Analogs
Analogs are sets of data containing the most important independent variables which relate to sales performance.  An analog dataset can be organized at the trade area level or at the map section level.  In sales forecasting for new sites, the analyst culls through the analogs to find those which are most similar in the key characteristics to the trade area or map section under examination.  Analyst judgments are integral to good sales estimates in the analog process.

BandQuads
BandQuads are unique SPR geographic units.  BandQuads focus on the key contributing regions outside the trade area from which a site will draw sales.  It is formed by creating two bands beyond the trade area, one at double the average distance used in the store’s dynamic trade area and another ten miles further out.  Each band is further divided into quadrants:  northwest, northeast, southeast and southwest.  In total each store can have eight BandQuad sectors.

Barrier
Barriers are natural or man-made features which constrain access from map sections to the subject site.  Features such as rivers with minimal cross points and a set of railroad tracks can act as barriers.  In some cases psychological barriers like high crime areas act as barriers.  SPR codes barriers on a four point scale.

Beyond beyond
This is the area which lies beyond the defined dynamic trade area and its outlying BandQuad area.  Every store attracts some proportion of business from great distances due to sales to customers in other counties, states, and countries.

Beyond sales (“beyond”)
Sales originating from customers living and/or working outside the defined trade area constitute beyond sales.  SPR divides up beyond sales into two parts, “BandQuad” sales, “beyond beyond” sales.

Bias
The concept known as “bias” represents the tendency of consumers to travel more or less frequently in a the direction of the subject site because of the influence of activity “magnets” such as shopping centers and major employment zones.  The number and distribution of access options also plays an important role in the influence of this variable on sales penetrations throughout a site’s trade area.  SPR has found this variable to be one of the most effective in sales forecasting.  Bias ratings are assigned by the analyst using a six-point scale.

Convergence
As used in sales forecasting, this term means the simultaneous use of multiple forecasting techniques to arrive at a single sales estimate.  The estimates emanating from the individual techniques (e.g., analogs, regression formulas, normal curves, market share, etc.) are compared by the analyst (or in automated routines in more sophisticated application systems) to determine the best final estimate based on the inherent strengths and weaknesses of each technique in the circumstances under review.  In some situations, the analyst may place more emphasis/weight on the output of a model formula than analogs and in other cases they may be treated as equally applicable.  Convergence forecasting has proven to be a valuable method in producing highly accurate sales forecasts. See Convergent Sales Forecasting for a more detailed explanation.

Customer source survey data
This information indicates the point of origin of customers who visit the store.  Most typically this is the location of the customers residence.  However, in some circumstances a work location may be more or equally as appropriate.  This information can come from a number of different sources depending on budget constraints, technological capacity, the nature of the business, etc..  Customer source survey information is the most vital component in an effective sales forecasting system.  If the survey contains a sufficient sample sizes the market analyst can determine the distribution of customers, an important step in defining trade area and sales penetration by map section.

Disaggregation
This is the process of dividing up the trade area into component geographic units.  The most common units are ZIP Codes, census tracts, block groups, and grids.  This technique allows the analyst to more precisely examine the influence of individual variables on sales performance.  Spatial relationships such as distance, access and bias orientation can be more effectively assessed in disaggregated datasets.  This process also has the statistical advantage of adding more observations to the dataset, allowing for greater accuracy.

Dynamic trade area
A primary trade area is the contiguous area around the store which contains the majority of its customers.  Sales penetrations tend to decline rapidly beyond the defined trade area.  A dynamic trade area is one whose geographic shape is responsive to the impact of market factors such as population density, competition and accessibility.  A dynamic trade area is significantly better for analysis than a standardized trade area (e.g., a 5 mile radius around the site).  For example, a dynamic trade area could extend 3 miles north, 5 miles to the east and west, and 8 miles to the south.  This more closely matches “reality” than a standardized trade area method where a unvarying “cookie cutter” such as a 5 miles radius is used.

Map section
A map section is the area of geography lying within an identified boundary.  Common map sections are block groups, census tracts, ZIP Codes, counties, states, and countries.
multivariate
This means a set of two or more variables used in a sales forecasting technique.

Normal curves
This is one of the oldest sales forecasting techniques.  Sales penetrations are graphed against one or more variables to depict their impact on sales performance as their values change.  For example, one common normal curve would show the decrease in sales per capita over distance for three different population density ranges as represented by three curvilinear lines (distance on the x-axis and sales per capita on the y-axis).  Normal curves are especially effective when used in conjunction with analogs.

Regression model formula (“model”)
Regression modeling is one of the most powerful sales forecasting techniques.  It allows the analyst to effectively explain the most statistically significant variables which impact sales performance.  Moreover, the impact of changes in individual or sets of variables can be readily observed.  Technically speaking, regression analysis means the modeling of a dependent variable (e.g., sales) as a function of a set of independent variables (e.g., demographics, competition, distance, bias, site characteristics, etc.).

Sales = X1 * 9.45 + X2 * -0.0034 + X3 * 12.45 + …