Abstract

On

A Study of Shopper Buying Behaviour in terms of ‘Selection of

Retail Outlets’ and the ‘Impact of Visual Merchandising’

Submitted by

Rajesh Panda

For the degree of

Doctor of Philosophy

(Faculty of Management)

Symbiosis International University

PUNE

In

June, 2011.

Under the guidance of

Professor Anil Keskar

Director, Management Programmes

Sinhgad Technical Education Society

19/15, Erandwane, Smt. Khilare Marg

Off. Karve Road, Pune-411004

 1. Introduction

Retailing consists of all business activities that involve in selling goods and services to
the final customer for personal and household use without any intention to resale.
Retailing has evolved globally into a high-tech business. Global Retail companies like
Wal-Mart have already topped fortune 500 companies in turnover. At the same time,
customers are gradually getting more quality conscious and preferring shopping in a
better environment. As the preferences of customers are changing, new business
opportunities are also coming up in retailing. This also gives rise to new areas of study in
retailing e.g. Shopper buying behavior so that retailers can modify their shopper targeting
strategy.

It is beyond doubt that Retailing in India is one of the most attractive sectors of this
decade. While the retailing industry itself has been present through history in our country,
it is only the recent past that has witnessed so much dynamism and growth. It's the latest
bandwagon that has witnessed hoards of players leaping onto it. While international retail
store chains have caught the fancy of many travelers abroad, the action was missing from
the Indian business scene, at least till last 5 to 7 years. But now because of availability of
more number of retail formats and malls, customers have got abundance of choice in
relation to selection of a retail store and purchase of merchandise.

Despite the current changes, there is a lack of studies that have focused on the nature of
shopping behavior exhibited in the Indian environment. Most research in this area is still
proprietary in nature and hence is outside the public domain. In such a scenario, there
remains a need for studying the shopper behaviour. Given the rapid rate at which new
retail formats have been introduced in the Indian market in recent times, many with
limited success, it is imperative for Indian businesses to understand changing shopping
behaviour among consumers, especially, with regard to their preferred points of purchase.
With growth in disposable incomes and improving infrastructure, consumers have a wide
choice of stores where they can choose to shop. It is therefore, necessary for retailers to
understand shoppers‟ motivations and to attract customers residing beyond the catchment
areas around their stores.

1
Assuring quality product or investment in brand building activities is not just enough to
acquire new customers or retain existing customers. As customers are bombarded with
more and more options, they are confused on what to buy. This leads to customers being
less brand loyal and more variety seekers. This often results in customers switching from
one brand to the other, or to a substitute product. In this case, product availability and
visual merchandising play important role. Moreover, visual merchandising is helpful in
inducing impulse purchase of unplanned merchandise in customers also. Though visual
merchandising, a critical decision area in retail, the importance of same has not yet been
felt in India. Most modern formats of retailing arrange their merchandising without any
scientific study or use of a planogram. Similarly, stores come up on the basis of
availability of real estate and price proximity. But hardly any study has been done to
understand the implications of retail location and visual merchandising.
So the research aims at studying the shopper decision process in selection of a retail
outlet and the impact of visual merchandising on select category of products. Retailers
can use the research in designing their merchandising and market targeting strategy.
The research first tries to understand the different attributes that influence shopper buying
behaviour. The understanding of the attributes is developed from existing literature and
the same verified by exploratory survey with shoppers in Pune. Once the attributes were
finalised, data from shoppers was collected on those attributes on an importance scale.
The data was processed through SPSS using factor analysis to create factors that affect
shopper buying behaviour and subsequently factor scores were computed. Factor scores
were used in cluster analysis to create shopper clusters or shopper segments to understand
their demographics and buying behaviour.

Aims and Objectives:

1. Explore consumer decision process in retail by understanding the

different triggers for shopper buying behavior

2. Identify the various factors influencing customer behaviour and

create an understanding of retailers‟ approach towards consumer
behaviour

2
 3. Understand the different factors (after clubbing different triggers
into factors) and their relative importance in shopper buying
behaviour.

