SUMMER INTERNSHIP REPORT

ON

Consumer Insights Of Beverage Industry

Using Data Analytics
Completed At

Under the Guidance of

Ms. SURBHI BANSAL

Chief Data Scientist
(Project Guide & Mentor)

Submitted At
Department of Management Studies, IIT Roorkee

By: SAURABH SINGH (13810069)

Acknowledgement

First of all I would like to thank Bigvue Consulting, New Delhi, for giving me this
opportunity to learn, share my ideas, views, and thoughts on various subjects
through this project, undertaken as a part of curriculum requirements.
It has been rightly said that we are built on the shoulders of others but the
satisfaction that accompanies the successful completion of any project would be
incomplete without the mention of the people who made it possible.
I am highly indebted to my project guide Ms. Surbhi Bansal Chief Data
Scientist for giving me an opportunity to work on this live project based on data
analytics. I would also like to thank her step-by-step guidance and timely inputs
during all junctures of the project made it a great learning experience.
I also express our gratitude to Dr. Santosh Rangnekar - HoD, Department of
Management Studies, IIT Roorkee for the moral support and motivation
provided to us.

SAURABH SINGH
MBA 2013-15
(13810069)

Table Of Contents
I. Completion Report ............................................................................................... 1
II. Acknowledgment ................................................................................................ 2
1. Introduction ........................................................................................................ 5
2. The Analytics Industry ........................................................................................ 7
3. Company Background ........................................................................................ 8
About ................................................................................................................... 8
Major Functional Aspects ................................................................................... 8
Major Strategic Aspects ...................................................................................... 8
Project: To Generate The Consumer Insights About Beverage Industry
In India, Using Data Analytics
4. Introduction ...................................................................................................... 9
4.1 Problem Statement .................................................................................... 9
4.2 Project Scope ............................................................................................. 9
4.3 Literature Review ...................................................................................... 10
5. Research Methodology ................................................................................... 12
5.1 Variable Identification .............................................................................. 12
5.2 Questionnaire Design ............................................................................... 13
5.3 Data Collection ......................................................................................... 14
5.4 Data Analysis ............................................................................................ 15
6. Research Hypothesis ...................................................................................... 17
7.

Analysis ........................................................................................................... 23
7.1 Demography Based Analysis ...................................................................... 23
7.2 Bivariate Analysis ....................................................................................... 25
7.3 Factor Analysis ........................................................................................... 31
7.4 Cluster Analysis .......................................................................................... 34

8. Recommendation .......................................................................................... 38

List of Tables & Figures

Table 1: Cross-Tabulation of frequency of beverage purchase by customers and
the quantity of beverage they prefer to buy
Table 2: Chi-Square Test Result
Table 3: Cross-Tabulation of frequency of beverage purchase by customers and
customer gender
Table 4: Chi-Square Test Result
Table 5: Cross-Tabulation of beverage kind purchased by customer & customer
profession
Table 6: Chi-Square Test Result
Table 7: KMO & Bartletts Test
Table 8: Total variance among the components
Table 9: VARIMAX - rotated Component Analysis Factor Matrix
Table 10: Results of Cluster analysis

List of Figures
Figure 1: Purchase quantity of beverage (for all categories of customers) Vs.
Frequency of purchase
Figure 2: Purchase frequency of beverage (for all categories of customers) Vs.
Gender count of customers
Figure 3: Kind of beverage purchased Vs. Profession of those customers (for all
categories of customers)
Figure 4: Age-wise distribution of customers of beverage
Figure 5: Gender-wise distribution of customers of beverage
Figure 6: Profession-wise distribution of customers of beverage
Figure 7: Distribution of beverage preferred by customers
Figure 8: Easy Availability At Retail Outlets As Factor Of Popularity
Figure 9: Nutritional Value/Calorie Content As Factor Of Popularity
Figure 10: Brand Value As Factor Of Popularity
Figure 11: Type of Beverage As Factor Of Popularity
Figure 12: Distribution of beverage quantity preferred by customers
Figure 13: Distribution of beverage consumption frequency of customers
Figure 14: Distribution of channels of beverage promotion
Figure 15: Health-consciousness while consuming beverage w.r.t. gender
Figure 16: Health-consciousness while consuming beverage w.r.t. age
Figure 17: Kinds of Beverages preferred by customers w.r.t. their age
Figure 18: Age-wise importance of brand value
Figure 19: Age-wise importance of price of beverage
Figure 20: Scree plot
Figure 21: Cluster sizes & their relative size ratio
Figure 22: No. of clusters & cluster quality
Figure 23: Cluster Analysis Input (Predicted) Importance

1.

Introduction

Knowing your customers need is the first step of business and getting proper
information to serve your customer better is a success mantra, which is an
inevitable process for any organization. This is an era where competitive
advantage for one company becomes a must- be for others and can be imitated
very soon. So the only thing where an organization can get an edge over others is
by understanding their customers by clearly specifying the target markets and
defining a proper strategy. Market research is only visible way, which can help in
these matters. Analytics is one step ahead where organizations analyze their
crucial information to get proper customer insights. Market research is any
organized effort to gather information about target markets or customers. It is a
very important component of business strategy.
Analytics is powerful tools that can help organizations find eye-opening
opportunities to improve sales, marketing and customer service. Some companies
find ways to cut advertising costs and increase sales. Others reduce customer
churn and improve satisfaction ratings. Smart analytics goes beyond number
crunching and can help companies develop ways to listen and respond to the
needs of customers and the employees who serve them. Standing in the way of
adopting an analytic approach may be a corporate culture that encourages
individuals to rely on their intuition and experience when making decisions. Or
perhaps the organization lacks the processes and technology needed to collect,
analyze and develop actionable insights. From there, companies may require the
ability to transform the analytic insights into offers, products, information and
services customers will value. Analytics is the processes of getting the findings
which help in decision-making process using powerful tools, can help
organizations find eye-opening opportunities to improve sales, marketing and
customer service. Some companies find ways to cut advertising costs and increase
sales. Others reduce customer churn and improve satisfaction ratings. Smart
analytics goes beyond number crunching and can help companies develop ways
to listen and respond to the needs of customers and the employees who serve
them. Standing in the way of adopting an analytic approach may be a corporate
culture that encourages individuals to rely on their intuition and experience when
making decisions. Or perhaps the organization lacks the processes and
technology needed to collect, analyze and develop actionable insights. From
there, companies may require the ability to transform the analytic insights into
offers, products, information and services customers will value.
Customer Insights and Analytics can increase bottom line of the business.
Customer analytics is playing a large and growing role in helping organizations
use data collected to make improved business decisions. Applying analytics to the
customer database allows an organization to analyze customer behavior,
customer loyalty, and reduce customer churn. Customer insights help on
organization to:

Generate a higher return on marketing and promotional investments.

