Can you imagine taking something that an established company is already doing, creating a duplicate product, and just marketing it differently for a different audience? This clearly works with paper towels. Can this work with Internet startups?
I think there are already some examples of this - the foremost being LinkedIn and Facebook. They are essentially the same thing: social networks. I am sure that there is a large percentage of overlap between some of my LinkedIn contacts and my Facebook contacts. Yet, the marketing experience, implicit in the product, and explicit in how they portray themselves, tells you that they are for different uses.
Most people do have two personas - their work and personal personas. LinkedIn and Facebook have managed to segment against these and target them mutually exclusively. Minor tweaks in the product allow them to fit those segments that little bit better - Facebook with more photo features, for example, and LinkedIn with more explicit company networks.
But are there only two personas that people have? Even more fundamentally, are these 'work' and 'personal' personas the only dimension on which you could segment Internet users? A number of companies have been started that take the basic social network concept and add tweaks for each of their niches. They each segment differently - some target academics, while others target music lovers. Yet, have these companies STP'd optimally? More broadly, have any of the social networks, including Facebook and LinkedIn, identified the optimal segment to target?