Last time I used it (I can still see it on my whiteboard) is when designing a spam filtering process. I wanted to get rid of the most spam as cost efficiently

Last time I used it (I can still see it on my whiteboard) is when designing a spam filtering process. I wanted to get rid of the most spam as cost efficiently as possible. So that meant figuring out the cost of each method (Big O) and then figuring out approximately how much spam the process would catch. Thus, I could run the most expensive processes last on the spam which still hadn't been caught.