The Location Analysis Tool is similar to a complex geospatial model known as the Huff Model. The model estimates the probability that a customer will visit one of a set of locations based on the distance to the location and the attractiveness of the store.
For example, if there are 10 stores in an area, the model predicts the probability that a customer will choose one of those locations. If seven stores are further from the customer than the maximum distance, those stores are immediately removed from evaluation. The remaining three locations are evaluated based on their distance and attractiveness.
The model defines that if a location is closer than another it is more likely to be chosen over another and if one location is more attractive than another then that will also be more likely to be chosen.
The tool offers additional control over the mathematical model. You can use these parameters to change the model based on how you would like to rate locations – either more strongly on distance or attractiveness.
‘Huff alpha’ influences how strongly attractiveness affects the model – a higher alpha will mean the lower attractiveness locations will be more strongly disadvantaged.
‘Huff beta’ influences how strongly the distance affects the model – a higher beta will mean greater distance will be more strongly disadvantaged.
Attractiveness is simply a single number, relative measure of how attractive each store compared to each other. Attractiveness may have a simple relative value of 0-10 or be a real-world value (such as the square footage of all the stores). It is important that all locations – existing, competitor and proposed – are scored on the same scale to achieve correct results. If a location attractiveness is not specified then it is given an attractiveness value of one. When new Proposed Locations are added to a scenario, by default those locations will receive the average of all other locations if an attractiveness metric is not explicitly provided.