Skip to main content

Most geomarketing researches are based on a basic Huff model. However, its usage is sometimes problematic. Read more on how  Geomatrix developers managed to overcome the most frequent difficulties and significantly improve the efficiency of working with the model.

  • Basic Huff Model, 1964

It states that the patronage probability Pij of a store j for inhabitants in a given point of origin i is equal to the proportional ratio between the size of this store (Sj) with the distance between a given region i and a given store j (Tij), compared to the total size generated by all N stores in the neighborhood of this area.

  • Model issues and Improvements by Geomatrix

Size of shopping centers / stores

Issue:The basic Huff models is limited by the definition of the size of the shopping center / store.

Solution:We have resolved this issue in proposing a dual approach to this problem, providing the choice to use either a size element (sales area, GBA, GLA), either an aggregated attractivity parameter to be define for each store.

This attractivity parameter for each store or for each brand is defined directly by the system’s users, and can be modified independently by the user, in order to test multiple scenario.

Selection of shopping centers / stores

Issue:The basic Huff model is limited by the taking in account of all shopping centers / stores within a specific accessibility area, non-regarding of their size / attractivity parameters.

Solution:We have assigned in the system a filter tool, so that the user can independently select only the shopping centers / stores fitting with its analysis.

Ignorance of consumer purchasing behavior

Issue:The basic Huff model is limited only to the calculation of a probability ratio for each customer in a point of origin.

Solution:We have developed and integrated into the system all our syndicated data, and client’s own analogical consumer purchasing behavior, in order to calculate automatically the potential client traffic and turnover in less than one minute.

Ignorance of consumer store selection behavior

Issue:The basic Huff model is limited by the fact that within a specific accessibility area, every shopping center / store will be considered as a potential competitor for all potential consumers of this very accessibility area. For instance, if in a specific area, we find 500 competition grocery stores, the probability of visiting one of these stores will be always superior to zero, meaning that a consumer will spend its grocery budget between 501 stores, which is perfectly impossible.

Solution: we allow the system’s user to choose the number of stores an average consumer will regularly spend in during a specific period of time (usually between 2 and 6 destinations), and reallocate the demand for each consumer between the best 2 – 6 stores.

Computation Time

Issue: when performing basic Huff model on desktop GIS, the calculation time and creation of one (1) single scenario – analytical report for an average shopping center / large store takes between half a day to one day, depending to company’ GIS analyst proficiency, non-regarding with the availability and quality of required socio-demographic and competition data, the availability and quality of road network, the availability and quality of consumer’s behavior.

And this time cost shall be multiplied for each new scenario. Hence, for some projects with necessity to determinate optimal commercial format for a specific location, taking in account different scenarii of competition, different scenarii of traffic congestion by day and by hour, and different commercial conceptions will take several weeks to be performed.

Solution:our system allows to perform every scenario in less than one (1) minute online, decreasing tremendously the cost and resources required.

  • Benefits

Result Description: Basic Huff Model

Result Description: Geomatrix Huff Model

Benefits

Through the integration of our algorithms, we’ve managed to enhance significantly the Huff model:

  • Technically: reduce the reporting production from 1 day offline to 1 minute online
  • Efficiency:we’ve made the model more realistic, taking into account final consumer behavior, as well as company internal analogical data
  • Resources:no longer requiring highly skilled GIS specialist to create / acquire data, and realize Huff model, any employee with 20 minutes training can perform it
  • Risk management: as the system is available to any non-GIS specialist, the control on report can be performed at any management level, different scenario can be performed in minutes, in order to take the best business decisions and minimize investments risks