This photo is one of my favorite on the internet - it looks confused and incomprehensible at first, but then, from the right point of view, it takes on perfect and surprising clarity.
This shift is sort of the point that Seth
Rogin Godin is makes in the video below (recently on BoingBoing). Something might start out with the best of intentions, but in reality comes out horribly broken. This can happen for many reasons, but the end result is the same: it is no good.
It's difficult to play out potential scenarios beforehand, and trying to anticipate every possibility can paralyze something before it even begins. But it's worth some effort, and it's also important to build in flexibility and allow for mid-stream course corrections.
Of course, that's not always possible. When starting something new, try to work out who the users will be, and what their needs are, and then incorporate that into the available resources. And if things aren't working, start to stray, or just plain break, take some time to get input, reassess, and try to fix it. But don't be afraid to scrap the whole thing and start over - try something, learn from it, and try something else.