Digital Marketing: Analyzing Customer Engagement with Google Merchandise Store

This report analyzes the customer engagement with Google Merchandise Store in different countries using digital marketing methods. It examines the relationship between bounce rate, average session duration, transactions, and revenue. The findings suggest that there is no significant relationship between bounce rate and average session duration, but there is a positive correlation between average session duration and revenue. The report also highlights the need for digital campaigns to increase customer engagement and transactions in different countries.

Digital marketing
1

Table of Contents
INTRODUCTION...........................................................................................................................3
CONCLUSION................................................................................................................................3
REFERENCES................................................................................................................................4
2

INTRODUCTION
At higher levels the digital marketing means advertising that are delivered using digital
channels like search engines, social media, websites, email and the mobile apps. Digital
marketing refers to method through which the companies will endorse the goods, services and
the brands. It uses online and internet based digital technologies like the mobile phones, desktop
computers and the other tools (Steils, 2021). The importance of digital media has been increased
after the invent of different platforms used for social interactions. Digital marketing methods
have proper combination of the SEO, SEM, influencer marketing, content marketing, campaign
marketing, e-commerce and such other activities. The present report is based on Google
Merchandise Store who aims at identifying the customer engagement in different countries
around the world. It will analyse the level of engagement to identify which country is highly
engaged and which is least engaged. It will provide the company to analyse where digital
marketing efforts are to be made for increasing the customer engagement with Google
Merchandise Store.
Objectives
 To analyse the consumer engagement with website in different countries
In order to evaluate consumer engagement with website in different countries, the data is
taken which consists of many things. Here, dataset is obtained of Google Merchandise store of 1
year that of 1st Jan 2020 to Jan 2021 (Desai and Phadtare, 2017). In that data of each country and
number of users is specified in it. Also, revenue, bounce rate, transactions, etc. are included in it.
So, this will enable in analysing that in which country organisation is able to generate high
revenue. The data is analysed with help of SPSS in which various test are applied in it. With that
it will give insight on which country there are large number of customers. Along with it, in this
various test such as regression, correlation, etc. is applied to find out relationship between
variables. In addition, frequency of data will be interpreted. This will show in how many
countries consumer are engaged more and what is reason behind it. SPSS will also help in taking
proper decisions. Also, on basis of that decision can be easily taken. Thus, data analysis is
explained as below
Frequencies
3

Regression analysis
In this regression the variable chosen is bounce rate and average session duration. This is
because it will help in finding out how average session duration depends on bounce rate. Besides
that, it will be easy to evaluate how session rate varies if there is change in bounce rate of
different countries (Massara, and Porcheddu, 2018) Hence, on basis of that consumer
engagement with website will be identified.
Descriptive Statistics
Mean Std. Deviation N
avgsessionduration 142.1720 33.44263 50
bouncerate .56 .056 50
Correlations
avgsessiondurati
on
bouncerate
Pearson Correlation avgsessionduration 1.000 -.717
bouncerate -.717 1.000
Sig. (1-tailed) avgsessionduration . .000
bouncerate .000 .
N avgsessionduration 50 50
bouncerate 50 50
Model Summary
Model R R Square Adjusted R
Square
Std. Error of the
Estimate
Change Statistics
R Square
Change
F Change df1
1 .717a .515 .505 23.53994 .515 50.898 1
Model Summary
Model Change Statistics
df2 Sig. F Change
1 48a .000
4

