I think you're right in that there are examples on both sides. I think Foss software is often powerful, but not always as simple to use. It's designed for functionality. Commercial software is designed to be easiest to use.

As we develop best practices more and more for UX, I think that is fading. I don't think marketing alone is the issue. Foss users often promote the free software and the price point is good. Take internet explorer. It's heavily marketed and free (not FOSS) but it's use is low. Software needs to fulfill a need for the user.

Foss is often filling a small niche. There are only a few large FOSS projects with broad appeal. Even then, it is difficult to not have them fracture and fork, which is better for options but usually poorer for the individual user after they have chosen.

I think for most things that are common, FOSS will end up the default eventually. There are few things that Microsoft can do to word to make it more attractive than librroffice. Excel is still ahead but the gap is closing for most users