Increase sales to existing customers

Lower new customer acquisition costs

Reduce customer churn and increase loyalty

Resolve customer service problems more efficiently or better yet, avoid

them

Increase the sales force effectiveness by targeting qualified prospects.

Customer Insight is the intersection between the interests of the
consumer and features of the brand. Its main purpose is to understand
why the consumer cares for the brand as well as their underlying mindsets, moods, motivation, desires, aspirations, and motivates that trigger
their attitude and actions.
Another definition of consumer insight is the collection, deployment and
interpretation of information that allows a business to acquire, develop
and retain their customers.
Firstly, the collected data must be audited to fully understand the
quality and opportunity within the database. Once this is done, there are a
number of different types of analysis that can be applied.
Impact Assessment will help a business to understand how actions taken
by the business affected their customer behavior, and also allow for some
predictions of customer reaction to proposed changes.
Customers as Assets measures the lifetime value of the customer base
and allows businesses to measure several factors such as the cost of
acquisition and the rate of churn.
Propensity Modeling predicts the future behavior of customers based on
previous actions and helps businesses understand how likely it is that a
customer will behave in a given way.
Cross-Sell Analysis identifies product and service relationships to better
understand which the most popular product combinations are. Any
identified relationships can then be used to cross-sell and up-sell in the
future.
Critical Lag allows a business to deliver specific customer
communications based on the individuals purchase patterns, helping to
increase loyalty and improve customer retention.

2.

The Analytics industry

Analytics is science of examining data (text, quantitative, qualitative, etc.) to

bring forth underlying information. This information can give us some
undiscovered patterns or can establish hidden relationships, which can shape the
decision-making capability of an organization. There are two important facets of
Analytics: First is practical intuitiveness, there can be hundreds of ways a given
data can be analyzed, but the beauty is that none can be completely correct
however it will give us some direction. The point is to chase that direction and to
keep it updating with the trend. Second is real time, if a Google search engine will
take one hour to list down all the possible matches, its enormousness couldnt
have been achieved. All the analysis has to be done on the fly. Company cant
compromise a customer to escape just because they were not able to understand
his purpose of visit. Change they say is the need of the hour; Flipkart analyzed
that if they have to stay in the business they have to expand their horizon from
books to all life-style products. They were fast enough to analyze their customers
purpose of visit.
Facebook gave us an option to be friend again with the person whom we lost
contact some 10 years back, wasnt they were playing with some form of big data?
Companies like Wal-Mart, AT&T, Colgate-Palmolive, Amazon and many more
had long access to vast amounts of data in the forms of scanner, serial logs but
were mostly unstructured. With the advent of many analyzing Software like
Hadoop, SAS, Cognos, Tableau it became easier and faster to analyze it.
Doldrums In global economic condition has barred companies vertical expansion
and thus to maintain their operating margin they were forced to shrink their
operating cost. That allowed companies an opportunity to think and employ
methods which previously were deemed as not so mandatory rather an
investment.
With wide range of competition customer retention needs focused approach for
almost all companies, losing a single customer is a threat and adding even one is
not an easy task. Analytics came as savior that allowed companies to target
customers at the right place and at the right time.
So, Analytics is now no more a luxury for an organization rather a hygiene factor.
Let us have small look at the current analytics industry of India
Size of the Indian analytics Market: 375 Million $ No. of companies operating
in this segment in India More than 500 Expected Indian Analytics market by
2015 1.15 Bn $ as per Business standard report

3.

Company Background

About
Bigvue Consulting is a consulting company, which provides business solutions to
various organizations ranging from start-ups to large size companies in the field
of marketing research and data analytics.
It has the expertise in analyzing the crucial data of any organization to get
information and findings, which help various organizations in making decisions
and formulating strategy for the success of their business.
Bigvue has done many surveys on different sectors to get the consumer behavior
and doing an in-depth analysis to find out major finding. It has clients from
different regions of the globe who have appreciated Bigvue for its efforts and
services. Bigvue believes in long term partnership with organizations in strategy
formulation and business transformation.
It works with companies from various sectors such as: Banking and Financial
Services, Retails and CPG, Media and entertainment, Technology and Internet
companies, Govt. services and non-profit organizations.
Major Functional Aspects
Bigvue Consulting brings expertise in Analytics and Data Science. It leverages
Analytics and business understanding in building solutions that help companies
in taking business decisions that are more robust and accurate.
The main functional aspect includes:
Analytics and Business Intelligence
o Leveraging Analytics and Data Science for decision making and
business transformation
Data Visualization
o Visualizing data to generate actionable insights and ensure a better
decision making
Survey Analytics and Strategic Insights
o Using surveys and field data collection as a medium to gather, analyze
and generate actionable insights
Analytics Marketing
o Using Analytics to customize marketing activities, improve ROI and
optimize cost
Major Strategic aspects
Bigvue nurtures a solution-focused approach while creating actionable insights
using various open source and proprietary tools and software like SAS,
RapidMiner, SPSS, Angoss, Affinium, CART, R and MS Excel.

Project: To Generate The Consumer Insights About

Beverage Consumption In India, Using Data Analytics
4.

