Predicting Profitability of Neighbourhood Stores in Colombia

Abstract

This paper examines the neighbourhood stores as part of the Colombian micro businesses landscape. Decision models were constructed to identify the profitability of neighbourhood stores, with the use of statistical exploration methods of analysis such as t-Test, Logistic Regression, Discriminatory Analysis, Decision Tree, Random Forest and Artificial Neural Networks. The database used for this study is the result of a survey carried out to the shopkeepers in 336 different neighbourhood stores in the city of Bucaramanga, Colombia, in 2014. The main finding from this research is that statistical analysis methods can offer significant accuracy when analysing micro business such as neighbourhood stores and should be considered in future research and in practice. The results indicate that prediction models can be a very useful decision-making tool for neighbourhood store owners, managers, suppliers, and investors. Therefore, practitioners and policy makers will also be able to make comparisons of past and present profitability, as well as future predictions. This will allow them to have a consistent perspective in the development of projects related to neighbourhood stores as part of the small business sector.