Browsing by Author "Dagnin, Monique"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
- ItemDetermining the mobile device offering at a large SA retailer(Stellenbosch : Stellenbosch University, 2023-03) Dagnin, Monique; Nieuwoudt, Isabelle; Stellenbosch University. Faculty of Economic and Management Sciences. Dept. of Logistics.ENGLISH SUMMARY: The retail industry is one of the biggest industries in the world and an important factor in the success of retailers, is carrying the correct products for their customers. The Retailer in this study, like many other retailers, provides a range of financial services and products to their customers to add value and improve the customers’ experience in their stores. Therefore, the aim of this study is to assist The Retailer in determining the best range of mobile devices to keep in stock in their stores. The Retailer has over 1 280 stores and was seeking device ranges for store groups, rather than a unique range for each individual store. Therefore, stores with similar characteristics are grouped based on several external factors that were identified. Hierarchical clustering is used to group similar stores within each supermarket type based on the number of landlines, rate of population change, population age and population income. Six clusters, two per supermarket type, are found with this method. For each group of stores, the range of mobile devices to keep in stock is determined using three performance measures, namely rate of sale, total units sold and average units in stock. These measures are calculated to evaluate the performance of mobile devices and rank the devices according to their performance. Two iterative approaches are followed to determine whether a device should be ranged in any of the six clusters. For mobile devices that have not been ranged in a particular store, but should be ranged according to their performance, the required stock level in these stores is determined by estimating the rate of sale per store using a regression tree for each mobile device. To build the regression trees, population age, rate of population change, population income, number of landlines, store size, province in which a store is located, adapted mobile device category rate of sale, average sales amount per store and total number of mobile devices sold in a store are used as independent variables. The methodology is illustrated using a selection of The Retailer’s devices currently in stock. Eleven of these 30 sampled mobile devices are ranged using this methodology, suggesting that this methodology succeeds in reducing the variety of mobile devices ranged by The Retailer by removing under performing devices.