Strategic Retail Location

Trade Areas

The notion of a Trade Area is appealing and popular.  It is easily visualized as a circle of say 5 km radius around a store—the exact radius is based on experience or analysis.  Given appropriate demographic data, GIS can easily calculate the size of the market, or profile the socio-economic composition of the area.  This is a quick and effective solution, at least superficially.

The Trade Area was a useful concept in the days when market analysis was limited to rough estimates and mental arithmetic.  But in a computing context, where we compare locations based on quantitative performance estimates, a number of difficult questions arise.

  • It assumes that people travel in straight lines. This conceptual error alone overestimates the market by nearly 60%. It's possible to get around this by calibration.
  • Constructing a circle, we imply that all people residing within that radius are our customers, and that nobody outside that radius shops at our store.  This is of course unrealistic.  It ignores traffic and the effect of nearby attractions.
  • It assumes away the competition.
  • It assumes that demand is evenly distributed. The truth is that residential density varies considerably.

Some analysts try to skirt these problems by counting only a portion of the population within the circle, or considering non-circular trade areas (“amoebas”), or primary and secondary trade areas, and a special derivative, the Thiessen polygon. Arbitrary ratios are applied to account for competition and distant traffic.

All these attempts are subject to the ultimate criticism that a trade area is in reality fuzzy and dynamic; it is a construct in the mind of the retail analyst, not the shopper. Trying to draw a boundary around it is artificial and arbitrary. The treatment of competitive effects is entirely inadequate.

In today's mobile society, shopping behaviour and store choice are subjective choice processes that depend on a range of factors such as advertising, ambience, complementary shopping opportunities and daytime traffic, as much as on proximity to one's place of work or residence. Dot density maps (right) or ideally customer spotting are far better suited to understanding retail behavior. A class of models known as Gravity Models (more properly, Spatial Interaction Models) is used to forecast attendance.

Back: Retail Site Selection Models

© Digital Geographic Research Corporation Contact us