Introduction
The beverage industry considered in this project includes carbonated
drinks, packed juices, milk based drinks, energy drinks, sports drinks &
traditional packed drinks (Jal-jeera, Nimbu-pani etc.). The market of
beverage has evolved a lot since last decade. The non-alcoholic beverage
market in India is around $ 5Bn today, which includes the health beverage
market of around $ 300Mn and is expected to grow with at least 16% rate.
4.1 Problem Statement
To generate the consumer insights of Beverage Industry in India, using
data analytics.

The primary objectives of the study can be described as follows:

To study & analyze various customer insights regarding beverage

industry & its various segments on base of drink types
To recommend various open market acquisition strategies for using
various channels (Print, Digital & etc.)

4.2 Project Scope

The scope of the study is as follows:

The study is specific only to beverage industry in Indian context

The study revolves around Consumer Behavior and its related
factors

The factors studied are:

Type of beverage & its availability

Nutritional features of beverage
Flavor options
Price bands
Brand equity
Brand promotion channel
Ad frequency
Packaging

4.3 Literature Review

This project use the following concepts:
Consumer Behavior
Customer Analytics
Behavioral Clustering
And following techniques were used to extract consumer Insights:
Hypothesis Testing
Factor Analysis
Cluster Analysis

Consumer Behavior
It is the study of individuals, groups, or organizations and the processes
used by them to select and dispose of products, services & experiences to
satisfy their needs and the affect that the processes have on the
consumer. It combines the concepts of consumer psychology, marketing,
economics & sociology. It explains the decision-making process of buyers
whether individual or groups, on base of characteristics like
Demography, Economic factors, Socio-Cultural factors, Technological
factors or Natural factors. It tries to assess the influences on the
consumer to understand their wants.
Customer Analytics
Its the process in which the consumer/customer behavior data is used to
extract insights and on base of which business decisions are made. The
insights/ information is used by business for target market selection,
segmentation, marketing, product improvement and customer
relationship management. It can be used to predict the customer
preferences over a range of choices, Increase customer loyalty, providing
effective & efficient solutions to customer, tapping the right customers
with relevant message & channels to ultimately increase RoI.
Behavioral Clustering
Its a statistical analysis method used for FMCG, Retail etc. industries to
understand the patterns of buyers & consumers by working on data
gathered from channels like retail stores, surveys etc. from where
consumers directly buy items. On base of different attributes of purchase
by customer e.g. price, advertising, product features etc. groups are
formed to identify psychographic, demographic & geographic similarities.
It allows the company to identify which type of product sales are
dominant in any store or specific stores in any location. It helps
companies decide to take the key decisions about the business.

10

Hypothesis Testing
Its a method to make statistical inference on base of quantitative
findings. Making an assumption called Null Hypothesis, about rejection
of an outcome, which is unlikely to occur by chance alone according to a
determined probability & significance level, performs it. If a deviation
from outcome w.r.t significance level & probability is expected, it leads to
accepting the hypothesis and termed as Alternative Hypothesis. This
technique helps to decide whether the information provides proof to
reject a conventional assumption or accept it.
Factor Analysis
Its a set of procedures, which allows us to look at groups of variables that
tend to be correlated to each other and identify underlying dimensions
that explain these correlations between the variables. It is based on the
correlation matrix of the variables involved, & correlations need a large
sample size before they stabilize to a few factors. The amount of variance
a variable shares with all other variables is called Communality. The total
variance explained by each factor is called Eigenvalue. When the
Eigenvalues are plotted against the no. of factors in order of extraction, it
is called Scree plot. Scree plot helps to find the no. of factors. There are
different types of rotations done after the initial extraction of factors,
including orthogonal rotations, such as Varimax and Equimax, which
impose the restriction that the factors cannot be correlated, and oblique
rotations, such as Promax, which allow the factors to be correlated with
one another. A test used to examine the hypothesis that variables are
uncorrelated in the population, is Bartletts test. The KMO measure is an
index used to examine the appropriateness of factor analysis. High values
(between 0.5 to 1.0) indicate factor analysis is appropriate. Whereas
value below 0.5 indicate that factor analysis may not be appropriate.
Cluster Analysis
Cluster analysis is a technique used to classify cases or objects into
groups or segments that are more like each other than they are members
of other groups or segments called clusters. Clustering procedure can be
Hierarchical or Non-hierarchical. When the clustering process
progresses by the development of a hierarchy like a tree, it is called
Hierarchical. But when the process first assigns a cluster center and then
groups all objects within a pre-specified threshold value from the center,
it is a Non-hierarchy clustering. Cluster centers are the initial starting
point in Non-hierarchical clustering. Cluster centroid is the mean value
of the variables for all the cases in a cluster. Icicle Diagram is a graphical
display of clustering result. The columns correspond to the objects being
clustered and the rows correspond to the no. of clusters. It is read from
top to bottom.

11

5.

Research Methodology

The purpose of this topic is to outline the methodological decisions taken during
the course of this research project.
Marketing research can be defined as:
"The collection, analysis, interpretation and presentation obtained from
individuals or groups of people in order to guide citizens on a wide range
of matters, affecting customers, either as buyers or as citizens" (Domegan
and Fleming, 2007, Pg. 3.)
In order to ensure that the research was conducted in a reliable, systematic
manner, the researcher adhered to a series of steps known as the research
process.
"The marketing research process provides a systematic, planned approach
to the research project and ensures that all aspects of the research project
are consistent with each other"(Kumar, Aaker and Day, 2002 Pg. 48.)
In this project I will analyze qualitative data. The objective of the project is to
gather the customer insights in the retail banking in India. To study about
banking sector we have gone through rigorous study various papers/article/
journals/blogs across different forums. By structured literature review provided
us the strong background knowledge about retail banking products & services in
India and our mentor was not to exceed the number of questions more than ten.
Questionnaire was designed with utmost care and made live for data collection on
online digital portal like Facebook, LinkedIn, Twitter & various other blogs
related to banking domain. We kept a data collection window of seven days was
kept for collection of data. The collected responses are recorded in spread sheet
format which was statistically analyses using MS-Excel and SPSS. Analysis for
reliability and response biasness was carried out using SPSS and MS-Excel.
Various hypotheses was formulated and verified, for rejection, using chi-square
in SPSS. Excel was use for basic Univariate and Bivariate analysis. The analyzed
data was used to draw conclusion and recommendation. Each stage from start to
end is explained in the following sections.

