Some of the worst thinking you can have is incremental thinking--where you go down one path, get stopped by something major, and then do some kind of hacky workaround to get to what you think is your end goal. Often times, what you wind up with us half a plan or product, and half garbage. If you don't use roadblocks as opportunities to reexamine the model and first principals from the beginning, then there's a good chance your'e going to wind up with a lot of wasted effort.
I've encountered that several times. It's so tempting to not want to start from scratch--to see if you can try and mold what you have into the answer, even though you know what needs to be the solution.
Like a lot of other theories, it seems, you can find this in both relationships and technology. How many times in a bad relationship do people just try and fix one superficial thing--the squeeky wheel as it were--when the extent of their problems start with the fact that they never should have went on date number two.
It happens in technology all the time--especially when it comes to confusing design problems for technology problems. When you hear someone say that they can't make a certain technology do something, it's just as often a design problem that started with the very first conception of the idea than it has anything to do that X technology can't scale or doesn't work with Y or whatever.
So, next time you're taking advice, and you're deciding between someone telling you, "We can fix that" versus "That's effed... you need to tear all that out and start from scratch", maybe you shouldn't dismiss the latter so quickly to save a buck or save your sunk cost.