Not sure why you downvoted, I thought you made some great points. However, your example helps illustrate what I've been saying.
Say you want a car. You may want to go to the dealership and just buy one (Toyota's automated process, thus lowering marginal cost). However, maybe you're the hotrod type and would rather build it and customize it yourself. You pay for someone to help you build it and troubleshoot any mistakes you make along the way. Sure, it took longer to build, and was probably more expensive, but it's exactly want you wanted. I think that's where the value of opensource/support lies.
Commercial open source companies (at least Redhat) use subscription models for their support (in addition to consulting, professional QA, etc) so I don't think they would consider troubleshooting software money earned. In fact, I think this gives them a greater incentive to maintain the quality of their software so they get less support tickets and scale to have more subscription fees per support person. I guess you can say that RedHat, through their support, sells the peace of mind of having that support person if you want to go the open source route.
I've no idea why I'm receiving downvotes either. I guess a few folks are angry about having their business model attacked with reality.
Addressing your post: Software isn't a car and the marginal cost to manufacture more software licenses is zero. When I direct an hour of effort to our commercial software, I'm gaining many dollars of future passive revenue thanks to that created wealth.
Contrast that with services and a 'free' product, where each hour of effort gets a support dollar but generates no wealth at all! Pitted against that means of generating revenue, a service company giving away its product is going to lose in the long-run every time.
To see this, just look at the software marketplace: Aside from the solitary black-swan exception of RedHat (which arguably sells licenses instead of services anyhow), where are the profitable open source support contract companies?
I can spool up 3,000 software licenses sold today without batting an eye (and add in support charges on top of that if needed) making a ton of cash with which to hire employees for product improvement, but 3,000 hours of support is a huge cost requiring a great many employees to manage before even a penny of profit is obtained.
Bringing this back to my original point: Commercial software companies can easily dominate 'free' open source alternatives given that reality, so they can safely ignore their free 'competition'.
Say you want a car. You may want to go to the dealership and just buy one (Toyota's automated process, thus lowering marginal cost). However, maybe you're the hotrod type and would rather build it and customize it yourself. You pay for someone to help you build it and troubleshoot any mistakes you make along the way. Sure, it took longer to build, and was probably more expensive, but it's exactly want you wanted. I think that's where the value of opensource/support lies.
Commercial open source companies (at least Redhat) use subscription models for their support (in addition to consulting, professional QA, etc) so I don't think they would consider troubleshooting software money earned. In fact, I think this gives them a greater incentive to maintain the quality of their software so they get less support tickets and scale to have more subscription fees per support person. I guess you can say that RedHat, through their support, sells the peace of mind of having that support person if you want to go the open source route.