تعامل رقابت فضایی قالب های فروشگاه های خرده فروشی : مدل سازی طرح ها و نتایج تجربی

This study builds on the spatial competitive interaction of store formats that are often considered generic retail positioning profiles. Specifically, a modeling proposal is developed that focuses on the role of the spatial coverage reached by particular store formats as a determining factor of intraurban market response. The proposal is assessed through an application to the Spanish retail context that tries to identify competitive frictions among the supermarket, the hypermarket, and the discount store. The results demonstrate the explanatory capabilities of the model and reveal interesting conclusions related to the impact of spatial growth by store formats on food shopping patterns. There is more intense competition between the discount store and the hypermarket, and the classic supermarket form maintains a balanced rivalry with both. The development of larger discount stores seems to affect the supermarket market share more severely, whereas the development of discount stores that offer parking subtracts more demand from the hypermarket.

Retailing has undergone an intense transformation during the past few decades. One consequence of this dynamism is the diversification of retail store formats characterized by diverse positioning profiles. For example, in the context of food shopping, the introduction of self-service has been followed by the expansion of the hypermarket and the superstore, the development of diverse forms of discount stores, and an emphasis on convenience stores. The retail offer has been enriched by a variety of retail forms aimed at satisfying the needs of different types of consumers in various shopping situations. This heterogeneity enables researchers to distinguish interformat competition in the rivalry among different store formats. The purpose of this paper is to contribute to the analytical methodology and the understanding of competitive interaction among store formats. Specifically, two sequential objectives are pursued: 1. To propose a market response model to assess the effect of changes in interformat retailing empirically. Our interest centers on the impact of the growth and expansion of store formats, and therefore the modeling emphasis is on the spatial dimension of retailing. Whereas previous attempts have related market share and retail format coverage statistics at an interurban or regional level, our modeling proposal focuses on capturing market response variations in an intraurban competitive scene. Furthermore, the modeling proposal embraces two methodological premises: (a) the use of standardized and comprehensible econometric tools and (b) the objective measurement of concepts that are compatible with usually available secondary data. Both aspects are critical to the adoption of quantitative models by retail managers (Simkin, 1996). 2. To assess intraurban competition patterns among three dominant formats in the food market: the classic supermarket and its larger and price-oriented versions; that is, the hypermarket and the discount store, respectively. The market coverage reached by these formats and their more specific forms produces differential effects across their competitors' market share, which should be taken into account both by retail firms in their location and diversification strategies and by public authorities in the assessment of possible retail planning policies and their economic impact. The relevance of the spatial perspective to the analysis of retail structure, growth, and evolution has been repeatedly pointed out in the marketing literature (Sparks, 1995). Spatial market coverage and location strategies constitute some of the most important and risky decisions of retail marketing management (Jones and Simmons, 1987). Along this line, research efforts to develop analytic methodologies to understand, explain, and even predict retail competitive interaction derived from location and spatial coverage strategies are especially significant (Babin et al., 1994). This paper contributes to this field of research by broaching the interformat level of competition rather than using traditional emphasis on retail chains or individual stores.

