The role of technology in assessment of shopping center location.

Vouk, Rudolf ; Simurina, Jurica ; Markovic, Milivoj 等

1. INTRODUCTION

The overview of body of literature concerning the retail location reveals the vast area of study that scientists have tried to make clear. First models that tried to explain the rationale behind choosing one retail location over another have emerged in the 1920s and were mainly mathematical models that in a certain way integrated key demand variables relevant for success of a retail store or a shopping centre. These mathematical models gave a single indicator of relative preference of one retail location over another. They mostly emerged during the first half of the twentieth century. Later models have brought various extensions of the existing models without highlighting the model that would substantively outperform other models. This lasted until the 1990s when technological advances enabled new approaches for assessing retail location potential. Specifically, the use of GIS (Geographic Information Systems) greatly contributed to the new directions in retail location assessment. GIS provided the opportunity to overlay spatial and non-spatial data on a single map and present the decision maker with a simplified overview of a real life situation.

2. SHOPPING CENTER LOCATION IN HISTORICAL PERSPECTIVE

By definition, shopping centre is a group of retail and other commercial establishments, planned, developed and managed as a single property under single ownership (Levy & Weitz, 2007). First modern shopping centre (conveying to former definition) was Country Club Plaza build in 1922 in Kansas City, United States. At that time the main venues of retail activities were the city centres. However, due to the gradual traffic congestion in city centres and relocation of population from city centres (following massive road development and availability of cars, especially during and after 1950s) suburban shopping centres started to emerge as the dominant place for retail activities (Cohen, 2002). Since free space was in abundance, and competition was very low, a question of shopping centre location was merely a question of proximity to the suburban population. Therefore, regional shopping centres were built a bit farther away from the suburb due to their vast space usage and the fact that they attracted shoppers from 50 or more miles away. On the other hand, neighbourhood shopping centres and convenience shopping centres were built closer to suburban population. During the 1970s shopping centre expansion in the United States was close to its peak and a good location for shopping centre was harder to find. Good assessment for new location played increasingly important role in shopping centre success. During the 1980s and 1990s shopping centre developers started to invest more in renovation of existing centres along with building super-regional centres encompassing more than 4 million sq. ft. (Mall of America).

3. SHOPPING CENTER LOCATION MODELS

Main models for assessing retail location date from 1920s and 1930s. These models were mainly developed in two different directions. One direction focused on subjective analysis by various observation approaches and could hardly be called scientific. On the other hand, alternative approaches used quantitative analysis in an effort to rigorously assess the potential of one retail location relative to the other. These models were called gravity models due to their analogy with Newton's law of gravity.

One of the first and most famous of the models is the Reilly's Law of Retail Gravitation. In 1931 William J. Reilly created a model that could be used to delineate a boundary between two cities trade areas using the distance between the two cities and the size of their populations. The theory stated that if two cities were of equal size than the trade area braking point (breaking point marks a line where 50% of residents buy in one city, and other 50% in other city) would be exactly halfway between the two cities. Thus, if one city is greater than the other, the breaking point would lay closer to the smaller city (Segetlija, 2006). The formula calculating the breaking point is:

[B.sub.b] = [D.sub.a,b]/1 + [square root of [P.sub.a]/[P.sub.b]] (1)

The formula calculates the breaking point between the cities a and b as a distance between cities divided by 1 plus square root of population of the city a divided by the population of city b. This formula presumes that all factors, e.g. as consumer preferences, rivers, mountains, political factors and others, are virtually nonexistent, i.e. they have the same effect on booth cities that cancel out. In spite their wide applicability, the formula was not without flaws.

Huff (1963) introduced a new formula with a goal of delineating specifically a shopping centre trade area. Huff's model included two variables. One was the variety of products that shopping centre offers and the other was the time it takes for the consumer to reach the shopping centre. His model relates these variables in the following way:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

P([C.sub.ij]) stands for the probability that the consumer in the location i visits shopping centre j; [S.sub.j] stands for the square footage of shopping area for certain product category in the shopping centre j; [T.sub.ij] stands for the time needed for a consumer to travel from i to shopping centre j and [gamma] marks an empirical parameter that encompasses the influence of time spent on trip with various types of purchases.

Later concepts have tried in certain way to amend for the rigidities imposed by previous models and more precisely explain the influence of location for shopping center success. However, these models did not deviate much from the concepts of gravity models explained earlier. New approaches did not gain much popularity until the 1990s and the use of GIS systems.

