A comprehensive experimental approach to the optimization of the process parameters for the production of methyl-isobutyl-ketone (MIBK) from acetone was developed and successfully implemented. This includes conventional laboratory analysis, in-line near infrared (NIR) spectroscopy (with associated principle component analysis) and statistical interpretation of the experimental results. The effect of temperature, pressure, feed rate and the H-2 to acetone molar ratio on several responses, i.e. acetone conversion, MIBK selectivity, MIBK yield and in particular MIBK productivity, was investigated on a continuous bench scale reactor system. Using experimental design methodology, empirical models were developed that describe each response as a function of the significant variables, enabling optimization of the process. The MIBK productivity model was validated with an independent experiment. In this paper the use of multidisciplinary chemometric techniques is demonstrated for the attainment of substantial industrial improvements. It is shown that the acetone to MIBK productivity is improved by 30% through operating at a higher pressure and feed rate, provided that the appropriate H-2 to acetone molar ratio is used. It was further shown that there is good agreement between the predictions of the response surface model derived from NIR-based transient catalysis testing, and conventional steady state results. (C) 2008 Elsevier B.V. All rights reserved.