a. Predictors: (Constant), bouncerate
ANOVAa
Model Sum of Squares df Mean Square F Sig.
1
Regression 28203.880 1 28203.880 50.898 .000b
Residual 26598.187 48 554.129
Total 54802.067 49
a. Dependent Variable: avgsessionduration
b. Predictors: (Constant), bouncerate
Coefficientsa
Model Unstandardized Coefficients Standardized
Coefficients
t Sig.
B Std. Error Beta
1 (Constant) 382.852 33.900 11.294 .000
bouncerate -429.506 60.203 -.717 -7.134 .000
a. Dependent Variable: avgsessionduration
Interpretation- From above table it can be analysed that the significance value obtained is
P= .000 that is less than P= 0.05. Thus, it can be evaluated that there is no relationship between
bounce rate and average session duration. This means session duration does not depend on how
much bounce rate is. Hence, even when there is change in bounce rate it does impact on avg
session duration.
Correlation
Here, the variables that are taken is avg session duration, transaction and revenue. This is
because it will help in finding out correlation between them that whether transaction and revenue
depends avg session duration or not (Miao, 2020). This means that if avg duration is high then
transactions are also more or not. So, it highly impact on overall revenue.
Descriptive Statistics
Mean Std. Deviation N
avgsessionduration 142.1720 33.44263 50
5

transaction 42.1400 281.73558 50
revenue 2581.0892 17015.32506 50
Correlations
avgsessiondurati
on
transaction revenue
avgsessionduration
Pearson Correlation 1 .409** .411**
Sig. (2-tailed) .003 .003
N 50 50 50
transaction
Pearson Correlation .409** 1 1.000**
Sig. (2-tailed) .003 .000
N 50 50 50
revenue
Pearson Correlation .411** 1.000** 1
Sig. (2-tailed) .003 .000
N 50 50 50
**. Correlation is significant at the 0.01 level (2-tailed).
Interpretation- It can be analysed from table that significance value obtained between avg
duration with transaction is .409 and with revenue is .411. So, it is less than P = 0.05. It means
that correlation is significant in it. However, it is found that correlation between transaction with
avg session duration is P = .409 and with revenue is 1. Here, both values are different but with
avg duration correlation is significant. Besides that, correlation of revenue with avg duration
is .411 that is less than P= 0.05.
Frequencies
Statistics
users newusers session bouncerate pagesession
N Valid 50 50 50 50 50
Missing 0 0 0 1 0
Mean 10671.9000 10500.1000 14864.5800 .56 3.3466
Median 3154.5000 3077.5000 4169.5000 .56 3.2350
Std. Deviation 33082.35462 32606.10345 50136.26165 0a .67016
Variance 1094442186.990 1063157982.173 2513644731.800 .056 .449
6

Statistics
avgsessionduration transaction revenue ecommerceconversionra
te
N Valid 50 50 50 0
Missing 0 0 0 50
Mean 142.1720 42.1400 2581.0892
Median 140.1700 1.0000 9.1750
Std. Deviation 33.44263 281.73558 17015.32506
Variance 1118.410 79374.939 289521286.750
7

Interpretation
It could be analysed from the above report that mean of users is 10761, median is 3154
and deviation is 33082.35. On average new users are 10500 with median of 3077. The session
has mean of 14864.58 with median 4169 and standard deviation of 50136.26. Further it has been
evaluated that bounce rate has mean of 0.56 with median of 0.56 and standard deviation of 0.056.
There is significant deviation from the mean results in bounce rate. The mean of average session
duration is 142.17 with median 140.17 and the standard deviation 33.43. It means that users
spend around 142 hours on merchandise store. It could be evaluated from the above analysis that
there is significant increase in the new users in every country however the increase in transaction
and revenues is very low (Azizi, and Hu, 2019). The transaction shows mean of 42.14 with
median of 1 and standard deviation of 281. The mean has significant variation from the actual
results. Revenues show mean of 2581.089, median of 9.175 and standard deviation of the
17015.32. It could be analysed that there is high deviation in results. The country that provides
highest revenues and transactions are US at first, then Canada and UK. People in these country
make highest use of merchandise store as compared with other stores. It could be analysed that
people needs to be made aware of the new modes through which business could be promoted and
brand recognition could be increased in different countries. Digital campaigns could be led to
increase the customer base and transactions in different countries with high number of users and
session period but have low or nil transactions throughout with merchandise store.
8

