Analyzing Retail Trade Areas Using the Huff Model

Introduction

The concept of a retail trade area has long been used by analysts and practitioners in retail site evaluation and market analysis. Trade area analysis focuses on identifying and describing a store’s target market and is essential for marketing, merchandising, and selecting new retail locations. In site evaluation, trade area analysis is combined with operational requirements of retail chains to support informed decision-making.

The use of GIS has significantly improved trade area analysis by enabling efficient data extraction and aggregation across multiple geographic levels. As a result, trade area analysis has become one of the most widely applied GIS techniques in business analytics. A commonly accepted definition describes a retail trade area as “the area, typically around the store, from which the store derives most of its patronage.”

Historically, a wide range of techniques have been used to delineate trade areas, from simple concentric rings to more advanced probabilistic surfaces. These techniques generally follow either a spatial monopoly approach or a market penetration approach.

Spatial Monopoly vs. Market Penetration Approaches

Traditional spatial monopoly methods—such as concentric rings, drive-time polygons, or Thiessen polygons—are simple and easy to implement. However, they assume that a store fully dominates its trade area and ignore competition. This oversimplification limits their usefulness in competitive retail environments.

In contrast, the market penetration approach accounts for spatial variation in customer behavior caused by competition. The most widely used method in this category is the Huff Trade Area Model, which represents a trade area as a probability surface rather than a fixed boundary.

Introduction to the Huff Model

The Huff Model, introduced by David Huff in 1963, estimates the probability that a consumer located at a given origin will choose to shop at a particular store, considering the presence of competing stores. The model incorporates:

• Store attractiveness
• Distance between consumer and store
• Competition from other stores

The result is a customer patronage probability surface, which can be contoured into probability regions and used as weighting factors for market profiling and revenue estimation.

In the first example, trade area analysis was conducted for a single shopping mall using Gross Leasable Area (GLA) as the sole attractiveness variable. A probability surface was generated assuming that potential customers are distributed across grid cells and choose stores based on attractiveness and distance.

The probability surface was converted into probability regions using contour lines. Census data were then aggregated for each region to create an unweighted market profile. To make the results more realistic, probability values were used as weights, producing a weighted market profile that better reflects actual customer behavior.

The second example expanded the model by incorporating multiple attractiveness variables, including GLA, number of stores, and parking capacity. These variables were standardized and combined into a composite attractiveness index.

The resulting probability surface showed increased drawing power for the analyzed mall, demonstrating how multiple variables can better explain variations in customer patronage. Weighted market profiles revealed a significantly larger estimated customer base compared to the single-variable model.

This example applied the Huff Model to estimate and compare potential revenue for two competing shopping centers. Consumer spending potential was rasterized for multiple retail categories and weighted by patronage probabilities for each center.

The aggregation of weighted spending grids revealed that one mall slightly outperformed the other, despite similar attractiveness characteristics—highlighting the value of probability-based analysis over simple trade area boundaries.

The final example demonstrated a more advanced application of the Huff Model using customer spotting data and model calibration. This approach allows for:

• Parameter estimation
• Market share analysis
• Scenario modeling (e.g., adding or closing stores)

After calibration, the model was used to simulate the impact of introducing a new store, showing how patronage and market share redistributed among existing locations.

Final Remarks

The Huff Model remains one of the most powerful tools for retail trade area analysis due to its ability to account for competition and spatial variation in consumer behavior. While not all GIS software supports parameter estimation and model calibration, recent integrations into commercial GIS platforms represent significant progress.

Beyond retail, the Huff Model has applications in banking, healthcare, public services, and other sectors where spatial choice and competition play a critical role. Its continued adoption reflects its superiority over traditional trade area methods and its relevance in modern spatial analytics.