Treating this as an either/or discussion overlooks the reality of business, and the world, which demonstrates that to be successful you need to do both.
Great Marketing + Crappy Code / Product = pets.com
Great Code + Crappy Marketing = The-next-big-thing-nobody's-heard-of-or-will-pay-for.
Probably the most telling part of this was talking about the "the world perception didn't seem to value the geek's contribution".
If you can't demonstrate value (which is marketing in a nutshell) then the value (which is business cashflow and profit in a nutshell) won't be there.
You can't tie features, product, and code quality together either. Even the most brilliant code is worthless if it isn't solving the customer's problem. This was the issue with Eric Ries' first version of IMVU - they built an epic piece of IM integration software that they trashed because no one wanted it.
I think that's a nice sequence for product-marketing development.
At some point you'll have to deal with all the business support stuff (Premises, HR, Accounts). And then when you get big enough and start hiring people to do most of the product-marketing work, then you'll need to have some business management aspects like Vision/Strategy, client base management etc.
But this process should serve you through ramen profitable, and beyond.
Great Marketing + Crappy Code / Product = pets.com
Great Code + Crappy Marketing = The-next-big-thing-nobody's-heard-of-or-will-pay-for.
Probably the most telling part of this was talking about the "the world perception didn't seem to value the geek's contribution".
If you can't demonstrate value (which is marketing in a nutshell) then the value (which is business cashflow and profit in a nutshell) won't be there.