Investigating the Combined Effect of Store Flyers Promotions on Brand Sales: A Novel Approach Based on Regression Trees

Flyers (i.e., printed means of weekly/monthly communication) constitute one of the most relevant forms of advertising for retailers. However, extant retailing research has only examined a limited number of promotion types advertised through store flyers, finding ambivalent results and neglecting to consider their combined effect on sales. Using machine learning methods that account for data dimensionality – that is, Regression Trees (RT) and stepwise selection – this paper aims to identify the combined effect of different promotion types and account for the negative impact of competitors’ promotions. The results show that promotions ensuring a dedicated section in the flyer (i.e., box branded) have a greater impact on brand sales. Moreover, when a main competitor pursues this promotion type, sales of the examined brand tend to decrease, thus providing empirical evidence for the phenomenon of competitive advertising in store flyers. The paper concludes with a discussion of its methodological and practical implications for retailers and scholars, as well as avenues for future research.