4. Impact of visual merchandising on select merchandise

5. Understand shopper clusters in terms of their buying behaviour

2. Review of Literature

The literature review focuses on the triggers of shopping identified by various researchers
across the globe, with specific emphasis on merchandising as an attribute and use of
factor analysis as a quantitative tool to create constructs affecting shopper behavior.
Moreover, it explores the different quantitative methods followed by different researchers
to reach at quantifiable factors under different conditions of study. The focus of the
literature review is to understand factors influencing shopper behavior and the scales
used by different researchers across the globe to address shopper buying behavior. This
also identifies the gap in the present research and derives hypotheses to address those
areas.

The journals that contributed the most are International Journal of Retail and distribution
Management, Journal of Consumer Marketing, European journal of marketing,
Sociological Methods & Research, Journal of Consumer Psychology, Journal of
Retailing, Harvard Business Review, Journal of Consumer Research, Journal of Shopping
Center Research, Journal of the Academy Marketing Sciences, Advances in Consumer
Research, Journal of Marketing Research, International Journal of Marketing and Journal
of Marketing Theory and Practice.

The research on shopper behavior goes back to 1969 when Dodge and Summer found out
that Store choice has been found dependent on socio-economic background of
consumers, their personality and past purchase experience. Later Kotler (1973), Cort and
Dominguez (1977), Mattson (1982), Meyer and Eagle (1982), Malhotra (1983),
Lumpkin et al. (1985), Zeithaml (1985), Fotheringham ( 1988), Sparks (1995), Davies
(1995) contributed a lot towards the research on shopper buying behavior. But most of

3
these researches were qualitative. McGoldrick and Betts (1995) conducted a research
named „ Consumer behavior and the retail sales‟ in UK and used quantitative analysis
(factor analysis) to bring out eight factors that influence shopper behavior in UK.
However till 2003, there was not much research done on Indian shopper behavior. Sinha
and Banerjee (2004) worked on “Store Choice Behavior in an evolving Market”. This
analysis is based on seven constructs and 43 variables. The seven constructs are
proximity, merchandise, ambience, service, patronized and others (two constructs). One
of the most recent Indian researches on shopper behavior is by Goswami and Mishra
(2008) who studied traditional kirana stores visa-a-vis organized Supermarkets for
grocery shopping and came out with factors related to grocery store patronage and
identified dimensions of customers‟ needs and desire which are relevant for grocery store
choice. In this study 44 attributes were considered in the questionnaire that was brought
down to 11 factors by factor analysis. A very recent research on shopper behavior is by
Swoboda, Haelsig, Klein and Morschett(2009) who used a 15 attribute model,
converting the attributes into 5 factors in their research “Moderating role of involvement
in building a retail brand”.

From the existing literature, it is quite evident that lot of work has been done on shopper
buying behavior globally, but except Sinha and Banerjee (2004) and Goswami and
Mishra (2008) there is hardly any relevant research being published on shopper buying
behavior in Indian context. Research done by Sinha and Banerjee (2004) on “Store
Choice Behavior in an evolving Market” gives deeper insight into Indian shopper
behavior. The analysis is based on seven constructs and 43 variables. This study
addresses the question of retail outlet selection, but it does a generic study of all modern
formats of retailing. So the seven factors justify the shopper behavior in outlet selection,
but don‟t specify the behavior for any specific outlet. The study by Goswami and
Mishra (2008) on traditional kirana stores visa-a-vis organized Supermarkets for
grocery shopping has identified factors related to grocery store patronage and identified
dimensions of customers‟ needs and desire which are relevant for grocery store choice. In
this study 44 attributes were considered that led to 11 factors. This research though based
on grocery store choice, the main objective is to differentiate Shopping in a kirana store
vs. in a supermarket.

