Ideate, refine, finalize, create. That's the way it's supposed to go, but, for me, it is rarely that linear.
I often find myself looking for a singular answer to a problem. While this method motivates me to do a great job with my final design, it hurts the intermediate steps. If I sketch something out that doesn't have obvious potential I throw it out (often literally.) Once I find "the one" I take that to final concept quite quickly. It would be much more helpful to attempt to develop multiple concepts before giving the bad ones the axe, leaving my mind open to possibilities throughout the whole process. This method of leapfrogging to the start to the finish probably arose from my first co-op, where I was critiqued for spending too much time exploring ideas.
Take my computer project for example: 200 ideations, 30 refinements, 3 further refinements - all thrown out the door for for my final concept, which was completed in three days. Though a part of it is still related to my early ideas, the main focus of my concept has no previous documentation. There aren't any sketches or explorative renderings. I'm fine with this, as I am confident that it's a good concept, but there is no proof of my thought process leading to my final design. It is hard to justify an idea that just came to me, even though that's what happened. It will probably come down to back sketching on an ongoing project. Weird, huh?
My only solution to this problem is to get much faster at the refinement stage so it is a much more useful design tool. If I am able to quickly refine concepts, I won't be worried about not having time the find "the one."