Back to search

University of Leeds

Evaluating alternative methods for forecasting convenience grocery store sales

Abstract

dc:description.abstract

Convenience grocery stores have become more commonplace in grocery retailing in Great Britain since the 1990s, with a substantial increase in the proportion of stores operated by the largest grocery retailers in Great Britain that can be defined as convenience grocery stores. Geographically, the convenience networks operated by the largest retailers are more spatially concentrated than their overall grocery networks bringing them into direct competition with retailers more traditionally associated with convenience retailing in some, but not all areas of Great Britain. As convenience stores have grown, so too has interest in site location research in finding techniques to best predict their success. This thesis is carried out with the support of Sainsbury’s and GMAP Ltd and specifically considers location based decision making for convenience grocery stores in Great Britain. Grocery retailers and their location planning teams employ models that are adept at predicting supermarket revenue. However, they find it more difficult to consistently estimate revenue to new or existing convenience store locations. From the outset of this research it was hypothesised that different locations in which convenience grocery stores are found may, in theory, require a different optimal methodology for forecasting revenue accurately. This thesis first offers a segmentation of the convenience market into 7 statistically distinct location types to begin to address this problem. Using the 7 location types as a framework, three methodologies for forecasting grocery sales are tested for their suitability for predicting convenience grocery sales in the different locations in which convenience grocery stores are found. These are: GIS buffer and overlay modelling, regression modelling and spatial interaction modelling. The different methods were found to have mixed success in predicting convenience store revenue. The regression model was found to be the most effective model on average whilst the spatial interaction model was found to be the best model for generating very good revenue forecasts. Contrary to popular belief, the GIS buffer and overlay model was outperformed by the regression model and spatial interaction model in the majority of locations in which convenience grocery stores are found. Overall, the modelling frameworks presented in this thesis provide a plausible kitbag of techniques which can be applied in different convenience location circumstances.

Degree

thesis:*
Name dc:type.qualificationname
Ph.D
Level dc:type.qualificationlevel
doctoral
Grantor dc:publisher.institution
University of Leeds
Year dc:date.issued
2016

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Hood, Nicholas Andrew
Advisors dc:contributor.advisor
  • Clarke, Graham P.
  • Clarke, Martin

Rights

Language dc:language.dcterms_iso639_2
eng

Identifiers

dc:identifier.*
Identifier
uk.bl.ethos.703357

Chain of custody

source
Harvested from
White Rose University Consortium
Base URL
etheses.whiterose.ac.uk/cgi/oai2
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
related terms
citation

Hood, Nicholas Andrew. Evaluating alternative methods for forecasting convenience grocery store sales. doctoral thesis, University of Leeds, 2016.