I've worked (as a support engineer) at software companies where, over a 7 or 8 year period, almost every member of the development team has been replaced, and those that weren't could have been. The same was true of the support team too. I then left and the company survived the shock just fine, so clearly I was replaceable too. So yes developers are vital, there would be no product without a development team to develop it, but then without a sales team it would never get sold. Without a support team it would be unsupported, etc. Removing any one team would be like cutting one of the organs out of someone's body. Sure you can outsource stuff, but the function being outsourced still needs to get done.
The problem comes when people from one world start making the decisions about how to run one of the other worlds. In my experience it's pretty rare for a technologist to try and tell the sales and marketing team how to do their job and expend their resources. Unfortunately I've seen several companies essentially collapse because really good sales and marketing people tried to run the engineering groups and make what are fundamentally technical decisions. I've never yet seen that end well.
I think it's partly because if you run a duff sales campaign, you can regroup and try an alternate strategy relatively quickly. The problem with development projects is they can take several years to come to fruition and if you're on the wrong path it can take several more years to fix the problems. By then it's often way too late.
The problem comes when people from one world start making the decisions about how to run one of the other worlds. In my experience it's pretty rare for a technologist to try and tell the sales and marketing team how to do their job and expend their resources. Unfortunately I've seen several companies essentially collapse because really good sales and marketing people tried to run the engineering groups and make what are fundamentally technical decisions. I've never yet seen that end well.
I think it's partly because if you run a duff sales campaign, you can regroup and try an alternate strategy relatively quickly. The problem with development projects is they can take several years to come to fruition and if you're on the wrong path it can take several more years to fix the problems. By then it's often way too late.