Microbial secondary metabolites are one of the sources of therapeutic molecules in the pharmaceutical industry. Product quality and high yields of secondary metabolites are the main goals for the commercial success of a fermentation process. Our novel approach was based on the decision-tree algorithm to determine the key variables correlated with the process outcome and on DOSY-NMR to identify both co-metabolites and impurities, and it improves fermentation systems and speeds up bioprocess development. The approach has been validated in the case of lovastatin production from Aspergillus terreus.