Locating New Pam and Susan’s Stores

Locating New Pam and Susan’s Stores

John Hopkins

July 5, 2015

Introduction

Pam and Susan’s is a discount department chain that is looking to open additional stores in the Southwest.  Having had success in this area in the past Pam and Susan’s is looking for the most desirable areas around the Southwest based on a collection of data that has been complied by the companies real estate experts.  There has been increasing importance on new store locations given the recent disappointing performance at a few new store locations.  In hopes to provide a more accurate assessment of Pam and Susan’s next store locations they have added seven “competitive types” of the trading zones.  There has been additional concern based on the new subjective data of the competitive types that has resulted in statistical analysis to assure the assessment is accurate.  Through careful statistical analysis using multiple regression techniques we will show clear direction to the leadership team as to where the next Pam and Susan’s stores shall be opened.

Data

A series of 32 data points were compiled of the 250 existing Pam and Susan’s stores by the real estate team.  This data consists of census based demographic data such as ethnicity, income categories per person, income per household, home value, household ownership, household demographics, population and family size.  The second part of the data is store based such as square feet of selling area, annual sales and percent of hard goods.  The last data category is what the team has added to this year’s assessment, competitive type.  The 7 competitive types are somewhat subjective but range from; location near shopping centers to locations that are in densely populated areas with relatively little direct competition.  In order to make the final decision on which store locations to open each of the independent variables mentioned above will be compared to the depended variable of sales.

There are three groups of independent variables in relation to sales.  Positively correlated, negatively correlated and not significantly correlated with sales at all. The following correlations are positively correlated with sales: %black, %spanishsp, %inc0-10, %inc10-14, %inc14-20, %nocars, %sch0-8, populat, sqrft.  An example of this correlation is shown below.

[pic 1]

The negatively correlated independent variables are; %inc20-30, %inc30-50, medianinc, medianrent, %owners, %washers, %dryers, %dishw, %aircond, %freezer, %sechome, %sch12, %sch12+, famsize, comtype.  As seen below the opposite relationship between the graph above.

[pic 2]

The last correlation is the ones that were not significantly correlated with sales; %inc50-100, %inc100+, medianhome, %1car, %tvs, %sch9-11, perhard.  These points can be excluded from a lot of our analysis due to their lack of significance with sales.  As shown below the scatter plot does not tell you much besides it does not correlate with sales.

[pic 3]

The other data points given by the real estate team are the competition types.  As shown below of the 7 total competition types 1 has the highest average sales, 3 is around the middle and 7 is the lowest on average.  By using these three completion types moving forward we will have a good idea of the range of competition types and their relationship to sales.