5.1 Variables Identification

It is necessary to study industry before you prepare the questionnaire to
customers insights. With extensive literature review exposed me to
various factors or sub factors related to retail banking products& services
to its customer; also to understand better and identify missing feature in
the offerings in Indian context. To get customers insights and
perceptions about the product and the service I have conducted focused
group interviews with few selected customers personally & over phone.
Through this interviews and research we gained various features or the

12

characteristic of the various services offering by the retail banks.

Comparative study between different players or retail banks exposed us
to the various quality characteristics & services that one offer better than
other. This identified characteristics or features were used to design the
questionnaire.

5.2 Questionnaire Design

The questionnaire design depends on whether the person wants to collect
exploratory information (i.e. qualitative information for the purposes of
better understanding or the generation of hypotheses on a subject) from
the respondent or quantitative information (to test specific hypotheses
that have previously been generated). A well-designed questionnaire
should meet the research objectives. But in many cases a research
surveys omit important aspects due to inadequate preparatory work, and
do not adequately probe particular issues due to poor understanding.
Every survey is bound to leave some questions unanswered and provide a
need for further research but the objective of good questionnaire design
is to 'minimize' these problems. A questionnaire should also obtain as
accurate information as possible. The questionnaire designer needs to
ensure that the respondents should fully understand the questions. The
questions need not sound boring or unimportant to the respondent. It
should take care that a respondent should not find any reason of avoiding
or falsely replying to any questions. A well-designed questionnaire should
make it easy for respondents to give the necessary information. A good
questionnaire is well organized and encourages respondents to provide
accurate, unbiased and complete information. In case of those
questionnaires whose responses are collected using an interview, it
should be designed in such a way that the interview is brief and to the
point and also it should be so arranged that the respondent(s) remain
interested throughout the interview. Before framing questions of a
questionnaire, articulate the questions that research is intended to
address and determine the hypotheses around which the questionnaire is
to be designed.
Before the design of the questionnaire is complete and made live for
responses, check the length of the questionnaire. It need not be too long
that a respondent finds it inconvenient to respond. Next is Pre-test the
questionnaire by doing a small survey and analyzing the consistency
using Cronbachs Alpha. After all these check the final survey form can be
made live for the respondent to respond.
The questionnaire was designed with utmost care including all the
important features that a questionnaire should have, as listed above.
From the ease of readability to proper sequencing of each and every
question utmost care was taken. However despite including all these

13

features, the only constraint that I faced was to not exceed the number of
questions more than twelve. And under such constraints only important
questions were included. As an example, the profile information was
limited to parameters like profession and city with email being optional.
The questions that became the part of the questionnaire were multiplechoice questions. Respondent could select one or two options. Hence it
can rightly be said that the questionnaire was qualitative in nature with
all qualitative questions to be analyzed. The co- operation and support of
the mentor was highly important during the entire design phase.
5.3 Data Collection
It is the process of gathering and measuring information on variables of
interest, in an established systematic fashion that enables one to answer
stated research questions, test hypotheses, and evaluate outcomes. The
data collection component of research is common to all fields of study
including physical and social sciences, humanities, business, etc. Though
the methods or analysis vary, the importance of ensuring accurate and
honest collection remains the same. The goal for all data collection is to
capture quality evidence and then translates it to rich data analysis and
allows the building of a convincing and credible answer to questions that
have been posed. Regardless of the field of study or preference for
defining data (quantitative, qualitative), accurate data collection is
essential to maintaining the integrity of research. Both the selection of
appropriate data collection instruments (existing, modified, or newly
developed) and clearly delineated instructions for their correct use
reduce the likelihood of errors occurring. A formal data collection process
is necessary as it ensures that data gathered are both defined and
accurate and that subsequent decisions based on arguments embodied in
the findings are valid. The process provides both a baseline from which to
measure and in certain cases a target on what to improve. Various
consequences that result from improperly collected data are inability to
answer research questions accurately and inability to repeat and validate
the study.
There are various data collection methods. Data are collected through
online survey by asking the respondent to fill a questionnaire. Data can
also be collected by acquiring the available data, the Point-of-Sale data or
interviewing the candidate/respondents. It can be collected by observing
or measuring the object of interest. Most of the analysis today is carried
out on electronic devices and if the data or the responses are collected
manually, as in case of form fill up, it need to be fed into excel for data
processing or analysis. The chance or the probability of error is higher in
case of manual feed data as compare to other methods. The method
adopted by us for collection of data was online over Internet.
After finalizing the questionnaire design, a Google-form was created to

14

collect the online responses. The form was made live for just seven days.
The online form was posted on all possible online media. Social network
like Facebook groups and pages, Mobile App- Whatsapp, LinkedIn, and
Twitter were used to collect the responses. Online banking related portals
& blogs were also used to reach out to more number of people. Also Email
was send to as many people as possible. The responses received over
these seven days were stored in spreadsheet format for further analysis.
There is no biasness in collection of response it was made live on the
online forums and social media and the responses received were random
in nature.
5.4 Data Analysis
Analysis of data is a process of inspecting, cleaning, transforming, and
modeling data with the goal of discovering useful information, suggesting
conclusions, and supporting decision-making. It is the most crucial stage
in a research project. Right analysis gives right outcome on the basis of
which right strategy can be framed. Wrong analysis leads to wrong
decision. Data analysis has multiple facets and approaches,
encompassing diverse techniques under a variety of names, in different
business, science, and social science domains. Analysis of data includes
data mining, a particular data analysis technique that focuses on
modeling and knowledge discovery for predictive rather than purely
descriptive purposes, business intelligence covers data analysis that relies
heavily on aggregation, focusing on business information. Predictive
analytics focuses on application of statistical or structural models for
predictive forecasting or classification, while text analytics applies
statistical, linguistic, and structural techniques to extract and classify
information from textual sources, a species of unstructured data. All are
varieties of data analysis.
Post the data collection phase; some initial checks were carried out to
ensure the accuracy and authenticity of the collected data. Data defining
and cleaning is a precursor to data analysis. Data analysis is closely
linked to data mining and visualization. All of these steps have been
carried out in the analysis of the response data for customer insight.
From statistical point of view, people divide data analysis into descriptive
statistics, exploratory data analysis, and confirmatory data analysis.
Exploratory data analysis focuses on discovering new features in the data
whereas confirmatory data analysis focuses on confirming or falsifying
existing hypothesis. Exploratory data analysis technique has been used to
analyze the qualitative data. Following analysis was carried out to extract
the customers insights based on categorical data:

