| With the booming development of the Internet and market segmentation in the FMCG industry,market demand is increasingly fragmented and scenario-based.In this context,how to gain timely and accurate insight into consumers’ needs in a frequently changing market,improve purchasing efficiency,optimize inventory management,and reduce inventory costs and consumer transfer costs is a pressing issue for FMCG giant O Enterprises.This paper firstly selects brand A as the research object from the many brands of enterprise O.According to the classification of brand A,we study its product structure,price distribution,sales distribution and other basic data,and then according to the product life cycle,6 months as a sales cycle,through the past three years of sales data and product characteristics to establish feature engineering,using neural network regression to sell products under brand A of enterprise O The model is then evaluated using performance metrics;secondly,before the start of the next sales cycle,based on the forecasted data,a primary order is placed,two two-stage ordering strategies are proposed based on the inventory situation at the end of the first stage of the plan,and a secondary update of demand is then made through the exposure of real market demand,where Bayesian updating is used as the demand update method;finally,based on the Finally,the second stage of ordering is based on the secondary order point provided by the supplier and the inventory situation at the end of the first stage.This paper compares the expected profit of a single-stage purchase under a normal distribution of demand with that of a two-stage purchase with demand update,using real data arithmetic examples of the star products of five categories under brand A of enterprise O as an example.When the secondary purchase price is greater than the primary purchase price,the expected profit of using two-stage procurement strategy1 for all five products is greater than the expected profit of single-stage procurement,with the expected profit of two-stage procurement for Care type products being 38 pp higher than the expected profit of single-stage procurement;when the secondary purchase price is less than the primary purchase price,the expected profit of using two-stage procurement strategy for all five products is greater than the expected profit of singlestage procurement,with the expected profit of using two-stage procurement strategy for Care type products being 38 pp higher than the expected profit of single-stage procurement.When the secondary purchase price is less than the primary purchase price,the expected profit for all five products using the two-stage procurement strategy is greater than the expected profit for single-stage procurement.In addition,based on the different secondary sourcing points provided by the suppliers,this paper also determines the optimal secondary sourcing points and the optimal sourcing volumes for the five different products.In summary,the use of two-stage sourcing with demand updates is better than single-stage sourcing in terms of both cost control and profitability. |