass markets. The higher the audience, the higher the CPTs.
Advertisers will not be ready to commit large budgets to mobile television until there is a significant audience reach. Our calculation shows that a service penetration of at least 30% would be necessary to cover even incremental network upgrades, which is unlikely given the poor indoor penetration such an incremental upgrade yields.
The only business case that stands up is to charge a subscription for mobile television. This could be a direct subscription for mobile television, or the MNO could include it in a premium bundle of voice and data services. In Italy H3G provides what appears to be a free to air service; suitable receivers are only available with a premium package, so its audience is paying a subscription for mobile television – they just do not realise it. Exactly the same happens in IPTV when basic users access a television service as part of a triple play bundle.
I emphasise that we have carefully looked at all the calculations. With a revenue of around eight euros a month a user and a reach of about 5% of the population, the business case for building a dedicated network remains sound. And it is far easier to win 5% of the audience with a good quality service that to reach 30% with a compromised offering.
Once the service has achieved critical mass an operator can look at more advanced business models, which could include personalised services – integrating 3G clips where appropriate with a broadcast feed – and appropriate targeted advertising. And of course in this piece I have not considered the other capital and operational costs of establishing a popular mobile television service, as these are largely constant however you deliver the content and earn revenue from it.
But there is no doubt that consumers are attracted by the idea of mobile television. Provided you give them the content they want to watch (and not the content they do not want to watch, like intrusive advertising), and give it to them in high quality form on a platform which is stable and reliable wherever they are watching, then there is no reason at all why the critical mass should not be reached quickly. Given the vision to invest up front, there is no reason why European consumers should not enjoy mobile television.