GIS Methods to Estimate Market Areas

The above figure represents different methods that can be applied in a GIS to estimate market areas:

• Buffer creation, a common GIS procedure, associates each concentric circle with a distance (or a time value). They can include the threshold and the range of a store. On the above figure three concentric circles have been created with 5 minutes distance increment each. 5 minutes is assumed to be the threshold of this activity, while the range is estimated to be 10 minutes. Another dimension of this method concerns Thiessen polygons where the market area is calculated as halfway distances from the location of other competitors.
• Share by polygon can be estimated in many ways. For instance, it can be an aggregation of individual customers within a geographical unit of reference (ZIP code, census bloc, etc.) or a statistical calculation based on a set of representative variables, such as distance, population, income and age. On the above figure a market share was calculated for each unit and then classified according to a graduated color expressing the level of membership to the store's market area.
• A star map is a vector creation where segments have different origins (one for each customer) but the destination is the same (store). It depicts a market area as a set of customers connected to a store. Qualitative and quantitative attributes can be attached to each vector.
• Spatial smoothing is the outcome of statistical modeling that interpolates data from a known set of points (customers) to a continuous surface. The density of customers thus becomes a statistical surface expressing the market area. On the above figure, spatial smoothing created a statistical surface expressing three levels of membership to the store's market area (high, average and low).
• Transport distance measures the accessibility, expressed in distance or time, of road segments to the store. It takes into consideration transport distance, often quite different from Euclidean distance, as well as the different capacities of road segments (number of lanes, driving speed, turn penalties, etc.). A new layer is created where each former road vector is segmented according to distance / time decay through a routing procedure that originates from the store. On the above figure, roads are segmented according to 5 minutes driving time increments from the store location.
• Manual polygons are created with tracing where the analyst evaluates a market area from a set of assumptions, often based on specific expertise and empirical knowledge about that market. For instance, the analyst may empirically know that for various reasons few customers may be coming from a nearby neighborhood, excluding it from the market area. It may also been known that few customers are coming from further away than a specific street, making that street a boundary for the market area.