The contributions of this research consist of theoretical and empirical formalization of a model (aimed at represent- ing competitive interaction among store formats) that is derived from spatial coverage and location patterns in intraurban markets. Specifically, the market share of store formats across intraurban residential zones and time periods is interpreted as a parameterized function of their spatial availability and accessibility. The spatial facet in the analysis of interformat competi- tion is relevant to retail firms in their growth strategies. The development of new store formats threatens the performance of existing ones. Only by anticipating potential consequen- ces is it possible to develop marketing actions that enhance the viability of a retail business in a changing environment. Moreover, because format diversification represents a com- mon growth strategy (Pellegrini, 1994) , an understanding of interformat spatial competitive interaction facilitates the optimization of the attraction and retention of customers while minimizing cannibalization effects between store formats. The spatial analysis of interformat competition is also relevant to local authorities in their retail planning, regula- tory, and controlling competence. Moreover, many legisla- tion initiatives to regulate retail development establish prerequisites and conditions regarding new locations for particular store formats, such as large-scale retailers or discount stores, and their impact on the competitive struc- ture (for examples, see Cliquet, 2000; Davies and Itoh, 2001 ). The empirical implementation of this model in the metropolitan area of Madrid has focused on analyzing the relationship between spatial coverage and the market share of three basic store formats—the supermarket, hypermarket, and discount store—using individual food budget allocation data. Although the results show only a moderate fit, prob- ably due to the incomplete nature of the model and the difficulty of quantifying the explanatory constructs, they provide the means to make interesting conclusions about the role of location for market response. In particular, compet- itive effects derived from location are not symmetric but differ across competing store formats. The results not only suggest the direction of these competitive frictions, but also indicate a quantifiable measure of their intensity. In the Madrid market, most of the discount stores in the intraurban scene are small discount stores, and the results concurrently indicate the greatest rivalry between discount stores and hypermarkets, which are oriented, respectively, toward money and time saving. There seems to be more overlap between the target segments of these versions of supermarket, whereas the conventional super- market competes equally against both. The development of medium-size discount stores seems to capture more market share from the supermarket than from the hyper- market, whereas larger discount stores with parking takemore market share from the hypermarket than from the supermarket. Several limitations of this research should be noted. It is first important to insist on the model’s incompleteness. It focuses exclusively on market coverage and obviates other relevant positioning dimensions whose roles may be im- plicitly considered as intrinsic to the attraction construct. It also ignores market spatial heterogeneity as a determinant of market response. However, these explanatory factors can be included in the model even while the same theoretical framework and a similar econometric formulation are main- tained. The role of additional determinant attributes easily can be contemplated in the model by splitting the intrinsic attraction into the effects of such variables. Heterogeneity can be explicitly considered by a priori segmentation, by including the consumers’ characteristics as covariates, or even by adopting more sophisticated and data-demanding methods such as random effects choice models (Chintagunta et al., 1991) . In this sense, the proposed model has great potential to evolve from simplicity to completeness, a key criterion for successful implementation of marketing models (Leeflang et al., 2000) . In addition, the operationalization of variables in this model involves some difficulties that derive from the translation of theoretical concepts into proxy variables. Although spatial accessibility and coverage are highly intuitive notions, they are not easy to define in practice. In particular, the proposal focuses on coverage of residential zones and obviates coverage of crowded zones that consti- tute a key issue to capture the effect of multipurpose and multidestination trips. Some additional limitations derive from data quantity and accuracy, including a generic defi- nition of store formats, poor spatial precision, and the omission of spatial barriers and trip routes in measures of accessibility. Although the proposed model uses objective measurements of the variables through available secondary data, subjective measurements of perceived accessibility through surveys could improve its explanatory possibilities. As a side note, secondary data derived from consumer panels, store census, or geographic information systems are continuously evolving toward improved quality and precision. Although the empirical results are probably specific to the studied retail context, the model’s emphasis on stan- dardized econometric tools and objective variable measure- ment makes it easily replicable in other retail intraurban contexts. Nevertheless, inferences about future contingen- cies should be made with caution. First, the model only proves relationships between variables, not causality, so causal inferences must rely on the theoretical background of the model. Second, significant relationships between variables are only reliable in the range of values considered in the estimation process so the predicted situation in conditions different from the current ones might not be reliable. Data from experiments with simulated retail sce- narios could provide interesting insights into this problemIn addition to the areas for improvement evoked by these limitations, two more research lines might build on the contribution of this study. First, it would be interesting to go deeper into the attraction dimensions that undergone asymmetric competitive effects among store formats. Pre- sumably, these effects are related to the overlapping of benefits supplied: accessibility, price, quality, variety, serv- ices, etc. Second, it would be interesting to explore the simultaneous modeling of competitive interaction at the inter- and intraformat levels of competition, namely, the integration of interformat competition into the whole retail structure and competitive interaction. The concept of store format plays a determining role in retail competition in that it agglutinates most relevant differences on image attrib- utes so its explicit consideration in modeling approaches may improve the explanatory possibilities but avoid ex- cessive parameterization