Correlation
Mean Std. Deviation N
bouncerate .56 .056 50
session 14864.5800 50136.26165 50
avgsessionduration 142.1720 33.44263 50
Correlations
9

bouncerate session a
v
g
s
e
s
s
i
o
n
d
u
r
a
t
i
o
n
bouncerate
Pearson Correlation 1 -.558** -.717**
Sig. (2-tailed) .000 .000
N 50 50 50
session
Pearson Correlation -.558** 1 .415**
Sig. (2-tailed) .000 .003
N 50 50 50
avgsessionduration
Pearson Correlation -.717** .415** 1
Sig. (2-tailed) .000 .003
N 50 50 50
Interpretation
Correlation is computed to identify strengths of linear relationship between the two
variables which are X and Y. Correlation which is higher than zero represents positive relation
between the variables where less than a zero shows negative relationship. In the present case
correlation is identified between bounce rate with session duration and average session duration.
The two variables are selected to identify relationship between bounce rate and session time and
their impact on customer engagement. Analysing the above table it could be seen that correlation
is -0.558 which is below zero. This shows that there is strongly negative correlation between the
10

two variables. This means that session time do not influence the customer engagement with
merchandise store (Ye, Feng, and Gao, 2019). When the correlation is highly negative other
factors having influence over the customer engagement should be researched. The correlation is
also calculated to identify relation between average session duration and bounce rate where
bounce rate is X and average session is Y. The computation of correlation has given output of -
0.717 which is also again strong negative relationship. It could be evaluated from the above
results that bounce rate and average session time are also not related with other. Further the
engagement of customer with merchandise store is not associated with session time and bounce
rate. There is strongly negative relation between both the variables.
Cluster analysis
Case
Processin
g
Summary
b,c
Cases
Valid Rejected T
o
t
a
l
Missing Value Out of Range Binary Valuea
N Percent N Percent N Percent N Percent
0 .0 51 100.0 0 .0 51 100.0
a. Value
different
from both
1 and 0.
b. Binary
Squared
Euclidean
Distance
used
11

c. Average
Linkage
(Between
Groups)
Hypothesis 1:
H0 (Null hypothesis): There is no significant relationship between the number of users and
revenue
H0 (Alternative hypothesis): There is a significant relationship between the number of users and
revenue
Model
Summary
Model R R Square Adjusted R
Square
Std. Error of
the Estimate
1 .973a .946 .945 4001.70
a.
Predictors:
(Constant),
users
ANOVAa
Model Sum of
Squares
df Mean Square F S
i
g
.
1
Regression 13417889418.9
4 1 13417889418.9
4 837.90 .000b
Residual 768653631.80 48 16013617.32
Total 14186543050.7
5 49
a.
Dependent
Variable:
revenue
12

b.
Predictors:
(Constant),
users
Coefficient
sa
Model Unstandardized Coefficients Standardized
Coefficients
t S
i
g
.
B Std. Error Beta
1 (Constant) -2757.044 595.215 -4.632 .000
users .500 .017 .973 28.947 .000
a.
Dependent
Variable:
revenue
Interpretation
As per the above table, it has been interpreted that there is a direct relationship between
the revenue and number of users because the value of p is lower than the standard criteria. That
is why, alternative hypothesis is accepted over other. Also, from the anova table, it is interpreted
that the strong relationship between both variables such as due to increase in number of users,
revenue of a Google Merchandise stores automatically increases. Apart from this, the table also
exhibits that if there is a fluctuation in number of users who frequently uses website, then there is
94% of chances to changes in company’s revenue (Zhang, and Rajasekaran, 2020).
Thus, it is clearly reflected that, individual who frequently uses the websites and purchase
the products online then company’s performance is also increases. This is analyzed that both
variables are directly related to each other and also affect the business performance in positive
manner as well.
Hypothesis 2:
H0 (Null hypothesis): There is no difference between the mean value country and revenue
H0 (Alternative hypothesis): There is no difference between the mean value country and revenue
13

