I think the main problem is that writing good software is not a commodity. There is no process or best practices that can ensure a successful software project.

I think the main problem is that writing good software is not a commodity. There is no process or best practices that can ensure a successful software project. Especially management is desperately seeking for a way to commoditize software "production". This would mean that with a cultivated process and a set of studiously selected tools and the right developers, the project would be guaranteed a success. The trouble is: every software project already tries to do that. What a business really aims is to be an exception on the "Law of large numbers". (http://en.wikipedia.org/wiki/Law_of_large_numbers http://en.wikipedia.org/wiki/Law_of_large_numbers) There is simply no static solution to be an exception.