This is a great observation. I don’t have the source for this available, but someone wrote a really great article on when open source licensing should be used vs closed-source. These were some of the take aways. The foundational issue to examine is “enabling technology.” Most companies are not in the business of software, but are enabled by it. So, software is a cost center and not a profit center. In regard to enabling technology, there are two broad varieties: differentiating and non-differentiating. Differentiating refers to whether the technology creates a competitive advantage for the developer.
It makes sense that all non-differentiating technology should be open source. That is why open source has taken over the server and backend. The software in the backend is infrastructure that enables every company that uses the web to be able to deliver the proprietary services they want to deliver. No one company gets an advantage from re-writing those services from the ground up, and everyone benefits from changes that contributors make to the software. (Theoretically.)
A system integrator also wants to use open source solutions where possible because it helps the mission of acquiring software and hardware as cheaply as possible; it also avoids vendor lock-in. This enables the system integrator to increase both its profits and its customer base.
Software vendors have an interest in using open source components for these kinds of non-differentiating technologies because those components reduce the cost of shipping the first copy of the software.
For an end-user, using open source is good for a non-differentiating technology (e.g., the infrastructure example), but that user is economically better off to continue using closed source for differentiating technologies.
As you can see, most of these kinds of issues are of interest to the enterprise.
On my Mac, I have a collection of open source tools and closed-source tools. I use VIM for all my plain text work. I use Scrivener or Word for my hardcore writing. Why? Because the closed source tool is more refined and capable (and pretty or approachable or whatever we measure UI with) for the reasons that @ibuys identified.
If you develop Scrivener or Paprika or Drafts or PDF Expert or Excel, you benefit economically if your solution is popular and people buy it. If you develop those apps, the technology is differentiating you from the competition and you should not develop it open source (unless you are just a true open source purist). If you develop it open source, you lose the opportunity to monetize it. UNLESS, your app would benefit from selling service or integration (e.g., Red Hat’s model).
I’m barely scratching the surface, but I hope it illustrates the point.
If I can find the article, I’ll link it back here because I think anybody reading this thread would probably also find the article an enjoyable read.