A national level retailer engaged us to conduct a study investigating the effects of proximity to other competitors on sales. The retailer identified Canadian Tire and Walmart as primary competitors. The objective of the study was to obtain business insights that could guide store openings in different provinces.
We were given access to confidential store specific sales data for several locations. These data were matched with information on local census level characteristics available from Statistics Canada. This is important to segregate the effects of local demographic and socio-economic factors from the presence of a Walmart or Canadian Tire location. Machine Learning methods were used to isolate the magnitude and statistical significance of a variety of factors that could plausibly impact trends in client store sales. A statistical model was developed with the ability to assess the effects of distance to either a Walmart or Canadian Tire retail location on store sales, controlling for other factors.
The analysis revealed that retailer sales were negatively associated with distance to Canadian Tire stores. Specifically, after adjusting for local population and income, sales in client stores in close proximity to a Canadian Tire retail outlet were much lower relative to stores that were further away. On the other hand, sales were actually higher in stores that were in close proximity to a Walmart. These results were used by the client to enhance its national expansion strategy and ensure greater success in new retail outlets. The findings demonstrate the capability of l statistical analysis to identify behavioural insights and the importance of conducting such research before the implementation of key business strategies.