4
The existing researches in India and abroad have identified many dimensions of Shopper
behavior in addressing the issues like retail branding, store location, differentiating kirana
store and organized retailing etc. But none of these researches address the question of
„shopper behavior in a supermarket‟ adequately. Either they talk of shopper behavior in
general (not specific to supermarket) or shopper behavior to specific stimuli
differentiating super markets from kirana stores. Contrary to above published research
findings, the proposed research is in between (neither too broad based, nor too specific)
that tries to address the triggers affecting shopper behavior in a super market, and studies
in specific shopper behavior in response to the triggers.

2.2 Formation of Hypothesis:

The reviews of literature gave deeper insights into the attributes affecting shopper
behavior, attributes and scales used by researchers and identify the gaps. Moreover this
helped develop hypotheses on the basis of understanding of shopper behavior in the retail
environment in India. The following hypotheses were proposed on the basis of literature
review:

H1: Shoppers go through a complex evaluation process before selecting a

retail outlet for purchase of merchandise.
H2: Shopper decision process is highly influenced by visual
merchandising and Point of Purchase Display.
H3: Shoppers are different in their buying behavior in selection of an
outlet.

3. Methodology

The literature review (both Indian and western) helped in creation of hypotheses for the
study. As the hypotheses were on shopper behavior, it necessitated to find out the
attributes that influence shopper buying behavior. On the basis of exploratory survey
twenty three attributes were identified and the same were discussed with subject experts
before finalization. Questionnaire was developed on identified attributes, and the data

5
was collected for pilot survey. SPSS was used to analyze the data (from pilot survey);
the tools used were Factor analysis and Cluster analysis to identify the factors and, then
to create shopper clusters respectively. The findings were presented and discussed with
subject experts. The pilot survey could clearly identify factors and also created distinct
clusters. On the basis of the pilot survey, minor modification was made to the
questionnaire; subsequently data was collected from 400 shoppers for analysis. The data
was subjected to factor analysis, to create constructs or factors that represent the twenty
three identified attributes. Subsequently Factor scores were created for individual
respondent from factor analysis that was consequently used for Cluster analysis to create
shopper clusters among the respondents. In addition to that, other pertinent numerical and
qualitative analysis was performed. The findings from the research were compared with
the hypotheses to draw relevant conclusions.

3.1 Exploratory Survey:

The main objective of exploratory survey was to understand different attributes that
influence shopper buying behavior in a supermarket. The attributes were first derived
from the literature survey (Sinha and Banerjee, 2004).The same were cross checked in an
Exploratory survey with 63 shoppers that reduced the original 43 attributes(Sinha and
Banerjee, 2004) to 23 key attributes for food and grocery segment. These attributes are:

1. Discounts offered at the outlet

2. Variety and assortment
3. Location of the store
4. Service level
5. Availability of parking facility
6. Store ambience
7. Availability of desired SKU
8. Home delivery facility
9. Status
10. Cleanliness/Hygiene
11. Brand choice

6
 12. Freshness of the merchandise
13. Merchandise display
14. Credit card acceptability
15. Behavior of salespeople
16. Adequate ventilation
17. Proper lighting
18. Price
19. Time to shop
20. Ease of shopping
21. Checkout
22. Signages
23. Loyalty programs

3.2 Primary Research:

The primary research was conducted on actual shoppers across outlets (supermarkets) in
Pune. The primary research was conducted over 54 shoppers on a pilot basis to validate
the questionnaire and the methodology, followed by which the actual primary research
was conducted with a valid sample size of 352 (out of total 400) where the unit of
analysis is the Shopper: the actual buyer who comes to the retail outlet to shop.
Convenience sampling was preferred over probability sampling in this case as the
objective of the research is to create shopper clusters on the basis of their behavior. So
creation of shopper clusters doesn‟t suffer from the disadvantages of convenience
sampling.