15

a. Data Defining
The received data was in text format and in this text format only limited
analysis could be carried out. Test for correlations, ANOVA test,
reliability test, etc. cannot be performed on text based data. Hence the
text data was converted to numeric categorical data by assigning numeric
value to the options. The values were assigned to the options in a
chronological order as they appeared in the questionnaire. Also few of the
variables were club together for better understanding and results. The
assignment has no impact on the output of the result.
b. Data Cleaning
Here I checked the data for missing response. Since the data was
collected online and all the questions in the questionnaire were marked
mandatory, no records were found to have any missing field.
c. Check for Response Biasness
It is described as a biasness that arises as a result of the influential
responses from the participating respondent. Biasness in responses is
said to have occur if any response deviates from normal and are not
accurate or truthful. Biasness is also said to have occur if the responses
are different from those collective responses from other respondents.
These biases occur more frequently in those types of study or research,
which involve participant self-report, like interview or survey. Response
biases can have a large impact on the validity of the questionnaire or
survey to which the participant is responding. The bias is caused by a
number of factors, all relating to the idea that human being do not
respond passively to stimuli, but rather actively integrate multiple
sources of information to generate a response in a given situation. As a
result of this almost any aspect of an experimental condition might be
able to influence a respondent in some way or the other. Few of the
commonly occurring response biasness reason include improper design
of the questionnaire, the biasness of the researcher, the methodology of
how responses are collected or the mood of the participant to be a good
or bed respondent and to provide socially desirable responses or correct
responses
The response biasness can lead to wrong outcome of the research and
hence it is important for a researcher to be aware of response bias and
the effect can have on their research so that they can attempt to prevent it
from impacting their findings in a negative manner. The respondent can
also result into response biasness by selecting only one type of option for
all the questions. Different reasons of response biasness are incomplete
information, incorrect questionnaire design, incorrect sample selection,
and wrong sample or target respondent and extreme response error or
biasness.

16

6.

Research Hypothesis

I have tested 3 hypotheses here, which are as follows:

a) To understand the dependence between the frequency of beverage purchase
by customers and the quantity of beverage they prefer to buy, a Chi-Square
test was conducted.
H0: There is No Association between purchase frequency and quantity preferred.
H1: There is Association between purchase frequency and quantity preferred.
Below image show cross tabulation between the above stated variables. The cross
tabulation was carried out using SPSS.
Table 1: Cross-Tabulation of frequency of beverage purchase by
customers and the quantity of beverage they prefer to buy
How often do you have a beverage? * What quantity do you usually prefer to
buy? Cross-Tabulation

How often do
you have a
beverage?

Total

1-3 times a
week
4-6 times a
week
More than 6
times a week
Once in 15 days
or more days

What quantity do you usually prefer

to buy?
200300 500
250 ml
ml
ml 1 litre 2 litre Total
6
6
29
12
2
55
28

28

64

34

23

61

17

77

60

44

13

197

The Table 2 below shows that Chi-Square test significant value is less than 0.05.
Hence we reject Null Hypothesis and accept Alternate Hypothesis concluding
that purchase frequency and quantity purchased are associated.

17

Table 2: Chi-Square Test Result

Chi-Square Tests
Value
95.648a
99.069
41.605

Asymp. Sig.
df
(2-sided)
12
.000
12
.000
1
.000

Pearson Chi-Square
Likelihood Ratio
Linear-by-Linear
Association
N of Valid Cases
197
* a: 9 cells (45.0%) have expected count less than 5. The minimum expected
count is 0.26

Figure 1: Purchase quantity of beverage (for all categories of customers) Vs.

Frequency of purchase

18

From Figure 1 above, it can also be observed that people preferring small
quantity, tend to purchase more as compared to others.
b) To understand the dependence between the frequency of beverage purchase
by customers and the gender of customer, a Chi-Square test was conducted.
H0: There is No Association between purchase frequency and customer gender.
H1: There is Association between purchase frequency and customer gender.
Below image show cross tabulation between the above stated variables. The cross
tabulation was carried out using SPSS.
Table 3: Cross-Tabulation of frequency of beverage purchase by
customers and customer gender
How often do you have a beverage? * Gender? Cross-Tabulation
Gender?
Male
Female
Total
How often do you
1-3 times a week
45
10
55
have a beverage?
4-6 times a week
48
16
64
More than 6 times a
46
15
61
week
Once in 15 days or
7
10
17
more days
Total
146
51
197
The Table 4 below shows that Chi-Square test significant value is less than 0.05.
Hence we reject Null Hypothesis and accept Alternate Hypothesis concluding
that purchase frequency and customer gender are associated.
Table 4: Chi-Square Test Result
Chi-Square Tests
Value
11.394a
10.104
6.323

Asymp. Sig.
df
(2-sided)
3
.010
3
.018
1
.012

Pearson Chi-Square
Likelihood Ratio
Linear-by-Linear
Association
N of Valid Cases
197
a: 1 cells (12.5%) have expected count less than 5. The
minimum expected count is 4.40

19

Figure 2: Purchase frequency of beverage (for all categories of customers) Vs.

Gender count of customers
It can be observed from Figure 2 that male customers buy beverages more
frequently than female customers.

c) To understand the dependence between the kind of beverage purchase by

customers and profession type of customer, a Chi-Square test was conducted.
H0: There is No Association between kind of beverage purchase and customer
profession type.
H1: There is Association between kind of beverage purchase and customer
profession type.
Below image show cross tabulation between the above stated variables.
The cross tabulation was carried out using SPSS.

