The Huff Model

“The Huff Model is an established theory in spatial analysis. It is based on the principle that the probability of a given consumer visiting and purchasing at a given site is a function of the distance to that site, its attractiveness, and the distance and attractiveness of competing sites.”

To put it in simpler terms, the further away customers are from your store, and the less attractive your store is, we expect it to receive fewer visits from customers. For example, let’s say you need to buy something like eggs, and you have two different stores near you that both carried eggs. One is 5 miles away and the other is 10 miles away. You most likely would go to the store that is 5 miles away rather than bypassing it and driving double the distance to go to the other store.

There are some factors that might make you drive to the 10 mile away store that we call “attractiveness.” Maybe the store that is 10 miles away has a better parking lot or is bigger. Then you might consider driving the extra miles for what you deem a more pleasant shopping experience.

The Huff Model allows you to simulate these circumstances so you can understand how a store would perform in the existing market.

The Huff Model has various applications in retail, marketing, and urban planning. It works best for use cases where a individual would want to travel for a particular service or product.

Here are some of the best use cases for the Huff Model:

1. Retail site selection: The model helps retailers identify optimal locations for new stores by considering factors such as competition, population density, transportation infrastructure, and consumer preferences.
2. Trade area analysis: The Huff Model can be used to analyze trade areas around existing retail stores or shopping centers to understand customer behavior, estimate market share and potential sales volumes.
3. Market expansion strategy: By using the model's predictions about consumer behavior in different locations, businesses can evaluate potential markets for expansion or relocation.
4. Marketing campaign optimization: Companies can utilize the Huff Model to target specific geographic regions where advertising campaigns or promotional activities will likely have higher impacts based on consumer preferences and accessibility.
5. Urban planning and policy-making: Planners can employ this model to assess how changes in land-use patterns or transportation infrastructure might affect consumer choices regarding shopping destinations within a city or region.
6. Evaluating customer loyalty programs: By incorporating data on customer preferences and travel costs into the model, businesses can assess how loyalty programs may influence consumer behavior towards preferred store locations.
7. Evaluating impact of competitors: The Huff Model enables businesses to quantify the impact of competitor store openings/closings on market shares within a given trade area.
8. E-commerce logistics optimization: Online retailers can use this model to determine optimal warehouse locations based on factors like demand patterns from different regions and delivery costs.
9. Healthcare service planning: Hospitals or clinics can utilize this model to analyze patient flows from different catchment areas when determining facility locations or expanding services into new areas.
10. Transportation planning: Transportation planners could apply this model while analyzing travel patterns related to commuting behavior, public transit usage trends between various origins-destinations pairs within a region/city.

Every tool that is run in ArcGIS Pro usually requires some form of input data and other parameters that need to be filled in. The quality of the data you put in will correlate with the quality of your results. The Huff Model has quite a few parameters so let's breakdown what some of them mean:

Input Facility Features: The current locations in the market you want to compare your candidate feature to. This could be other existing facilities you have, or competitor locations.

Input Candidate Features: The location you are evaluating the current, or potential future, performance of.

Input Sale Potential Features: This is the market area you are evaluating. Are you evaluating the ZIP codes, custom trade areas, census tracts? You can always pull these boundary layers from the ArcGIS Living Atlas and use the Enrich Layer Tool to derive the Input Sales Potential Features parameter.

Sales Potential Field: This is the field in the Sales Potential Features layer that tells us what the market for your particular good/service looks like. In other words, in this ZIP code what do average furniture sales look like? How much are people spending on my particular product or in this particular industry? Don't worry if you don't have these, you can easily obtain market potential data through ArcGIS Business Analyst.

Attractiveness Variables: These are the things that bring people to your locations, and possibly your competitors. For hospitals this could be number of beds, number of services offered, number of doctors etc. Retailers can look at store size, number of products offered, employee count.

Exponents: We have two exponents to consider: one for attractiveness and one for distance. Notice the distance exponent has a negative value. This is because we assume distance has a negative impact on a customer's effort to go to a location. Attractiveness has a positive value as the more attractive the store, the more effort people will put forth and overcome the distance barrier.

Once the Huff Model is done running, we get a choropleth map of our results showing the probability of the people who reside in our Sales Potential Features visiting our location (The examples below show this in more detail). The calculation is run for each of the boundaries. Opening the attribute table for this new layer we can see it also added a field called "Total." This field shows us the forecasted total amount of sales to be generated from each boundary for our store.

Competition Analysis - Example #2

London - Example #3

The example below shows a comparison between looking at just a disposable income dataset and the results of a Huff Model Analysis. Higher disposable income areas are show in darker green postal codes. Using the swipe with the second map, we are looking at a Huff Model result for one potential new site (darker purple areas show more sales potential). We can see that we get a more accurate view of where sales could come from not only looking at disposable income, but with our current footprint of sites too.

Another tool that could be used to add more value to your Huff Model workflow, if you want to create an index from multiple values (this could be used for the sales potential surface or for the attractiveness variables),  The Composite Index tool !

This could add more value to your analysis if you want to incorporate more variables to your analysis and combine them in an overall index to get an even better idea of the attractiveness of your site.