You can try to improve small things, like reducing a lot of small losses in a plant (using improved insulation, stopping leaks, reducing friction, and so on). By contrast, you can redesign the entire process using new technologies or techniques to realize quantum gains.

I learned about this in a class called Analysis, Synthesis, and Evaluation during my junior year in college. It took me a long time to understand some of the material in that course, but this concept was obvious the minute I heard it. I tend to look for the big kill whenever possible, but that doesn't mean the incremental improvements should be ignored.

Big Kill improvements require a lot of insights and may involve many other concepts on this list. They often require reframing the problem entirely.

Big Kill opportunities don't always exist, or may not be useful even if they do exist (e.g., for commodity products in declining industries.)

Conversely, organizations may be afraid to capitalize on Big Kill breakthroughs because they cannibalize existing businesses. (see: Kodak)