This blog takes an informal look into the debates and methods related to business GIS and mapping
Author: Dr. Murray Rice
This is the second of three posts that explore the perspectives that retail geographers have developed on how to identify the best locations for a business. The previous post introduced some important context for understanding the changing role of location in the operation of retail and service businesses. The first post in this mini-series also introduced the most simple of methodologies that geographers and retail managers have developed to assist businesses in identifying good locations. This second post goes on to discuss two more elaborate approaches to business location selection: use of analogs, and regression modelling.
So far we have seen two approaches to site selection that work by leveraging the massive investment that businesses have made in locating their operations to date. The basic perspective advanced by both rules of thumb and ranking is that insides from the previous operations of the business can be used to optimize operations planned for the future. The means of encoding previous business operations in both cases is on the simple slide, but the basic idea as much intuitive appeal. A prerequisite for using these two methods is some kind of experience on the part of business management or ownership, as the results of rules of thumb and ranking can be no better than the information at hand. In other words, if a site is to be judged based on five criteria, then the five criteria better be based on solid business intelligence.
The next business location perspective we will examine here shares this focus on past results with the first two methods. However, analogs switch the focus from an internal perspective on the businesses own decisions to an emphasis on the market response. The key insight here is that the analog approach hypothesizes the past financial performance of the business can be used to provide a prediction of future performance. However, the analog approach calls for care in setting up comparisons used for such predictions.
The key here is for the business to identify existing locations that are highly similar to new locations under consideration. Thus, the comparison between previous operations and future potential operations is a highly guided and information intensive venture.
Typically, the analog approach begins by identifying market conditions that characterize the businesses operations at a series of proposed locations. Characterizing the new location possibility accurately is crucial at this stage. Multiple levels of market characteristics enter into the thinking here. Highly macro location characteristics, such as the size of the city, its economic growth track record and future prospects, and characteristics of the region that hosts the site are important variables in this consideration. For example, the business would categorize a potential new location opportunity in Bismarck, North Dakota by dimensions such as
Further adjustments to this analog comparison need to come from examination of more micro-location characteristics associated with the potential location. For example, how accessible is the location (is it at a major street corner, or in the middle of a long block), is it a standalone business opportunity (such as a street front parcel) or is it part of a planned development (such as a big-box or mall development)?
In total, a comprehensive list of market and location conditions need to be collected for every proposed site (Figure 1).
Figure 1: Market Characteristics Assembly for a Proposed Location
Then, the business would seek out analogs from its current portfolio of locations that most closely match this particular macro and micro profile. The idea here is for the business to identify the locations from their current portfolio that best replicate the conditions at each proposed site.
Figure 2: Seeking a Match from Existing Locations
Once these existing and proposed location pairings are established, the business can then connect known performance outcomes at current sites with predicted outcomes that may be possible to achieve at the proposed sites.
Figure 3: Model Projected Results using the Closest Matching Analog
In total, the analog approach involves:
Like the first three site selection approaches that we have surveyed, regression modelling involves leveraging the value of the business’ knowledge of its locations and markets, as built up over the life of the business. Regression modelling differs from first three approaches by using an advanced multivariate statistical methodology that provides a sophisticated way of combining all of the insights into a coherent overall package. Regression modelling focuses on placing the key site factors identified by the business in a quantitative model which provides a direct prediction of financial results associated with the specific site.
For example, Figure 4 provides a sample framework for regression modelling that identifies variables for regression analysis in the case of a restaurant chain. The figure identifies the variables included for consideration in the regression model, including a mix of demand and supply variables needed to be considered.
Figure 4: Identifying the Variables for Analysis
Note, I’m not going to even try to do justice to the work needed in step 2. Please refer to any of a number of sources that provide an excellent overview of the process involved in multiple regression modelling. Overall, regression modelling is a topic that goes well beyond a blog post. However, a couple of notes will help to define the value of this approach:
For more insight into the analogs and regression analysis described here, please refer to the site selection content in my new applied handbook Pathways to Actionable Results with Business GIS
Home | Products | Contact | Secure Store
Call Us 
