There is a good example in the high end digital art niche. A few years ago a little known software company released a program called ZBrush. Their 2.0 release rocked the industry and became a must have tool for CG animation in film and games. The company, Pixelogic, suddenly became a player in a tight space competing with the likes of AutoDesk.
The success catalyst, to me, was their lateral approach to solving a long time industry problem (real time polygon count). They innovated a new paradigm. And they executed with a rock solid product.