One of the big problems with car companies is that it takes so much time and money to change models in response to changes in taste. Car companies make bets on design and technology up to a decade before the products reflecting those choices debut on the market. Markets change but the car companies hardly turn on a dime in response. They are more like those movie car chase scenes in which cars miss turns and run over lawns, onto sidewalks, and through parks and run over small cabanas, driving straight while the road turns.
Modern technology might make it a little easier for car companies to move with market demand.
It'd be interesting to develop simple-to-use car design software and make it publicly available on the Internet. Users - everyday customers - could make their own designs and people could vote on these - or even tweak them to win more votes.
Coincident with this, at a particular stage you could have people "bet" on the designs by making deposits, putting earnest money down on a car before GM or Ford or Toyota even begins to make production plans. Money "bet" on designs would fund next steps of development. Designs that made it progressively through various stages of development to actually make it to market would then offer payback to investors in this early stage. Betting on the right design at the onset might yield, say, a 100 to 1 return, whereas getting the "bet" wrong would result in a loss of the investment. Investors would be able to bet more precisely than just buying GM or Ford stock, steering the company towards specific models and designs.
Further, if your design was one that moved into production, you might win $1 million or even more. You might get a portion of the profits. (And yes - this could even attract the attention of car designers from inside your company. So what? Why shouldn't employees who help save you from billions in losses get an extra million or more?) So the car companies would turn their production capacity into a tool to reflect demand by rewarding both design and trend spotters who invest.
It would seem that if car companies could make as much of their at-risk design work less risky, they’d be less likely to put billions into creating cars that no one wanted. Let the designs fail when they are still virtual rather than, five years later, when those designs are sitting on show room floors.