Ford uses Personas...

Does that mean they're played out? Or is this validation that they can be useful?
A model depicting Antonella, the imaginary woman who was the guiding personality for the Ford Verve, a design study that was the basis for the new Fiesta.

A while back a 37 Signals blog post started a whole debate in the UX community about the usefulness of personas as a design tool. As someone who uses them myself, I find that they can be very valuable, especially if they are based on real human beings with real needs & wants who you can ask questions and present concepts to. (I think 37 Signals would say something similar--build for a real person, not an imaginary one.) I have always shied away from personas based on demographic data, or fiction, rather using demographics to point me in the right direction to find a real human being to interview. But here Ford seems to be using demographic data to create fantasy people to make fantasy products for, and it seems to be working. Perhaps because cars are fetish objects as well as tools it's OK to design for an idealized fictional person. Especially if you've done the research right, and can attempt to understand desire through data. But this brings up another debate that is often discussed in the UX community--the cool vs. the usable. Do we lose something by focusing solely on the usable and useful? (that's for another post...)

ANYWAY, what do you think about this issue? Personas or no? Demographics based or real people? Or something completely different? Who do you design for?