Estimating Retail Sales with GIS & the Huff Gravity Model (Sample)
If you are planning a new store or shopping mall/retail development, estimating how much in retail sales your project may make in its first years of operation is essential.
If you may need help estimating retail sales for your project in Korea (or elsewhere) I am able to use GIS in order to run Huff Gravity Model, which will provide baseline retail sales figures, which can then be further broken down into sales by specific retail category (e.g: apparel, accessories, footwear, F&B, etc).
SAMPLE Case:
STEP 1: The first step in running the Huff Model is ensuring you have supply/competitor data.
Sample competitor map for a sample/planned 30,000 sm retail development in Geomdan New City, marked by the red/white star. Competitor data is blurred.
Size is a common metric to run a Huff Model on, but if sales data from competitors may be available the model will be more accurate; sales are more of an accurate indicator to as how "attractive" a specific competitor is.
STEP 2: The second step is ensuring you have demand data - retail spending (number of households and spending per household).
Sample map showing total market spending (2020) for each neighborhood. KOSIS, the country's census bureau was utilized for household/population and spending data. Spending data was aggregated for DSTM (department store type merchandise), things that would otherwise be sold at typical retail centers/shopping malls/etc.
STEP 3: Calculate "Probability" by running the Huff Model in GIS
The map above shows the probability of residents in each neighborhood visiting the Geomdan Site given other competitors and their straight-line distances to/from the project site. As evident, areas to the south, south-east, and north-east of the Geomdan Site with high competition show lower probabilities as residents in those areas are less likely to travel further distances to visit the Geomdan Site when there are other retail destinations that serve them.
STEP 4: Aggregate sales data
Sample outputs include total retail market for each neighborhood, the probability that those residents will go to the project site, total potential retail sales, and percentages regarding retail sales.
Total retail sales for each neighborhood can be derived once probability has been calculated through the Huff Model. Estimations about retail spending by specific retail categories can then be made by assessing current trends in retail spending:
Sample outputs from deriving retail sales by category after calculating total retail sales.
STEP 5: Create trade areas based on sources of retail sales
Sample trade area map; Primary trade area (60% of total sales) is in red. Competitors are overlaid.
After aggregating sales data you can assess what neighborhoods can be considered part of the primary and secondary trade areas.
Closing Thoughts
The Huff Gravity Model is an important and useful tool for forecasting retail sales, especially when run by a GIS analyst who has a background in assessing whether the numbers "make sense" given the project's location in the market.
Outputs from the Huff Gravity Model can be mapped in an interactive way so stakeholders and decision makers may access and interact with the data. A sample of the kinds of interactive maps that I can develop can be viewed below:
