Kevin Maney has a book out in September about the trade off between delivering extraordinary experiences (which he calls fidelity) and doing it in a way that's cheap and easy (convenience). The book takes this simple idea and supports it with dozens of examples.
The simplest example is movies. You pay to go to a theatre when you want the fidelity of the big screen and the crowd and the speakers. You stay home when you want the convenience of Netflix and the pause button. Vinyl records and live concerts offer fidelity, MP3 on your iPod is convenient.
In the words of Bill Gross, in order to win with a new product, you need to be on one axis or another, and ten times better than what you're aiming to replace. Which means ten times more high impact or ten times cheaper and easier.
A refrigerator is ten times more convenient than an icebox. A cell phone is ten times more convenient than a pay phone. A private jet is ten times more joyful/fidelity than first class for the executive that can afford it. A backstage pass at a Cat Power concert is ten times higher fidelity than a ripped MP3.
There are interesting ways to define 'fidelity'. Wikipedia is certainly more convenient, and the presence of a million articles that aren't even in the Brittanica makes it higher fidelity as well. At the start, though, they were neither. It took a (relatively) tiny group of passionate people to create enough content and enough quality that they could be high enough fidelity to be considered an alternative to the printed encyclopedia. It's interesting to watch Luddite teachers refuse to accept it as a source, claiming that convenience shouldn't trump their definition of fidelity…
The mistake that's so easy to make is to be a little bit higher fidelity and a little bit more convenient. Incumbents fall into this trap all the time, assuming that you'll stick with what you've got because they're sorta both. And insurgents almost always fail because as geeky insiders they think that twice the convenience is enough to persuade anyone who cares. Not going to work.