My bias:
Prices ought to reflect actual cost, which lets people make informed decisions about risk (or cost) and benefit. That should apply to big companies, small companies, and consumers.
If flood insurance were priced based on actuarial cost instead of some politically acceptable number, very few people would choose to live in areas likely to get flooded. Those who chose to do so would have been paying their share of the actual risk, meaning less for the tax payer to pick up.
If something like Skuzzy's outline or the NZ system were in place, then you'd basically get what you paid for - and you'd only be paying for what you got. It allows niche providers to survive when they might get buried by megacorp. Also, we wouldn't pay for Netflix related ISP upgrades, except to the degree that we actually used that kind of bandwidth. The little old ladies wouldn't be stuck subsidizing the kid who was P2P downloading every movie he could find whether he wanted to watch them or not. And the kid (or his parents) might get surprised with a bill that encouraged them to change their behavior.
It's an approach based on something called "signaling theory", and it seems to me that a great degree of economic problems could be solved by allowing prices to reflect actual costs - and then relying on people to make the judgments that make the best sense for them as individuals. Anything that distorts the actual costs (subsidies, collusion between corporations, monopolistic behavior) causes inefficient allocation of resources and reduces the amount of good stuff the economy can produce for all of us.