I regularly get vendors suggesting ways to monetise networks by providing higher guarantees of QoS, either offering users a "turbo button", or perhaps suggesting that an online video provider might want to pay extra for a partitioned part of a broadband pipe.
Sounds fair-enough in principle... but how would it actually work in practice?
I'm reminded of two roads that go through the centre of the UK - the main M6 Motorway, and a "sister" road called the M6 Toll which bypasses Birmingham. One charges for "quality of service" and the other is "best efforts". That's great - if you're a busy haulier delivering refrigerated goods, or perhaps a salesman trying to get to Liverpool in a hurry, you can avoid the endless roadworks & hordes of slow caravan-draggers on the main road.
But if you were driving at 3am... would you still pay extra to use it? Unsurprisingly, the overnight rates are cheaper. But if you had realtime traffic reports, and you knew that the free road was completely clear, why would you bother?
Something similar will happen with the various policy and QoS initiatives being discussed now. At certain times, best-effort will be perfectly good. At other time, it will be almost good enough. And at peak periods, it will be congested and pretty useless.
I think that there will need to be a good amount of intelligence in applications to know when it's necessary to pay the toll, or when the network is clear enough to use without goldplating the service.
Of course, it's possible that the operators might try to impose differential speed restrictions or place "speed humps" on the best-efforts route. But frankly, if everyone knows that these are simply money-gouging tactics, they'll get dealt with harshly by the highwaymen of competition (OK, I'll stop the overextended analogy there).
All of this is going to be difficult to apply to mobile anyway, as the gating factor in QoS is usually going to be the radio network. Having 99.9% QoS is worthless if it's only available 70% of the time.