3.3 Determination of sample size:

For Factor analysis, studies have revealed that adequate sample size is partly determined
by the nature of the data (Fabrigar et al., 1999; MacCallum, Widaman, Zhang, & Hong,
1999). In general, the stronger the data, the smaller the sample can be for an accurate
analysis. “Strong data” in factor analysis means uniformly high communalities without
cross loadings, plus several variables loading strongly on each factor. The best method of
determining the sample size for Factor Analysis is subject to item ratio (Costello,

7
Osborne 2005). As per Costello and Osborne “In a majority of the studies (62.9%)
researchers performed analyses with subject to item ratios (ratio of respondents and
attributes under study) of 10:1 or less”. In the present study, data is collected from 400
customers out of which 352 are only considered for the final analysis. Rests 42 were
rejected because of inconsistency. At a subject size of 352 for 23 items under
investigation, the subject to item ratio stands at 15: 1 which is well above average among
all researches done using factor analysis.

3.4 Questionnaire development:

The questionnaire was developed on the basis of the 23 key attributes identified during
the exploratory survey. For each attribute, there was one question in the questionnaire
and shoppers were asked to choose between one to five for every question (one being of
least importance and five being extremely important to them in choosing a retail outlet).
In addition to the above, the questionnaire contained information regarding the
demographics (age, income, occupation) and behavior (frequency of purchase, monthly
expenditure) of shoppers. It also contained question on features that attract a shopper to
an outlet, the features that appeal to customers inside the store, what makes a shopper to
brand switch, selection of merchandise in an outlet, type of merchandise arrangement
preferred by a shopper, shopper view on store loyalty and brand loyalty, features that can
put a store on negative light, visual merchandising and commitment towards a brand etc.

An individual shopper being the sampling unit and the retail outlet being the sampling
frame, the 1st task in data collection was to list down the retail outlets from where the
data to be collected. The study involved a field survey conducted across different stores
in the city of Pune, in Maharastra. The respondents were approached at the shop after
they had finished shopping and were leaving the store. It was felt that shop intercept (exit
interviews) would capture the mindset of the shopper more effectively compared to an
interview away from the shop that might bring only “visualized perception” and not the
real experience. The respondents were administered a structured questionnaire. The
responses were recorded using a set of 23 statements measured on importance scale in
addition to other relevant information. The selection of retail outlets were done to ensure
coverage of most formats of supermarket in Pune.

8
3.5 Scope and limitation:
 The study pertaining to retail consumer behaviour is restricted to Staple
Merchandise only.
 Retail format covered in this study are restricted to supermarkets only.
 The primary data collection will be limited to retail outlets in Pune only.

4. Analysis

The analysis of the data revolves around two types a statistical analysis i.e. Factor
Analysis and Cluster Analysis. Factor analysis tries identifying the underlining constructs
that influence the responses on a number of measured variables. Factor analysis has been
used to group different items or attributes into common factors that influence shopper
buying behaviour. The data collected from respondents on the 23 key identified attributes
( on a scale of one to five) was converted to „Factors‟ or „Constructs‟ using Factor
analysis. So attributes or items that are close to each other in terms of respondents‟
opinion tend to group together into a common factor. Furthermore, „Factor Scores‟ were
created for each respondent from factor analysis. Factor scores represent the cumulative
response of a respondent for all the attributes that group into a factor. Factor scores were
used to create „Shopper Clusters‟ using cluster analysis. Shopper clusters represent
groups of respondent that are similar within the cluster and the respondents are different
across clusters in terms their response to the underlying factors or constructs. The various
demographic and behavioural data collected from the respondents have been used to
define clusters characteristics. The outcomes of Factor analysis and Cluster analysis
helped to test the proposed Hypotheses and draw relevant conclusions.