Model
Summary
Model R R Square Adjusted R
Square
Std. Error of
the Estimate
1 .252a .064 .044 16636.28869
a.
Predictors:
(Constant),
country
ANOVAa
Model Sum of
Squares
df Mean Square F S
i
g
.
1
Regression 901770179.436 1 901770179.436 3.258 .077b
Residual 13284772871.3
17 48 276766101.486
Total 14186543050.7
53 49
a.
Dependent
Variable:
revenue
b.
Predictors:
(Constant),
country
Coefficient
sa
Model Unstandardized Coefficients Standardized
Coefficients
t S
i
g
.
B Std. Error Beta
1 (Constant) 10085.396 4776.932 2.111 .040
country -294.287 163.034 -.252 -1.805 .077
14

a.
Dependent
Variable:
revenue
Interpretation: In accordance with the above table, it has been identified that there is no
relationship between the users who visit websites in any country with revenue. It is so because
the value of p (0.07) is greater than the standard criteria i.e. P > 0.05. Therefore, it can be stated
that alternative hypothesis is rejected over null, so there is inversely relationship between users at
different country with sales of a Google Merchandize store.
Apart from this, as per the model summary table, it is interpreted that there is low
correlation between both variables and also if there is any fluctuation in users accessed websites
at different countries, then it impacts around 4% in the revenue of a company. Therefore, it is
clearly examined that users at different country will accessed websites is not affected the sales of
a company (Mannem, 2019). But if there is a sudden increase in number of users (without any
bounce) then it affects the business performance directly. Therefore, it clearly exhibits that
customer’s engagement with a website at different country is changing and there is no impact
upon the sales of a business group. Thus, to increase their engagement, it is suggested to the firm
to use creative content which helps to attract customers and as a result, company sales directly
affected effectually.
One same T test
One-
Sample
Statistics
N Mean Std. Deviation Std. Error
Mean
revenue 50 2581.08 17015.32506 2406.33035
country 50 25.50 14.577 2.062
Users 50 1.071.90 33082.35 4678.55
One-Sample
Test
Test Value = 0
15

t df Sig. (2-tailed) Mean
Difference
95% Confidence
Interval of the
Difference
Lower Upper
revenue 1.073 49 .289 2581.08920 -2254.6127 7416.7911
country 12.369 49 .000 25.500 21.36 29.64
Users 2.281 49 0.27 10671.90 1269.99 20073.8012
Interpretation
From the above statistic table, it is interpreted that the average value of revenue is
2581.08, whereas mean value of country is 25.50. On the other side, there is 14% chances of
fluctuation of country within sample size of 50. Through the applied tool, it is identified that
there is a significant relationship of country because 0.00 is lower than 0.05 and as a result,
alternative hypothesis is accepted over other. Also, the value of p in the term of revenue is higher
than the standard criteria and that is why, null hypothesis is accepted. Apart from this, in order to
analyze the customer engagement, it is identified that alternative hypothesis is accepted due to
lower p value from 0.05. This reflected that changes in the pattern of usage of website within a
country, it directly impacts upon sales or revenue of Google Merchandize Store and that is why,
there is a need to focused upon the selling strategy of a company so that it does not affect the
overall business in opposite manner (Jurišić, 2020).
Hence, it can be evaluated that bounce rate in many countries is high of Google
Merchandise Store. So, due to that avg duration session is low. By that revenue generated is low.
However, in nations such as Argentina, Switzerland, etc. there is no transaction which has
occurred. It is because of high bounce rate. In these session rate is moderate but no transaction
has occurred. In addition, bounce rate is not related to session duration. In US, there are highest
number of users (Richard, Stewart,. and Sackett, 2017). Thus, revenue generate is huge in that
and transactions are more. Moreover, conversion rate is high in US as compared to all others
countries. This is result of rise in new users as well. The implication of doing regression is that it
shows relationship between bounce rate and session. This helps in easy analysis of what is to be
done, how, etc. by
RECOMMENDATION
 The Google Merchandise store can focus on reducing its bounce rate so that more new
users are attracted and conversion rate increases as well. For that the company can
16

