[6ren]

"The Design of Everyday Things" talks primarily about functionality, and frames design so broadly as to be almost indistinguishable from "solving a user problem".

As an example, Roy Fielding describes the URLs that a RESTful webservice includes in its representation of a resource (for what transitions are available to other states) as "affordances". It could even be argued that Codd's relational model was a better "design" for thinking about databases, which he presented in terms of the problem of data models being too closely coupled with storage representation.

Of course, even this broad sense of design doesn't address whether there's a market for a solution; but it does address whether you can make a solution that's better. I can see the sense in seeking a problem that needs to be solved - in being "market-driven"... but personally, I'm much more excited about creating something better (which is only possible when you already know the problem and some existing solution, because "better than" takes two operands). And that seems to be the history of all the products I admire.