5. Results and Discussions:

5.1 Factor Analysis

The data collected from 352 respondents was subjected to factor Analysis using SPSS.
The initial challenge was to determine the number of factors. “Kaiser Criterion” and
“Scree test" were used to understand the underlining constructs and determine the
number of factors. The principal component analysis resulted in five factors explaining

9
83% of the total variation. Factors having eigenvalue more than one are considered for
analysis. The outcome of the Principal Component Analysis can be summarized as
follows:

 Factor-1 has an eigenvalues of 6.586 explaining 28.6% of variance

 Factor-2 has an eigenvalues of 4.373 explaining 19% of variance
 Factor-3 has an eigenvalues of 4.103 explaining 17.8% of variance
 Factor-4 has an eigenvalues of 2.473 explaining 10.75% of variance
 Factor-5 has an eigenvalues of 1.615 explaining 7% of variance

The 23 attributed in the questionnaire grouped into 5 factors. The rotated component
matrix below shows the factor loadings (correlation between the factors and the
attributes). In other words, the factor loadings describe the strength of relationship
between the factors and the attributes grouped under that factor. Any factor loading more
than 0.5 is supposed to be a high correlation, and considered for grouping purpose.
Moreover, the attribute having highest correlation with the factor is grouped under that
factor. For an example, the attribute „discounts‟ has a correlation of 0.707 with the factor
number three. This attribute has highest correlation with factor number three and the
correlation is more than 0.5. So, this is a valid inclusion and „discounts‟ belong to the
third factor. Likewise, the grouping of attributes was done under various factors.

Table: Rotated Component Matrix

Component/ Factor

1 2 3 4 5

Discounts -.176 -.098 .707 .184 .039

Varassort .921 -.048 -.048 -.140 .061
Location .055 .037 .934 .202 -.029
Service -.310 .073 .343 .834 -.035
Parking .041 .020 .922 .222 -.014
Ambience .944 -.083 -.060 -.122 .051

10
 Availability .132 -.009 .000 .014 .889
Homedelivery -.251 .087 .262 .854 -.021
Status -.086 .974 -.043 .032 -.020
Cleanliness .955 -.073 -.028 -.157 .019
Brandchoice -.053 .975 -.036 .040 -.014
Freshness .045 .037 .009 -.071 .890
Display .952 -.062 -.030 -.146 .037
Creditcard -.058 .884 .009 -.025 -.009
Salespeople .037 .028 .927 .195 -.044
Ventilation -.009 -.145 .631 -.101 .051
Lighting .945 -.044 -.034 -.147 .029
Price -.113 .007 .585 .150 -.011
Time to shop -.076 .977 -.034 .029 -.013
Ease of shopping .928 -.074 -.021 -.146 .041
Checkout -.307 .024 .275 .837 -.034
Signages .962 -.073 -.043 -.106 .066
Loyaltyprograms -.084 .812 -.065 .097 .081

Extraction Method: Principal Component Analysis.

Rotation Method: Varimax with Kaiser Normalization.
These factors or constructs describe all the attributes that have grouped into those
constructs. The importance of the constructs or factors is based on the percentage
variance explained by them.

1
Factor 1: Merchandising (28.635)

1. Variety and assortment (0.921)

2. Ambience (0.944)

1The values in bracket against factors explain the total variance explained by the factor, the values in
bracket against the attributes explain the factor loadings of the respective attributes

11
 3. Cleanliness (0.955)
4. Display (0.952)
5. Lighting (0.945)
6. Ease of shopping (0.928)
7. Signages (0.962)

Factor 2: Status/ Shopper Image (19.013)

1. Status symbol (0.974)

2. Brand choice availability (0.975)
3. Credit card(0.884)
4. Time to shop (0.977)
5. Loyalty Programme (0.812)

Factor 3: Convenience/Location (17.84)

1. Discount offered (0.707)

2. Convenient Location (0.934)
3. Parking space availability (0.922)
4. Friendly sales people (0.927)
5. Proper ventilation (0.631)
6. Price (0.585)

Factor 4: Service (10.75)

1. Services offered (0.834)

2. Home delivery (0.854)
3. Speedy check out (0.837)

Factor 5: Availability (7.002)

1. Availability (0.889)
2. Freshness (0.890)

As per the outcomes of the factor analysis, Factor 1 (Merchandising), consisting of seven
attributes, comes out as the most critical factor explaining 28.6% of the total variation.

