Capability of huff model to predict market share

Abstract

Gravity model is widely used in the field of transportation planning, trade and allocation of public facilities. Originating from the famous Newton’s Gravitational Law, gravity model has been modified and adapted to various purposes and applications in many fields. However, there are too many varieties of variable being used, especially various types of distance and non-physical variables, putting the viability of the model to explain an ongoing phenomenon and to predict future conditions into question. Application of the gravity model also requires extensive socioeconomic data which, in some countries, are difficult to obtain. Thus, can we use a basic gravity model that only uses the size and distance variables to explain a phenomenon? For that reason, this research compares the accuracy of four (4) modified Huff Gravity Model variants using only i) hypermarket floor area data, and ii) either Euclidean distance or actual distance to predict market share of three selected hypermarkets. The predicted results were then compared to the actual market share of each hypermarket, obtained through customer survey. Results show that the variant that uses Euclidean distance with parameter; power of two (2) function come out with the most accurate market share explanation. The modified Huff gravity model generated by using spatial analyst capability in geographical information system (GIS) further shows the gravitation surface of each hypermarket, and even successfully helped to explain the spatial interaction between each competing hypermarket and their market share - proving the viability of the gravity model to explain existing interaction phenomena and forecast future interaction pattern.