The goal of packing optimization is to provide a foundation for decisions related to inventory allocation as merchandise is brought to warehouses and then dispatched. Major retail chains must fulfill requests from hundreds of stores by dispatching items stored in their warehouses. The demand for clothing items may vary to a considerable extent from one store to the next. To take this into account, the warehouse must pack “boxes” containing different mixes of clothing items. The number of distinct box types has a major impact on the operating costs. Thus, the PrePack problem consists in determining the number and contents of the box types, as well as the allocation of boxes to stores. This paper introduces the PrePack problem and proposes CP and MIP models and a metaheuristic approach to address it.