Share of Local Assortments in Revenue for a Large Scale National Food Retailer : Assessment of Long and Short Time Series Patterns

Recent market trends show an increasing strength and number of competitors, creating a more challenging environment for national retail chains. In response, these chains are adopting more curated assortment management strategies to maximize operational efficiency. Particularly, the management of local product assortments has emerged as an effective strategy to attract customers interested in products from their neighborhood, ultimately increasing local outlet revenues. Our study utilizes advanced econometric time-series methods and their machine learning counterparts to assess the trend of local product assortment share in the revenue of organized national supermarkets. Our primary goal is to identify local neighborhoods more inclined towards this product group using prescriptive analytics. A dynamic root-cause analytics framework is applied on trend of local product sales. Additionally, we aim to forecast the share of local assortments in total revenue. Given the low dominance and standard deviation of this product group’s sales share, we employed problem-specific success metrics to evaluate our model. Understanding and projecting the pattern of improvement in local product assortment enables retailers to devise the right strategy for managing procurement from local suppliers and sales activities more effectively.