MACHINE-LEARNING–BASED PRODUCT ASSORTMENT RECOMMENDATION FOR RETAIL OUTLETS IN DISTRIBUTION: CROSS-SELL AND ASSORTMENT-WIDTH GROWTH (CONCEPTUAL MODEL)

When an agent or van-seller visits an outlet, an important decision is to offer a product that the outlet does not yet sell but is likely to accept (cross-sell), thereby increasing assortment width and the average order. The concluding part of this three-part series proposes a conceptual model for recommending products to each outlet based on item-item similarity (market-basket) and describes the methodology for comparison with a popularity-based baseline (Precision@k, Recall@k).