Natural products are biologically validated starting points for the design of combinatorial libraries, as they have a proven record of biological relevance. This special role of natural products in medicinal chemistry and chemical biology can be interpreted in the light of new insights about the domain architecture of proteins gained by structural biology and bioinformatics. In order to fulfil the specific requirements of the individual binding pocket within a domain family it is necessary to optimise the natural product structure by chemical variation. Solid-phase chemistry is becoming an efficient tool for this optimisation process, and recent advances in this field are highlighted in this review article.