The problem can be fairly mundane and not even unique. You just have to do it well, be enthusiastic and put your all into it.
Also, it's fairly likely you may need to try out 100 ideas, maybe more, before you find one that is good.
It's way easier to try 100 ideas until one succeeds, than doing a comprehensive 'on paper' analysis of ideas and assessing the risk of each one, building the one you've chosen, launching it and finding that it falls flat.
I seriously shy away from "saleable product", so my thoughts are purely from a conventional ad supported website model.