Make room for randomness in drug development

Working as a medicinal chemist in early-stage drug discovery means running lots of experiments, because you’re going to be making lots of compounds. A typical project will generate hundreds of new compounds at the very least, and could well run into the low thousands of analogues before things are finished. And there’s never been a drug research effort where there weren’t more compounds that could still be made at the end of it.

How, then, do you figure out which compounds to make?