Basically, it's a credibility problem. Credibility, once lost, is damn hard to recover. Among the best cars I've ever owned is a GM product from the early 1990s -- a 1993 3.8 liter V-6 Oldsmobile Achieva. I put more than 120,000 miles on the car, and it had only two real flaws. It was a little over-engined and consequently it ran through the lower motor mount called a "dog bone" because of its shape, about every 30,000 miles. And after I had about had put about 115,000 miles on the car (meaning the car had seen 145,000 as it wasn't new when I bought it), the driver's side power window quit working. It was still running strong when I sold it to a teenager (I don't know if he ever fixed the window). I'd frankly love to have a new one just like it.
The flip side of that story, of course, is that I once owned a truly iconic foreign car -- a 1973 1600 cc Volkswagen Super Beetle. It was a true $#@!box, and I've worked hard ever since not to judge all Volkswagen's by the one experience.
Speaking of Olds et al, I think one big problem GM had back in the day, vis a vis the likes of say Honda, was in their repair/service model. GM, in concert and contract with their dealers, assumed too much of their bottomline to come from service/repair.
I'm not necessarily thinking in terms of their designing repair issues into their cars. It goes more towards them not thinking in terms of minimizing the customer's exposure to their dealer's service departments. As a for instance, we all remember back in the 80's, people with GM/Chryslers having to have their "computer" or "engine module" replaced. The downside was that the computer was expensive. The good side (so to speak), was that replacing the computer usually did tend to "fix" the car. People would
about the price of the computer, but they were somewhat mollified by the fact that at least their car ran fine again. But of course..when they bought their NEXT car, they'd go with another brand.
Here's where it didn't make sense though: Per the back-channel GM data, of ALL the computer/control models that were returned to the factory under warranty, fewer than 3% were ACTUALLY defective, even as they continued to compensate their dealers for replacing them.
So...why in the world would GM put up with that? And, why were they willing to leave their customers with the perception of a poor-quality product, that wasn't actually of poor quality? I'm not inclined to the various conspiracy theories that might pop into people's heads around this sort of thing. I'm MORE inclined to putting it down more broadly, to their old profit model.