20

Table 5: Cross-Tabulation of beverage kind purchased by customer &

customer profession
What kind of beverage do you prefer most? * Your Profession Cross-Tabulation
Your Profession

Service
What
kind of
beverage
do you
prefer
most?

Fruit Juices and

Natural Drinks
like
lemonade/Fruit
Juice

Count

Energy drinks

Count

41

12

16

29.5

14.2

24.2

12

5.1

2.4

4.1

.1

.3

12.0

2.5

1.2

2.1

.0

.2

6.0

12

30

47

Expected
Count

19.8

9.5

16.2

.2 1.2

47.0

Milkshakes/Milk Count
based drinks
Expected
Count

20

16.4

7.9

13.5

11

19

8.0

3.9

6.6

.1

.5

19.0

Expected
Count

1.7

.8

1.4

.0

.1

4.0

Count

83

40

68

197

83.0

40.0

68.0

1.0 5.0

197.0

Expected
Count

Expected
Count
Sports drinks

Count
Expected
Count

Carbonated
drinks (Pepsi,
Coke etc.)

Tea/Coffee

Count

Count
Expected
Count

Other

Total

Count

Expected
Count

Ho
use
Self
employed/
wif
Business Student Retired e Total
0

70

.4 1.8

70.0

39

.2 1.0

39.0

21

The Table 6 below shows that Chi-Square test significant value is less than 0.05.
Hence we reject Null Hypothesis and accept Alternate Hypothesis concluding
that kind of beverage purchased by customers and their profession are associated
[for a couple of categories].
Table 6: Chi-Square Test Result
Chi-Square Tests
Value
49.102a
55.863
6.032

Asymp. Sig.
df
(2-sided)
24
.002
24
.000
1
.014

Pearson Chi-Square
Likelihood Ratio
Linear-by-Linear
Association
N of Valid Cases
197
a. 23 cells (65.7%) have expected count less than 5. The
minimum expected count is .02.

Figure 3: Kind of beverage purchased Vs. Profession of those customers (for all
categories of customers)
It can be observed from Figure 3 that customers having Service & Students as
profession have preference for different kind of beverages e.g. Natural drinks &
carbonated drinks.

22

7.

Analysis

For analysis I have conducted online survey using various social media platforms
like Facebook, Twitter, LinkedIn and Google form survey page links
(www.bigvue.in). Through the survey we have collected response samples across
pan-cross India covering around 27 cities both Tier-I & Tier-II cities. Following
chart showing the respondents profile based on there Tier-Wise.
We have done bivariate analysis using MS-Excel tool for observing better
visualization. Similarly we had performed various analyses, which includes factor
& cluster etc. for assessing overall degree of accuracy of collected data.
Statistically analyze inter-relationships among large number of variables and to
explain these variables in terms of their common underlying dimensions
(factors). With using cluster analysis we use multivariate procedure for
segmentation.

7.1 Demography Based Analysis:

Customer Distribution
2.03, 2%
9.14, 9%

11.68, 12%

37.56, 37%

15-19 years

20-25 years

39.59, 40%

26-35 years

36-45 years

45 years and above

Figure 4: Age-wise distribution of customers of beverage

The age-wise participation of the customers in the project had the following
distribution.

23

Gender-wise Distribution

Customer Distribution
Male
Female

26.40, 26%

73.60, 74%

Figure 5: Gender-wise distribution of customers of beverage

The customers on whom survey was done majorly included males in comparison
to females.
Profession-wise Distribution

Customer Distribution

Total

42.64%
34.52%
19.80%
3.05%
Housewife

Self employed/
Business

Service

Student

Figure 6: Profession-wise distribution of customers of beverage

The customers on whom survey was done included majorly from service or
student profession.

24

7.2 Bivariate Analysis

Most Preferred Beverage:
The most preferred beverages by customers can b shown by the following
Pie chart:

Most Preferred Beverage

3%
2%

Carbonated drinks (Pepsi,

Coke etc.)
10%

Energy drinks
24%

20%
6%

Fruit Juices and Natural

Drinks like lemonade/Fruit
Juice
Milkshakes/Milk based
drinks
Other
Sports drinks

35%

Tea/Coffee

Figure 7: Distribution of beverage preferred by customers

It was observed that the most preferred beverage by customers was Fruit Juices
and Natural Drinks. It was also observed that if observed the collaborative
customer preference on base of Fruit Juices/Natural Drinks and Milk based
drinks, they constitute a major market share. It indicates that people prefer
drinks, considering the health reasons.

25

Factors Contributing to Beverage Popularity

It was observed that the popularity of beverage depends majorly on the
following factors:
Response For
Nutritional
Value/
Calorie
content
100

90

90

80

80

60

20

10

10

40

20

Responses as Not Important

20

Responses as Not Important

30

30

Responses as Somewhat Important

40

Responses as Very Important

Responses as Not Important

Responses as Somewhat Important

Responses as Very Important

20

40

Responses as Somewhat Important

50

40

60

50

Responses as Very Important

60
60

80

70

70

80

100

Responses as Not Important

100

120

100

Responses as Somewhat Important

120

Response
For Type of
Beverage

Response
For Brand
Value /
Brand Name

Responses as Very Important

Response
For Easy
availability
at retail
outlets

Figure 8: Easy Availability At Retail Outlets As Factor Of Popularity

Figure 9: Nutritional Value/Calorie Content As Factor Of Popularity
Figure 10: Brand Value As Factor Of Popularity
Figure 11: Type of Beverage As Factor Of Popularity

26

Quantity Preferred By Customers

It was observed that 200-250ml is the most preferred quantity by the
customers and as the packaged quantity increases, customers prefer the
less.

