When a chromatographic separation process is numerically optimized, a number of input parameters to the column model need to be accurately determined, and the accuracy will affect the reliability of the predicted optimal operational conditions. Furthermore, the numerical accuracy of the solution to the column model will have similar impact. The input parameters holdup volume, injection profiles, and the selection of an algorithm for solving the column model were investigated. Errors in parameters or a numerical error in the solution of the column model were found to lead to a prediction of a product fraction where the set purity and/or yield requirements are not met.