Despite the initial promise of combinatorial chemistry, particularly large library combinatorial chemistry, to greatly accelerate drug discovery, this approach has not been fully utilized as a means to build the compound collections of pharmaceutical and biotechnology companies. This review highlights some of the strengths of large library combinatorial chemistry as a means of generating molecules for lead discovery, such as providing rich and robust structure-activity relationships around each hit series. The challenges and concepts emerging from traditional high-throughput screening and fragment-based drug design, how these methods influence the design of large combinatorial libraries and the interpretation of the ensuing high-throughput screening data are also highlighted.