The McKinsey Quarterly has an article on pathway modeling, and how to make marketing more scientific. The article makes some good points, but what really struck me was this paragraph about the problems with innovative brand building.
Once marketers have their eye on the most promising future segments, they must rethink the brand-an increasingly complex process. Brand proliferation and rapid imitation have diminished the return on clever advertising and "breakthrough ideas," such as adding a "miracle ingredient" to a detergent or associating a sports star with a particular brand of athletic shoes. Today, cost-effective brand building depends on knowing precisely what consumers care about and tailoring the brand accordingly. Sophisticated new analytic approaches provide the precision but only when coupled with conceptual clarity in first defining a brand and then actually delivering it through a variety of what marketers call "touchpoints," the sites where consumers interact with it.
I think Seth Godin has the right idea, that Purple Cows are needed to really differentiate you from competitors. The problem is that purple cows are increasingly easier to copy, and companies can do it faster all the time. I guess that is why innovation is so important, so that you can always stay one step ahead of the copycats.