Welcome to RequireRequirements.com, a place dedicated to software excellence and results.
Anything worth doing is not necessarily worth doing well. Cost is always a factor. It doesn't make sense to build a million dollar mousetrap when your home is the stomping ground for a single rodent. But software often outlives its intended lifespan, as evidenced by the level of quality in much of the software in the world today. Software is, still, consistently late or less-than-expected. But the committment to build better software is everywhere. We know we have an opportunity to improve the state of affairs.
Some things worth doing are worth doing very well; especially things that get reused, revised, and revisited time and time again. Unless we're specifically building throw-away software we've got to start with, at least, a definition of our expectations for the software. Anything less is a shot in the dark.
And if something is worth doing well, we should study how to do that thing well and constantly work on how to do it better.