4. INFLUENCE OF GIS ON SHOPPING CENTER LOCATION ASSESSMENT

Geographic information systems can be defined as "a computer based system that provides four sets of capabilities to handle geo-referenced data: input; data management; manipulation and analysis; and output" (Aronoff, 1989.). Depending on the project in question various approaches are feasible. Digital map of the area delineating potential mall location is necessary. Size of the area it represents depends on the particular case, however, target market and potential competitors greatly determine that size.

Secondly, decision maker must determine which variable it wants to include in output. Besides the roads one should probably include tabular data such as household income, frequency of purchasing trips, average amount spent etc. If the shopping centre in question is neighbourhood or convenience centre, located in the urban area, fig.1 could easily represent various layers included in the final output. It is the sole judgment of the decision maker to include other or exclude existing layers as to his assessment of the informative power of each layer.

What makes GIS so appealing is its ability to present various spatial and non-spatial data. Yang (2002) demonstrated how GIS can help in determining RQS (Retail Supply Quality) index for the entire town. Inputting shopping centre data and population data the author calculated RQS for different parts of town clearly indicating on a 3D output map which parts of the town lack shopping centre space and which have it in abundance.

Besides the natural objects (rivers, mountains), transport routs, city boundaries, it can include into its output variables such as demographic data, existing shopping centre locations, delineation of primary market area etc. Furthermore, demographic data can show, beside the location of the consumers, their household income, as well as relative market share of existing shopping centres etc. (Cheng, Li, & Ling, 2007).

Type of analysis previously described can be called external analysis since it includes general and industry environment of shopping centres.

[FIGURE 1 OMITTED]

Extension of GIS for analysing shopping centres is further possible in tracking consumer flow inside a centre. GIS would simply map each retail establishment inside the centre and then, based on consumer data flow, spot which stores have the highest customer flow. It can further include turnover data and profitability data for each store into the output and then simulate new store opening or expansion of an existing store with consequences for the remaining stores. This can easily be done for each floor of the shopping centre.

Furthermore, shopping centre developers and managers can use GIS to keep track of historical evolution of the tenant mix. Consumers often patronize a shopping centre because of the key retailer within a shopping centre (usually a department store) and GIS can keep track of the customer flow key retailers have in each shopping centre. Also historical data can be a good indicator of shopping centre (re)development potential and even could aid in the decision to close a mall (Jones, Pearce, & Biasiotto, 1995).

5. CONCLUSION

In this article, we provide a brief overview of the evolution of models for deciding on shopping centre location and further elaborate how contemporary advances in technology influence this process. First models dating more than 50 year ago had simple assumptions and were easy to use; hence the results were often not accurate. Later models contributed to accuracy of assessment but at the same time their simplicity diminished the use for managers.

During the 1990s technological advances equipped decision makers with an ability to summarize real life data and present them in a unified graphical output. GIS systems, adding more useful features every day, provide managers with tools that have mathematical precision of earlier models but user friendly graphical interface that visually display important data in an easy and comprehensive way.

Usefulness and popularity of GIS systems for shopping centre developers as well as shopping centre managers is predicted to increase due to ability to tackle more problems. However further research is needed to provide new insight not just with matters concerning location issues, but also with problems of store layout within the centre, consumer traffic and consumer response to various management initiatives.

6. REFERENCES

Aronoff, S. (1989). Geographic Information Systems: A Management Perspective. Ottawa: WDL Publications

Cheng, E. W.; Li, H., & Ling, Y. (2007). A GIS approach to shopping mall location selection. Building and Environment (42), pp. 884-892

Cohen, N. (2002). America's Marketplace: The history of shopping centres. New York: International Council of Shopping Centres

Huff, D. L. (1963, Februar). A Probabilistic Analysis of Shopping Centar Trade Areas. Land Economics, 39 (1), pp. 81-90

Jones, K.; Pearce, M. & Biasiotto, M. (1995). The Management and Evaluation of Shopping Center Mall Dynamics and Competitive Positioning Using a GIS Technology. Journal of Shopping Centter Research, 2 (1), pp. 49-82

Levy, M. & Weitz, B. (2007). Retailing Management. New York: McGraw Hill

(n.d.). 2008. Mall of America, Available from: http://www.mallofamerica.com/about_moa_facts.aspx Accessed: 2008-06-22

Segetlija, Z. (2006). Trgovinsko poslovanje (Trade business). Osijek: Ekonomski fakultet u Osijeku

Yang, Z. (2002). Microanalysis of Shopping Center Location in Terms of Retail Supply Quality and Environmental Impact. Journal of Urban Planning and Development, 128 (3)