Broadcast versus on-demand
I believe that on-demand will replace broadcast when >95% of the pop has broadband & there is a good “feed me TV” mode for on-demand
Now, I don’t technically disagree with this statement, but there are a few hidden conditions which means it’s a lot further off than availability of the raw technology belies.
First, there’s not just the general provision of broadband, but the means by which it’s architected. As Kate alluded to in an earlier tweet, we need smart “routers” which can cache content; effectively, CDN nodes within very very easy reach of everyone. Technically, this isn’t hugely difficult, and there are lots of ways by which content can be fed to them (which is where effective use of multicast can help, too), but in many current ISP’s networks, there actually isn’t somewhere that you can place such a node which won’t cost you just as much for the consumer to access as it would were they to stream the content directly from source. Of course, BT will be happy to sell you a solution which avoids this (which again, in part is built on multicast).
It’s not all doom and gloom, of course. Some ISPs could do this today, and it’s not hugely different from how cable TV works. But actually deploying this on a useful scale requires having enough demand from both on-demand providers and ISPs to make the cost of the nodes affordable, which relies on the ISP’s networks to be amenable to their placement, and so on, and so forth. This is not a “it won’t happen”, but it probably won’t happen within the next five years without some serious game-changer appearing in the domestic ISP market.
Next, in order to “feed” content in place of editorially-chosen programming, you need to have rich enough metadata — not just for new programmes, which only make up a comparatively small proportion of what is broadcast and people watch, but also for all the things we’re making and broadcasting today, all the way back to those episodes of Fawlty Towers and classic Star Trek.
This is a huge challenge, much bigger than trying to get intelligent lightweight content-delivery nodes into every ISP in the land. The technical aspects of matching preferences to metadata are fairly straightforward, but to deploy this as a replacement for, say, BBC One (with access to the same content that the controller of BBC One has) would require a monstrous amount of effort. But, on the plus side, it’s effort which is being undertaken.
We’ve already seen Joost try — and fail — to do precisely this, and they failed for a few reasons (chief amongst them being availability of content that people want to watch). Getting content that people want to watch and getting the metadata attached to it appropriately and getting enough of it that’s a serious contender and packaging all of this up into a device people can just plug into their TVs cost-effectively is a big challenge. A huge challenge, in fact.
Finally, there’s the matter of identity: even with the prevalence of iPlayer, people still differentiate between the BBC’s different channels (even without the BBC’s own brand identity efforts). It’s a safe bet that a documentary on the life and works of Tolstoy is probably BBC Four fare, while a new 18-35 comedy series is at home on BBC Three. Technology aside, there’s something to be said for the editorial choices of others. This is relatively easy to solve, though: rather than building a profile for a complete channel, you start with one of the existing ones and add/remove as desired (e.g., “BBC One without the property shows and more drama”).
All in all, I don’t think it’s something that’s pie-in-the-sky, but by the same token I think it’ll be a good ten years (a lifetime in technology, and a lifetime in some respects in the broadcast world) before we see anything which achieves it with a serious chance of success and survival.
