To: Mary Kazinczi GM Alfonsos Department Store

From: Muhammad Aqeel Ashraf - Consultant

Date: 22nd Sep, 2016

Subject: Assessing effectiveness of Alfonsos newspaper advertising

Purpose of doing this study was to analyze the effectiveness of Alfonsos newspaper advertising
and making recommendations for the future of newspaper advertising.
In order to accomplish this, data spanning over 26 weeks was collected such that per week sales
and newspaper advertising expenditures for Alfonsos and competitors were recorded. (Exhibit 1)
After collecting this data, regression analysis was run on this data in SPSS software to find
answers to above mention questions. (Exhibit 2- Exhibit 3)
Regression: Alfonsos - Advertising expense vs Sales
First of all, scatter plot was made to see for any visual evidence of any relatedness in
independent variable (per week advertising expense) and dependent variable (respective sales)
for Alfonsos Stores. That graph show that there is no visible relationship between the two.
(Exhibit 2)
Then simple regression analysis was run between advertising expenditure) and respective sales
of Alfonsos Stores. According to Exhibit 2, correlation between these two variable is -.01. This
suggest that sales of department stores do not increase in linear fashion with the expenditure
incurred on newspaper advertising.
Furthermore an F-Test was conducted (Exhibit 2-continued) to verify our above conclusion
regarding relatedness of advertising expenditure and respective sales at Alfonsos Stores. This
test further confirmed our result.
R-squared value in this regression was .000 (Exhibit 2) whose interpretation is that no variability
in sales data per week of Alfonsos store is explained by advertising expense.
Regression: Alfonsos sales vs All stores sales
Scatter plot was made to see for any visual evidence of any relatedness in Alfonsos per week
sales and all stores per week sales. The graph show that there is a strong visible relationship
between the two variables. (Exhibit 3)
Then simple regression analysis was run between Alfonsos per week sales and all stores per
week sales. According to Exhibit 3, correlation between these two variable is .955. This suggest

Page 1 of 7

that there is a very strong linear relationship between Alfonsos per week sales and all stores per
week sales.
Furthermore an F-Test was conducted (Exhibit 3-continued) to verify our above conclusion
regarding relatedness of Alfonsos per week sales and all stores per week sales. This test further
confirmed our conclusion.
R-squared value in this regression was .912 (Exhibit 3 ) whose interpretation is that 91.2%
variability Alfonsos per week sales is being explained by all stores per week sales.
This high positive correlation between the increases in sales of Alfonsos and sales of all other
stores in a week despite the fact that other stores spend far less on advertising suggesting that
there are additional factors affecting the increase in sales of Alfonsos stores.
Recommendations

Rather than increasing true sales, newspaper advertising is costing store profits by
transferring sales from one department to another and from on point in time to another.
Management must come up with other means of advertising that ensures increase in
true sales.
In order to do more accurate studies, sales data should be collected for that particular item
for which advertising is being done rather than whole stores data. This will give us better
picture.

Page 2 of 7

Exhibit 1

Page 3 of 7

Exhibit 2
Regression: Alfonsos - Advertising expense vs Sales
Correlations
aSales
Pearson Correlation

Sig. (1-tailed)

aSales

1.000

-.010

aAdvert

-.010

1.000

.481

.481

aSales

26

26

aAdvert

26

26

aSales
aAdvert

aAdvert

Model Summaryb

Model
1

R Square
a

.010

.000

Adjusted R

Std. Error of the

Square

Estimate

-.042

222476.42282

a. Predictors: (Constant), aAdvert

b. Dependent Variable: aSales

Page 4 of 7

Exhibit 2 (continued)

F-Test
Ho: 1 = 0

Ha: 1 0
ANOVAa

Model
1

Sum of Squares
Regression
Residual

116827000.694
1187898209095.
345

Total

1188015036096.
038

df

Mean Square
1
24

116827000.694

Sig.
.002

.962b

49495758712.3
06

25

a. Dependent Variable: aSales

b. Predictors: (Constant), aAdvert

Test Statistic:
The p-value=0.962 in the last column in above table indicates that we cannot reject H o: 1 = 0,
since p-value is greater than = 0.05. Thus relationship is insignificant.

Page 5 of 7

Correlations
aSales
Pearson Correlation

Sig. (1-tailed)

aSales

1.000

.955

allSales

.955

1.000

.000

.000

aSales
allSales

Exhibit 3
Regression:
vs All stores

Model

allSales

aSales
26
26
Model Summaryb
allSales
26
26
R Square
Adjusted R
Std. Error of the

Square
1

.955

.912

a. Predictors: (Constant), allSales

Estimate
.908

66103.44708

Alfonsos sales
sales

b. Dependent Variable: aSales

Page 6 of 7

ANOVAa
Model
1

Sum of Squares
Regression

1083143058903
.608

Residual

104871977192.
431

Total

1188015036096.
038

df

Mean Square
1

24

1083143058903
.608

247.878

Sig.
.000b

4369665716.35
1

25

a. Dependent Variable: aSales

b. Predictors: (Constant), allSales

Exhibit 3 (Continued)
F-Test
Ho: 1 = 0

Ha: 1 0

Test Statistic:
The p-value=0.000 in the last column in above table indicates that we can reject H o:
1 = 0, since p-value is less than = 0.05. Thus relationship is significant.

Page 7 of 7