The most important aspect of implementing an effective site evaluation system involves enabling the analyst to use the best and most appropriate forecasting tools for each distinct site/map section situation.  This has proved to be superior to over reliance on one specific method or technique.
 

Convergence sales forecasting is the simultaneous use of multiple forecasting techniques to arrive at a sales estimate.  The sales estimates emanating from the individual methods, including analogs, regression formulas, normal curves, market share are compared and contrasted.  The analyst determines the best forecast for a given situation based on a specific siteís location and market area characteristics in relation to the inherent strengths and weaknesses of each technique.

To successfully employ any method the analyst must know the strengths and weakness of each method as it applies to real-world situations.  The analyst must also understand the limitations and peculiarities upon which each method was based, i.e., ìthe databaseî.

A trained and qualified analyst rapidly becomes familiar with the datasets and the best use of each method.  Our system always puts out a baseline forecast which gives the analyst a good starting point from which to raise, lower, or apply no change to the initial forecast.  We find that in many cases the original system output requires little or no adjustment.

In some cases the analyst may weigh each methodís output equally, and in many cases, the analyst will use the initial forecast given by the system, as all indicators are that the system forecast is accurate.

Example: The analyst may need to place more emphasis on the output of a regression model estimate than analogs because there are no ìgoodî analogs (no existing stores) which reflect several conditions simultaneously: a very high positive demographic (like income) in a trade area with very low population and with no existing competition.

Example: A good analog group may offer a convincing argument to raise the baseline model forecast when the majority of the analogs (12 out of the 15 pulled by the system) are very similar to the map section/trade area under review and indicate larger sales volumes.  The comparison with normal curves supports the increase based on a few additional analogs which verify the final higher sales estimate.

During the development process we determine which variables most powerfully drive (predict) sales and customize our proprietary sales forecasting system to automatically pull the most relevant collection of datapoints for each site/map section under review.  The analysts can then interactively apply their specific knowledge of the site under review with the most valid forecasting methods.

The final piece to the puzzle is utilizing the power of GIS technology. Besides producing pretty maps, GIS enables the analyst to rapidly and accurately describe and analyze the complex spatial dynamics impacting sales performance.

The process of ìconvergingî on the final sales forecast has proven to be very accurate, especially when applied to the variety of situation which retailers encounter everyday.