improve their website pages. They can put more relevant content in website along with
product description.
 Here, the company can focus on new market where there exist large number of
opportunities for them to gain new users and attract them. Thus, they can use various
social media tools to do promotion and advertisement (Peng, 2019).
 The can analyse data and then take relevant decision that to enhance their website
webpages. Thus, it will be useful in increasing session duration and conversion rate.
CONCLUSION
It has been analysed from the above report that digital customer engagement refers to the
method for interaction of customers with business. The importance of digital marketing is
increasing everyday in the modern world to promote their business and attract new customers
around the world. Customer engagement through digital media has been increasing but growth is
seen in few developed countries. US is the biggest user of merchandise store followed by Canada
and UK. The number of users are increasing with double rate but increase in transactions is not
seen. Highest revenues are earned from US, Canada and UK. It is also found that many
developed countries have not yet made single transaction with store. Company is required to
create awareness among such countries regarding the benefits that could be derived by using the
digital merchandise store. India has second largest users however they are not generating
required revenues and transactions from them. It should bring new offers and deals that would
attract the users and business to explore towards new platform. It is to be suggested to the firm
that they can focused upon different countries who has a strong economy. Also, company has to
identify the emerging trends of a country so that it developed strategy accordingly. This in turn
assists the firm to raise the position of a company at global market and also creates awareness
pertaining to use official website of Google Merchandize store. Therefore, with the help of this
strategy, quoted firm enhanced the performance and also increase its business operations
effectually. Further, focusing upon different trend will provide an opportunity to a business to
identify new areas of development that creates positive impact as well.
17

18

REFERENCES
Books and journals
Azizi, V. and Hu, G., 2019, June. Machine Learning Methods for Revenue Prediction in Google
Merchandise Store. In INFORMS International Conference on Service Science (pp. 65-
75). Springer, Cham.
Desai, D. and Phadtare, M., 2017. Attributes influencing retail store choice decision of shoppers:
A case of Pune City. Vision, 21(4), pp.436-448.
Jurišić, N., 2020. Analiza poslovanja alatima Google Analytics i Google Data Studio (Doctoral
dissertation, The Polytechnic of Rijeka).
Mannem, V.K.R., 2019. google analytics customer revenue prediction.
Massara, F. and Porcheddu, D., 2018. Affect transfer from national brands to store brands in
multi-brand stores. Journal of Retailing and Consumer Services, 45, pp.103-110.
Miao, Y., 2020, November. A Machine-Learning Based Store Layout Strategy in Shopping Mall.
In International Conference on Machine Learning and Big Data Analytics for IoT
Security and Privacy (pp. 170-176). Springer, Cham.
Peng, W., 2019. Prediction of purchasing power of Google store based on deep ensemble
learning model. Automation and Machine Learning, 1(1), pp.1-4.
Richard, O.C., Stewart, M.M. and Sackett, T.W., 2017. The impact of store-unit–community
racial diversity congruence on store-unit sales performance. Journal of
Management, 43(7), pp.2386-2403.
Steils, N., 2021. Using in-store customer education to act upon the negative effects of
impulsiveness in relation to unhealthy food consumption. Journal of Retailing and
Consumer Services, 59, p.102375.
Ye, Z., Feng, A. and Gao, H., 2019, November. Prediction of Customer Purchasing Power of
Google Merchandise Store. In International Conference on Advanced Data Mining and
Applications (pp. 839-852). Springer, Cham.
Zhang, S. and Rajasekaran, R., 2020. Exploratory Data Analysis, Descriptive and Predictive
Analytics Approach For Improving Google Merchandise Store Revenue Based on Google
Analytics Data.
Online
Dataset link, 2020. [online] Available through : < https://analytics.google.com/analytics/web/?
utm_source=demoaccount&utm_medium=demoaccount&utm_campaign=demoaccount#/
report/visitors-geo/a54516992w87479473p92320289/
_u.date00=20200101&_u.date01=20210101&geo-table.plotKeys=%5B%5D&geo-
table.rowCount=50/ >
19

APPENDIX
20