The impact of travel time savings on shopping location choice or how far do people go to shop cheaply

Abstract

The determinants of store choice have always been a key topic in the marketing and especially retailing literature. It is clear that not only the consumer’s perception of directly store related criteria like pricing, quality or market positioning strategy but also the store location itself affects store choice substantially. Different studies demonstrated this in the past but were not focussing on the implied willingness to pay for travel time reductions on the way to the store. In this work two independent shopping location choice stated preference surveys are brought together to estimate a joint shopping location choice model. One was conducted in 2004, the year of the market entry (2004) of the German discounters Aldi and Lidl in Switzerland, for which customers were interviewed in the tri-national Basle region in three supermarkets, one in Switzerland, Germany and France each. The other survey was conducted in Gottingen, Germany 2006. Both surveys focus on the weekly food shopping trip which is very common in both countries and are composed of two parts: In the first part people were asked about their shopping behaviour in general. In the German case, the stated preference questionnaire was handed and filled out directly in the store. In the Swiss case the respondents filled it out at home and returned by mail. Overall 2447 (1608 in Gottingen, 839 in the region of Basle) decisions were obtained and are the basis for the estimations. The descriptive analysis of the data shows large differences in the valuation of the different variables trough the respondents, especially concerning distance and price. Therefore the model includes socio-demographic variables in non-linear functions as well as different interaction terms. Focussing on the willingness to pay for travel time reductions, the impact of income, distance and other variables are examined and identified. As these values diverge strongly between the different study regions and persons a qualitative assessment of the difference leads to the recommendation of possible improvements in modeling using different kinds of market segmentations like latent class models. Furthermore the hypothesis of Louviere et al. (2001) of preference homogeneity between countries cannot be confirmed.