Quite a lot of the discussion at last week's World Handset Forum was around Mobile TV... DVB-H, DAB/DMB, Mediaflo and all the rest. I also dropped in on the UMTS-TDD Alliance meeting which was occurring (by coincidence) across the other side of Prague, where they were discussing their TDtv technology.
I still like the underlying proposition - "It's TV, on your mobile" . Nice, easy to communicate to customers, and apparently with good results in customer acceptance during trials.
But there are a few issues.
I don't have time to write about all the issues about spectrum here, but clearly that's a big problem.
Similarly, it looks like getting good indoor coverage may also be an issue - at low frequencies (sub 1GHz), it may be do-able with a 50% premium in infrastructure spend vs. "outdoor + best-efforts indoor" coverage. But at higher frequencies like 3G-band 2.1GHz, this is likely to be much worse.
There also hasn't been much discussion about the handsets. Looking at some early devices, it seems that typically they have to have 30-70% larger batteries than ordinary 3G phones (which lets face it aren't the most power-efficient in the first place). Various commentators have said that they can get 2-4 hours viewing time, although this presumably excludes people using the devices for other 3G stuff like music downloads or serious gaming as well.
There's also the old favourite problem of user interface design, and integration with all the other bits of software on the phone. Unless we have a "velcro-type" (trademark acknowledged...) solution, it seems likely that we'll have to have a huge amount of integration work knitting in the various bits & pieces. Want interactive viewer-response? OK, customise the SMS client to behave in the right way. Want to have a Tivo-like experience? Get some more software (& memory). Want the TV to pause when you get a call coming in? Or maybe picture-in-picture when you get an MMS? Get writing some custom code....
Bottom line is that I reckon a handset may cost a lot more than just the additional hardware bits & pieces. Obviously it will depend on volumes, but I reckon that by the time you've done additional software work, added in a TV chipset / antenna / better screen / more memory / clever hinge etc, done some "useability" consulting and so on, it may add $50-120 to the price of a handset.
Now, supposedly people are prepared to pay a monthly fee of $10 / €10 for mobile TV. Some trials bear this out, although a speaker from Finland (ie pretty affluent mobile-centric country) thought €5-8 was more likely. I'd be surprised if massmarket punters would be willing to pay much more than €5 (ie roughly what they pay now for SMS per month). €60-100 per year, maybe. And unless you can persuade them to pay more for a device (€50 was mentioned as being acceptable), operators will need to subsidise it - so probably an extra €10-€50 per sub per year, given a 18/24 month upgrade cycle. Then there's the capex/depreciation cost of the network, marketing & sales & support, opex for the backhaul network, revenue-share with content owners or broadcasters etc.
Bottom line: I'm getting sceptical about the business model ever working, even at €10 / $10 per month. Unless there's a hefty wedge coming from advertising - maybe $30-50 per sub per year. (I've got no idea what the comparative figure is for TV advertising although that would need to be multiplied up by average # of viewers per household subscription anyway)
The other possibility is getting additional voice / SMS calls from people talking about TV programmes they're watching, SMS'ing in for interactive votes etc. Which is all very well, but has to be weighed against the risk of people making fewer calls or SMSs or music/game downloads, because they're busy watching TV.