12
The factor „Merchandising‟ consists of attributes like Variety and assortment, Ambience,
Cleanliness, Display, Lighting, Ease of shopping and Signages. The second factor
extracted in the Factor analysis can be named as Status or Shopper Image. The various
attributes that converged with this factor are Status, Brand Choice, Credit card
acceptance, Time to Shop and provision of Loyalty Programmes. This factor explains
19% of the total variance by grouping 5 different attributes. The third factor extracted in
the Factor analysis can be named as Convenience/Location. The various attributes that
converged with this factor are Discounts offered by the retailer, Location of the
Supermarket, Parking, number of Sales People, Ventilation and Price. This factor
explains 17.8% of the total variance by grouping 6 different attributes. As all these
attributes are linked to the convenience of the shopper and hence can be named as
“Convenience/Location‟. Price is the only attribute that doesn‟t seem suitable to be a part
of this construct. If we analyze the factor loadings, it becomes apparent that Price has a
very small factor loading 0.584), though significant, compared to all other attributes. The
fourth factor extracted in the Factor analysis can be named as „Service‟. The various
attributes that converged with this factor are quick and efficient service, Home Delivery
and Check out time. This factor explains 10.75% of the total variance by grouping 3
different attributes. The fifth factor extracted in the Factor analysis can be named as
„Availability‟. The various attributes that converged with this factor are Availability and
Freshness of products. This factor explains 7% of the total variance by grouping 2
different attributes. This also indicates that the 1st three factors collectively are
responsible for 65% of the total variation. Moreover all significant factors together (all
five) explain 83% of the variation. This indicated that 83% of the Shopper behavior can
be explained by these five factors where merchandising is the most critical followed by
Status and Location.

5.2 Hypothesis testing using Factor Analysis

Factor analysis examines how underlying constructs influence the responses on a number
of measured variables. Here the measured variables are the 23 attributes on which data
was collected from respondents. The 1st two hypothesis of the research are:
1. Shoppers go through a complex evaluation process before selecting a retail
outlet for purchase of merchandise.

13
 2. Shopper decision process is highly influenced by visual merchandising and
Point of Purchase Display

The output of the factor analysis indicated five underlying Constructs (referred as
Factors) as Merchandising, Status, Convenience, Service and Availability. This indicates
that there are 5 basic constructs that consumers evaluate while selecting an outlet for
purchase of staple merchandise signifying the acceptance of the 1st Hypothesis. The
hypothesized factor “Visual Merchandising” came out as one of the predicted factors
explaining 28% of the total variance and grouping 6 different attributes justifying the
acceptance of 2nd Hypothesis.

Though the factor analysis justifies the Hypotheses, before accepting the Hypotheses the
reliability of the factors can be verified by Cronbach‟s alpha. The value of Cronbach‟s
Alpha indicates the total variance explained by the grouped attributes in a factor
compared to the total variance. The Value of Cronbach‟ Aplha more than 0.7 is
considered as strong enough to justify the association among the attributes converged to a
factor. The table below indicates the SPSS Output indicating the values of Cronbach‟s
Alpha for the five extracted factors justifying the reliability and statistical significance.

Table: Reliability test Summary

Factor/Construct Name Cronbach‟s Alpha

Merchandising 0.985
Status 0.959
Convenience 0.883
Service 0.937
Availability 0.753

Cronbach‟s Alpha indicates very high level of reliability (more than 0.85) for the 1st four
constructs (Merchandising, Status, Convenience, Service) and relatively week but
significant reliability for the 5th construct (availability).

14
Moreover, the extracted five factors explain more than 83% of the Variance
cumulatively.

Total variance explained and the Reliability test support the five constructs emphasizing
that shoppers evaluate a retail outlet based on five constructs while making purchase
decision. This explains the acceptance of 1st Hypothesis that “Shoppers go through a
complex evaluation process before selecting a retail outlet for purchase of merchandise”.
Out of the five constructs or factors extracted in factor Analysis, Merchandising is one of
the most critical explaining 28% of Variance and having a Cronbach‟s Alpha of 0.985
supports the acceptance of the 2nd hypothesis that “Shopper decision process is highly
influenced by visual merchandising and Point of Purchase Display”.