Quantity Preferred
Custmer Percentage

40.00%
35.00%
30.00%
25.00%
20.00%
15.00%
10.00%
5.00%
0.00%
2 litre

200-250 ml

300 ml

500 ml

Quantity Preferred
Figure 12: Distribution of beverage quantity preferred by customers

Consumption Frequency Of Beverage

It was observed that majority of the customers prefer to consume
beverage more than 4 times a week.

Beverage Consumption Frequency

8.63%
27.92%
30.96%

1-3 times a week

4-6 times a week

32.49%

more than 6 times a

week
once in 15 days or
more days

Figure 13: Distribution of beverage consumption frequency of customers

27

Channels Of Promotion
It was observed that the most effective channel of promotion of beverage
is TV/Radio advertisements followed by word of mouth.

Channel of Promotion
Total

Word of mouth

28.43%

TV/ Radio advertisements

Stalls or promotions (In exhibitions
and trade fairs)

40.61%
8.12%

Social media and Internet

6.60%

Newspaper and magazines

7.11%

Hoardings and banners

9.14%

Figure 14: Distribution of channels of beverage promotion

Health-consciousness w.r.t. Gender of Customers
It was observed that Male are more health-conscious while consuming a
beverage.

Health-consciousness w.r.t. Gender

72.19%

80.00%
60.00%
40.00%

27.81%

20.00%
0.00%
Female

Male

Figure 15: Health-consciousness while consuming beverage w.r.t. gender

28

Health-consciousness w.r.t. Age of Customers

It was observed that customers are health conscious, specifically above
the age of 25 years.

Age-wise Health Conciousness

100.00%

80.85%

80.00%

72.82%

60.00%
40.00%
20.00%

Above 25
19.15%

27.18%

Under 25

0.00%
No

Yes

Figure 16: Health-consciousness while consuming beverage w.r.t. age

Kind of Beverage preferred by customers w.r.t. their age
It was observed that customers above the age of 25 years prefer more
juices and milk-based drinks as compared to other drinks while
customers below the age of 25 years majorly prefer juices and carbonated
drinks.

Age-wise Kind of Beverage Preferred

45.00%
38.30%
40.00%
30.10%
33.01%
35.00%
24.47%
30.00%
25.00%
17.02%
15.53%
20.00%
10.64%
15.00%
8.74%
6.80%
4.85%
10.00%
5.32%
3.19% 1.06%
5.00%
0.97%
0.00%

Above 25

Under 25

Figure 17: Kinds of Beverages preferred by customers w.r.t. their age

29

Age-wise Importance of Brand Value:

It was observed that irrespective of age, customers give importance to a
known brand while consuming.

Age-wise Importance of Brand Value

60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%

45.74% 48.54%

54.26%
50.49%

0.00% 0.97%
Not Important

Somewhat Important
Very Important
Above 25
Under 25

Figure 18: Age-wise importance of brand value

Importance of Price Range of Beverage w.r.t. Age
It was observed that for the customers below the age of 25 years, price of
beverage was more important than the customers who were above 25
years of age. It explains the behavior that since most of them are
students, price of beverage is important for them.

Age-wise Importance of Price of

Beverage
Above 25

Under 25
53.40%

60.00%

47.87%

50.00%

47.87%

40.00%

38.83%

30.00%
20.00%
10.00%

7.77%
4.26%

0.00%

Not Important Somewhat Important Very Important

Figure 19: Age-wise importance of price of beverage

30

7.3 Factor Analysis

To assess the overall significance of the correlation matrix with the Bartlett test,
considering the data in this research, the correlations, when taken overall, are
significant at the 0.0001 level which is 81.97. However, by doing so we test only
the presence of non-zero correlations and not the pattern of the correlations.
The other overall test is the measure of sampling adequacy (MSA), which as far as
this research is concerned; fortunately, it falls in the acceptable range (of above
0.5) with a value of 0.536.
Table 7: KMO Index & Bartletts Test of Sphericity
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling
.536
Adequacy.
Bartlett's Test of
Approx. Chi-Square
81.974
Sphericity
df
36
Sig.
.000
Table 8: Total variance among the components
Total Variance Explained
Initial Eigenvalues
Compone
nt

Tot
al

1.57
6
1.35
6
1.28
9
.954
.89
3
.818
.801
.736
.576

2
3
4
5
6
7
8
9

Extraction Sums of
Squared Loadings

% of
Varian Cumulati Tot
ce
ve %
al
17.508
15.068
14.324
10.601
9.926
9.086
8.905
8.182
6.401

Rotation Sums of
Squared Loadings

% of
Varian Cumulati Tot
ce
ve %
al

17.508 1.57 17.508

6
32.576 1.35 15.068
6
46.899 1.28 14.324
9
57.500
67.426

% of
Varian Cumulati
ce
ve %

17.508 1.46 16.250

2
32.576 1.38 15.411
7
46.899 1.37 15.239
1

16.250
31.661
46.899

76.512
85.417
93.599
100.000

31

Figure 20: Scree plot

From Scree plot it is quite visible that the number of factors from factor
analysis is at least 3.

32

Table 9: VARIMAX - rotated Component Analysis Factor Matrix

Rotated Component Matrixa
Component
1
2
3
Flavor options
-.107
.007
.762
Type of
.658
-.165
-.073
Beverage
Ease of
.598
.187
.318
Availability
Price bands
.111
-.016
.809
Nutritional
-.009
.597
-.104
Brand value
.063
.677
.020
Packaging
.432
.355
-.037
Promotion
-.113
.614
.127
Ad frequency
.667
-.085
-.043
Extraction Method: Principal Component Analysis
Rotation Method: Varimax with Kaiser Normalization
(Rotation converged in 5 iterations)
Here we can observe from Table 9 that the 3 factors emerged from factor analysis
are as follows:
a) Variety & Convenience
i. Beverage Availability
ii. Beverage Type
b) Health & Media Influence
i. Nutritional Content of Beverage
ii. Brand value
iii. Promotion
c) Product Features
i. Price bands
ii. Flavor Options
Attributes that were not important are:
a) Packaging
b) Ad frequency

33

Key points to observe:

The perceptual maps show the mapping of different factors in the minds of
the customers and the association between them. The data was reduced
from many attributes to 3 factors that have similar attributes

The packaging and ad frequency are not a major attribute in any

of the factors. This justifies the fact that almost all the beverages have
high ad frequency but still that category e.g. milk-based drinks are doing
well in market. Neither the type of the packaging affects the customer
preference while making a purchase for beverage.

