The garment manufacturing industry is reported to be experiencing customer complains of misfit. Studies on this problem seem not to have provided solutions that satisfy customers and manufacturers of clothing. In this study, Customer Pairwise Fit matrix (CPF) was explored to group a given customers population for the purpose of garment sizing. A mathematical fit function expressed as Percentage Fit (PF) between pair of customers was developed, CPF matrix for a number of customers indicating the degree of fit between any pair of customers was derived. A grouping algorithm from literature was adapted with the CPF matrix to divide a given population of customers to garment sizes, and demonstrated using garment anthropometric data related to trousers of 100 male potential customers in Nigeria. The results obtained for the PF between any pair of customers indicated 3 possible results of ‘no fit’ (PF = 0.0%), ‘partial fit’ (0.0% < PF <100%) and ‘perfect fit’ (PF=100%). The outcome of the adapted algorithm to the 100 potential customers led to 19 set of customers but 18 sizes appeared to give the ‘best’ possible fit of 99.31%. The procedure developed in this work may be useful for sizing of larger customer population and any garment type.