A conic optimization approach for SKU rationalization

Expanding variety and the number of offered products is attractive for a firm to fit customer needs. Nevertheless, the greater complexity and the proliferation of stock-keeping units (SKUs) without substantial differentiation may not substantially improve customer satisfaction while raising costs. Based on the principle that the product-line size involves operational implications and particularly manufacturing and holding costs, this paper develops a mixed-integer nonlinear mathematical program (MINLP) to support efficient product portfolio reductions. Basically, the fixed costs elimination and the risk-pooling effects must balance the demand contraction due to customer dissatisfaction. Off-the-shelve Mixed-Integer Quadratic Problem (MIQP) solver provides optimal solution to the proposed conic quadratic reformulation, and a real-life industrial case illustrates the program and the algorithm efficiency. Findings show that our mathematical programming subject to various assumptions and estimations is efficient to rationalize portfolios up to at least 400 SKUs.