While well dimensioned 3G networks are offering fast Internet access today, some somewhat underdimensioned networks show the first signs of overload. Some industry observers argue that the answer is to introduce tiered subscriptions, i.e. the user gets a guaranteed bandwidth or a higher bandwidth if he pays more. But I am not sure that this will work well in practice for two reasons: The first reason is that when some users are preferred over others in already overloaded cells, the experience for the majority gets even worse. And second, if such higher priced subscrptions get more popular because the standard service is no good, it won't be possible at some point to statisfy even these subscribers. So such gold subscriptions just push out the problem a bit in time but otherwise don't help a lot. There is just no way around sufficient capacity or your subscibers will migrate to network operators who have made their homework. So instead of only investing in QoS subscription management I would rather also invest in analysis software that reports which cells are often overloaded. That gives the operator the ability to react quickly and increase the bandwidth in the areas covered by such cells. Having said all of this, what do you think?