7.4 Cluster Analysis

In this project it was observed that 2 clusters emerged from 12 variables
as seen from Table 21. It was also observed that the cluster quality was
fair as observed from Table 11.
Figure 21: Cluster sizes & their relative size ratio

Figure 22: No. of clusters & cluster quality

34

It can be observed that Cluster validity is fair as observed in above summary

chart and hence cluster quality is good for segmentation purpose.

Figure 23: Cluster Analysis Input (Predicted) Importance

35

Table 10: Results of Cluster analysis

Variables
Age Group
Age
Profession
Through which medium did you
came to know your preferred drink
What kind of beverage you
preferred most
Health factors you consider for
beverage
How often you have beverage
Most preferred channel of
purchasing
Price fluctuations preference
changes
Quantity prefer to buy
What kind of packaging you mostly
prefer

Cluster 1
Under-25 (100%)
20-25 (77%)
Student (57.3%)
TV / Radio
advertisement
(43.7%)
Fruit juice & Natural
drinks (33%)
Yes (72%)

Cluster 2
Above-25 (100%)
26-35 (75.5%)
Service (56.4%)
TV / Radio
advertisement
(37.2%)
Fruit juice &
Natural drinks
(38.3%)
Yes (80.9%)

6 times in a week
(35.9%)
General store/ Retail
(60.2%)
No (55.3%)

4-6 times in a week

(31.9%)
General store/
Retail (59.1%)
No (58.5%)

200-250ml (37.9%)

200-250ml
(40.4%)
Plastic (28.7%)

Plastic (32%)

Using cluster analysis, 2 segments are derived:

CLUSTER 1:
It majorly comprises of respondents below 25 years of age. Most of respondents
are from the student profession. The channel from which they get most
influenced about any product is TV / Radio advertisements with a maximum
percentage of 43.7% among all the channels.
In this segment 72% of respondents of these segments consider health as the
important factors while choosing any beverage. It also becomes obvious from the
fact that most preferred drinks type is fruit juice & natural drinks with 33 %
dominance.
The average consumption of beverage of this cluster is 35.9% with a frequency of
6 times in a week and most preferred channel of purchasing is general stores &
retail market stores comprising a major part of total respondents (60.2%). Also
respondents who prefer packaging in plastic bottles are highest with 32% and the
maximum no. of respondents prefer to buy quantity of 200-250ml. This cluster of
customers preference in comparison to other clusters customers is affected more
by any fluctuations in price.

36

CLUSTER 2:
It majorly comprises of respondents above 25 years of age. Majority of
respondents are from the service profession. The channel from which they get
most influenced about any product is TV / Radio advertisements with a
maximum percentage of 37.2% among all the channels.
In this segment 80.9% of respondents of these segments consider health as the
important factors while choosing any beverage. It also becomes obvious from the
fact that most preferred drinks type is fruit juice & natural drinks with 38.3 %
dominance.
The average consumption of beverage of this cluster is 31.9% with a frequency of
6 times in a week and most preferred channel of purchasing is general stores &
retail market stores comprising a major part of total respondents (59.1%). Also
respondents who prefer packaging in plastic bottles are 28.7% and the maximum
no. of respondents prefer to buy quantity of 200-250ml. Majority of this segment
customers preference is unaffected by any fluctuations in price.

37

8.

Recommendations:
The majority of the customers consume a beverage more than 6 times a
week. It was also observed that customers prefer small packaging of about
250 ml but tend to buy it more frequently. So a beverage firm should
manufacture & market small quantity packaging aggressively.
The student class of customers has an inclination to buy carbonated drinks
so promotion of those drinks should be done keeping demographic factors
of students. Whereas the service class of customers showed inclination
towards fruit juices / natural drinks and milk-based drinks, so marketing
of these drinks should be done keeping demographics of service class.
To purchase a beverage, customers give importance to its easy availability,
nutritional content and brand value; hence a beverage firm should focus
on market penetration and make efforts to improve its brand image. Any
beverage firm should also be cautious about the nutritional content of
beverage since customers have now become health conscious.
The customer segment below the age group of 25 years give more
importance to price fluctuations so a beverage firm should vary the price
of carbonated drinks & fruit juices only when necessary and it should be
done carefully. Whereas the customer segment of above 25 years of age,
comparatively gives less importance to price fluctuations.
The customers get influenced by the TV / Radio advertisements but not by
the advertisements frequency, so the beverage firms should focus on the
quality of beverage advertisements.

38

9. References:
Malhotra N. K. and Dash Satyabhushan (2010), Marketing
Research-An Applied Orientation, 5th Edition, Prentice-Hall,
Pearson-India.
Hair J F, Black W C, Babin B J and Tatham R L (2006),
Multivariate Data Analysis, 6th Edition, Prentice-Hall, Englewood
Cliffs, NJ.
Gupta.S.L and Gupta H. (2011), SPSS 17.0 for Researchers, 2nd
Edition, International Book House Pvt. Ltd.
Marketing Management (14th Edition) Philip Kotler, Kevin Keller,
Pearson Education
IBM SPSS Statistics 22 Core System User's Guide, Version 21
Wayne L. Winston, Microsoft Excel 2010: Data Analysis and
Business Modeling, Microsoft Press (2011)
http://hbswk.hbs.edu/topics/marketing.html
Consumer Behavior (3rd Edition) Michael Solomon, Gary Bamossy,
Soren Askegaard, Margaret K. Hogg, Pearson Education Ltd.
http://www.referenceforbusiness.com/management/CompDe/Consumer-Behavior.html
http://www-01.ibm.com/software/in/analytics/rte/an/customeranalytics/
Statistics for Management (6th Edition), Richard Levin, David S.
Rubin
http://www.ats.ucla.edu/stat/spss/output/factor1.htm

39