Advances in methods to support store location and design decisions

PhD ceremony: Mr. A. Hunneman, 14.45 uur, Academiegebouw, Broerstraat 5, Groningen

Title: Advances in methods to support store location and design decisions

Promotor(s): prof. T.H.A. Bijmolt

Faculty: Economics and Business

This dissertation develops three models to improve evaluations of existing store locations and help determine where to locate new retail stores. The proposed quantitative models assist retailers in deciding whether to invest in a particular location, because they can forecast the potential contributions of proposed locations to firm performance. Thus, this dissertation responds to recent appeals to improve the accountability of marketing activities.

The proposed models advance current methods for store location decisions in several ways. First, they account for spatial heterogeneity in consumer characteristics, even if they cannot be observed by the researcher, by using disaggregate data at the zip code level. Second, they exploit the fact that consumers living in close proximity often behave similarly when it comes to shopping. Third, the models decompose the effects of each predictor variable of store sales into separate components, offering richer insights than a single model of just sales. For example, travel distance to the store has a negative impact on loyalty card penetration and the average number of store visits, whereas average expenditures are higher for consumers living farther from the store. The results of three empirical studies thus show that the enhanced model specifications lead to better predictions, which underlines their usefulness for store location and evaluation decisions.

The proposed models support location decisions in marketing practice in general as well, because they can be adjusted to the idiosyncrasies of a particular market, and the necessary data are collected on a regular basis in many industries today.