Especially for the production of active pharmaceutical ingredients, the use of process analytical technology (PAT) is highly encouraged by the U.S. Food and Drug Administration. In crystallization and granulation processes, in situ particle characterization is the most important PAT. The technique of Focused Beam Reflectance Measurement (FBRM) is very well-suited for in situ particle characterization. A large community of users successfully applies FBRM technology for monitoring, fault detection, and quality control of dynamic processes. However, FBRM measurements are not easy to interpret, because the measured chord-length distribution (CLD) is different from any type of particle size distribution (PSD). For monitoring purposes, moments of the PSD are usually correlated empirically to moments of the CLD. Alternatively, process phenomena such as secondary nucleation and particle growth can be attributed to the time evolution of the number of chords detected in a length interval. To the authors' knowledge, no publication has examined the accuracy of such correlations or presented a methodology to set the boundaries for the chord-length intervals. In this work, a mathematical method is presented with which a set of measured CLDs can be reduced to a small number of chord length classes, so that the class boundaries are chosen in an optimal way. A reconstruction of PSDs from FBRM data is not used for process monitoring in this work, because such a reconstruction may lack accuracy, as shown in earlier work [Kail et al. Chem. Eng. Sci. 2009]. The method presented in this work relies either on a simulation using the optical FBRM model presented in earlier work [Kail et al. Powder Technol. 2008, 185 (3), 211-222] or on reference experiments. With the presented methods, a batch crystallization of alpha-lactose monohydrate is analyzed. In addition, measurement artifacts observed in FBRM data are explained and discussed.