Predicting supermarket sales: The use of regression trees

Abstract

-reassessment of a point of sale's performance, when competing chains open new stores in its 'area of influence'; -restructuring the layout of points of sale in order to increase their performance; -the location of already planned new points of sale; -possible closures of points of sale.Simple rules and easy to use operational specifications for company decision INTRODUCTIONThis paper is part of a study performed in 2002 involving a supermarket chain. Its primary goal is the search for predictors that are relevant in explaining the performance of the chain's points of sale. In addition, this study aims to gain some understanding concerning the relationships between relevant predictors of supermarkets' performance.This study also intends to contribute to the management decision-making process in connection with: Abstract The performance of points of sale is influenced by multiple factors, including the characteristics of their external environment. The main focus of this study is the analysis of a supermarket chain store's external environment and its influence on the store's performance. Principal components analysis is used for data reduction of store attributes (eg visibility and accessibility). A regression tree is used for stores' sales prediction and clustering. The first tree branching divides smaller and larger stores. Factors like the stores' lack of visibility and heavy road traffic contribute especially to the decrease of sales in smaller stores. Factors like the availability of public transport and the absence of a specific competitor in the neighbourhood contribute to a better performance of larger points of sale.In addition to the insight it allowed concerning the performance of the existing stores, and notwithstanding some limitations (eg a static modelling approach is considered), the proposed tree model has been used successfully in predicting the performance of new stores.

Leveraging locational insights within retail store development? Assessing the use of location planners’ knowledge in retail marketing

“…Multiple Regression Attempts to define a correlation between store sales and variables within the catchment that influence performance Silva and Cardoso (2005) Geographical Information Systems (GIS)…”

A Real-Options Approach to Modeling Investments in Competitive, Dynamic Retail Markets

“…Trees (CART) [35]: the financial performance of a retail store is influenced by a large number of factors. Principal Components Analysis (PCA) is used for dimension reduction in the number of variables that are used to explain variations in stores' sales.…”

Comparative Analysis of Regression Algorithms used to Predict the Sales of Big Marts

“…It is necessary to have a proper inventory management system for the store to be able to meet the demands of their customers. The inventory of the store is tracked through a software which is updated every day so as to make sure that the store is not running out of any item that is demanded by the customers [2].…”