There are three objectives that should be satisfied by the biscuit packaging process. Every packet of biscuits should ideally be exactly at the declared weight and in addition each pack should contain the same number of biscuits and the length of the biscuit pack should be invariant. Any deviations from these desired targets can impose costs on the manufacturer or have negative quality implications. However, because there is random variability in the weight and thickness of individual biscuits, it is not possible to achieve uniformity in packet parameters using existing packaging methods. A methodology is proposed whereby other possible approaches in assembling packets of column wrapped biscuits can be investigated. A Monte-Carlo computer model, sampling from the probability distribution functions of biscuit weight and thickness, has been developed to quantify the efficiencies of alternative packaging strategies. Packaging by fixed biscuit count is shown to be superior to the existing approach.