The final product quality of crystallization and precipitation processes is often determined by the interplay between mixing and relevant phenomena such as, nucleation, molecular growth, aggregation and breakage. In this work the attention is focused on the aggregation and breakage processes and on the level of details needed to accurately describe them. Different modeling approaches are tested and compared for two different shear devices: a Taylor-Couette cell and a stirred tank. Results show that when aggregation and breakage processes are much slower than mixing a simplified homogeneous model can be used whereas when aggregation and breakage occur on a time-scale comparable with that of mixing then a detailed approach based on computational fluid dynamics (CFD) is needed. The very common lumped model, used to derive kinetic expressions for aggregation and breakage rates, can eventually lead to serious errors.