Not that I disagree with your basic premise but I wonder why some companies get away with this type of thing and others don't. Your example of Apple is an obvious one but they have plenty of "failures" on their hands as well. Whether it is something that is dead on arrival like Ping or the company letting one of its old marquee products die a slow death like the MacPro or iPod classic, Apple seems to be able to brush off these issues like no one else.
The services Apple kills are usually rubbish. Ping, Mobile Me, those iPod speakers. Google is killing things that are actually good. I think that is where the Reader backlash came from. Reader was good even if the relative popularity was low.
Hardware is kind of a different thing. The iPod was pretty nice. But now I want a shinier iPod with more features. Production of the Classic just has to match demand from those with niche uses for it. So a slow death makes sense?
I think it's different when there're concrete numbers that the company can use to inform their decisions. Apple only does profitable things; if they kill a product, it's because it's no longer making them enough money.
