Split run does not appear to be a widely understood term; however, in my personal chemical process development experience it is a common term with a well defined meaning. Split runs are experiments done to test different work-up/isolation and/or purification procedures starting with equal portions from a single large reaction mixture. As a consequence, even if a reaction mixture is both complicated and uncharacterized, a chemist can compare different follow-on procedures for working it up without being concerned about reaction differences that may still exist from batch to batch because the reaction process itself has not yet been finalized and critical parameters completely controlled.
In the general case, the isolated yield fraction, (which when expressed as a percentage is often called simply the percentage yield), is the product of the reaction yield fraction (determined by an assay of the crude reaction mixture before work-up) multiplied by the isolation yield fraction (the fraction representing the effectiveness of the recovery of pure product from the crude reaction mixture). In the case of comparing split runs, the ratio of isolated split run yields is proportional to the ratio of the isolation yields because the pure reaction yield is the same for each, since they come from the same reaction batch. Thus, in comparing the recovered product from split runs, the difference represents only the effectiveness of the isolation protocols….it is independent of and not confounded by the exact details of the reaction conditions. In this way, even at an early stage in developing the particular reaction step, the development chemist can get some idea of the comparative effectiveness of different treatments of the crude reaction mixture.
Of course it is not just the best isolation yield that is of interest and of importance to the process development chemist but also the relative purities of the two products and the amounts of different impurities within each.