A Multiple-Attribute Decision Model for Retail Store Location

Abstract

(ProQuest: ... denotes formulae omitted.)For a retail store planning expansion, site selection is critical. A poorly-sited location represents lost capital, a drain on profits, and potential harm to the company's reputation. Moreover, for the past several decades, the range of choices available to a retailer has continued to expand, stretching the retailer's resources available for site selection. According to the U.S. Census Bureau (2010), between 1986 and 2009 (the most recent data published), the number of shopping centers increased every year across all sizes of shopping centers, despite potentially adverse changes in the economy, demographics, and competition during the period. As in most complex decisions, site selection involves tradeoffs that reflect the decision maker's preferences. Advances in decision analytics aid the solution of such complex problems. This article describes the development of a multiple-attribute analytic model used by a chain of retail stores to assist in the selection of new store locations.The theory underlying multiple- attribute decision models was developed in the 1960s, 1970s and 1980s, as summarized in Keeney and Raiffa (1976), Zeleny (1982), Dyer (1992), and Zanjirani et al. (2010) with numerous immediate applications, such as Keeney (1973b). Multiattribute models are particularly applicable in decision situations in which no single objective, such as profit maximization, exists. Stimson (1969), for example, developed a multiattribute model for decision making in a public health facility. Zanjirani et al. (2010) specifically looked at multiattribute location models.The models are also applicable in cases with a natural objective, but the alternatives being evaluated cannot be expressed in terms of that objective in any practical manner. The scoring model by Lucas and Moore (1976) is an example of such an application. Dyer et al. (1992) noted that scoring models received much attention in Eastern Bloc countries because of their suitability to central planning. Their applicability to large government projects led pioneers in the field such as Keeney and von Winterfeldt to studies such as the disposal of nuclear waste (1994). Huber (1974a, 1974b) and Dreyer (1974) reviewed early studies in these areas.Saaty, in a series of books and articles from 1980 to the present (e.g., 1980, 2008), developed and refined the Analytic Hierarchy Process (AHP) and the Analytic Network Process (ANP) methodologies for analyzing multi-attribute problems. Unlike most multiple-criteria decision models, which assume a single decision maker, ANP also works well in group decision making. The citations for applications of AHP and ANP from 1980 to the present are too numerous to mention. A good source is the Proceedings of the biannual meeting of International Symposium on the Analytic Hierarchy Process.Another application, the one considered in this article, involves the case in which a single natural objective exists, but other considerations such as the attitude of the decision maker, the nature of the decision process, or higher goals preclude its use as a decision variable in the evaluation of a particular project. To a retailer, site selection is critical and complex and is an important part of the firm's overall strategic policy. Because alternative sites may each excel on different dimensions, multi-attribute decision modeling provides a means of assessing and quantifying the decision maker's preferences.In this study, the attitude of the decision maker (DM) was important because implicitly he already had a set of attributes upon which he based his decisions (i.e., his intuition or gut feeling). Furthermore, the nature of the decision process often demanded that he make quick decisions based upon the immediate available information. For an excellent review of the relationship between intuition and analytical decision making, see Dane and Pratt (2007). Because of the nature of the problem (including a single DM), the need for rapid, uncomplicated decision making and the need to reflect the intuition of the DM, the researchers chose to construct a scoring model based upon multi-attribute utility theory to assist the DM in locating his new stores. …