There is a fascinating case history of the development of five drugs in the current Nature Drug Discovery. Among other things, it concludes that target-oriented drug discovery isn’t working as well as its promoters say. It has partly to do with the notion of magic bullets and exclusivity — most molecules interact with multiple targets — as well as the desirability of exclusivity itself, where compounds that interact with a few selected targets are often more desirable anyway (i.e., in cancer).
As an aside, I love drug names. They all sound like minor devils from The Omen VIII: matinib, natalizumab, bevacizumab, rofecoxib and natalizumab.