5.3 Cluster Analysis:

Cluster analysis is a collection of statistical methods, which identifies groups of samples

that behave similarly or show similar characteristics (Nethra Sambamoorthi, 2003). So if
the cluster analysis segregates the respondents into clear clusters or groups, it can be
concluded that shoppers are not uniform in their behavior, rather there exists clear
groups that are similar within while differ across groups in their shopping behaviour.

In this research, to create clusters, factor scores have been used instead of Individual
attributes. Factor scores represent a shopper‟s response to all attributes captured in a
factor. In this research shopper clusters have been created on the basis of factor scores
derived from the five identified factors (from factor analysis). Cluster analysis was run in
SPSS using Hierarchical clustering Technique. The dendogram plot (cluster structure) in
Hierarchical clustering indicated presence of five clusters in the data. To confirm the
same, cluster analysis was repeated using K Mean Clustering (non-hierarchical). The
table below is the SPSS output for K Mean clustering indicating the Dominant factors in
different cluster centers.

15
Table: Final Cluster Centers
Cluster
1 2 3 4 5
REGR factor score 1
.82718 .44558 -1.02444 -.62002 .20061
(Merchandising)
REGR factor score 2
-.56660 -.20776 -.32901 -.15176 1.66809
(Status)
REGR factor score 3
-.50576 .42045 -.40748 1.95718 -.46843
(Convenience)
REGR factor score 4
.49030 -.81406 -.24658 1.07720 .28937
(Service)
REGR factor score 5
.32411 -.60724 .39603 -.21783 -.02686
(Availability)

The cluster analysis resulted in five clusters or five shopper segments. The table above
(Final Cluster Centers) indicates the dominance of different factors in different clusters.
The 1st cluster is dominated by the factor „Merchandising‟ (with a value of 0.82718); and
the factor „Availability‟ (with a value of 0.32411) has least importance in cluster-1. The
importance of the factors for cluster-1 goes down in the following order of the factors-
Merchandising, Status, Convenience, Service and Availability. The following table
indicates the relative importance of different factors on different clusters.
Table: Relative importance of the Factors on different clusters:
Different factors in the decreasing order of their importance
Cluster-1 Merchandising Status Convenience Service Availability
Cluster-2 Service Availability Merchandising Convenience Status
Cluster-3 Merchandising Convenience Availability Status Service
Cluster-4 Convenience Service Merchandising Availability Status
Cluster-5 Status Convenience Service Merchandising Availability

The respondents are distributed across the clusters. As per the SPSS output, the 1st
cluster has 79 respondents, 2nd cluster has 87, 3rd cluster has 92, 4th cluster has 35 and 5th
cluster has 59 respondents.

16
 5.4 Hypothesis testing using Cluster Analysis

Cluster Analysis has been used to test the third hypothesis i.e. “Shoppers are different in
their buying behavior in selection of an outlet”.

Cluster Analysis clearly defines 5 clusters or shopper segments. The clusters are different
among them but similar within each cluster in terms of their buying behavior. The
buying behavior of the shoppers has been studied based on the factors extracted in factor
Analysis. Different clusters are centered as per the importance given to the factor by the
shoppers in the respective clusters. Though the cluster analysis justifies the Hypotheses,
accepting the Hypotheses will depend upon the statistical significance of the outcome of
the cluster Analysis. The validity of the clusters can be tested for statistical significance
using Anova table below.

ANOVA- Cluster Analysis

Sum of Mean
Squares Df Square F Sig.
REGR Between
factor Groups 183.708 4 45.927 95.263 0.00
score 1 Within
Groups 167.292 347 0.482
Total 351 351
REGR Between
factor Groups 204.051 4 51.013 120.459 0.00
score 2 Within
Groups 146.949 347 0.423
Total 351 351
REGR Between
factor Groups 197.879 4 49.47 112.107 0.00
score 3 Within
Groups 153.121 347 0.441
Total 351 351
REGR Between
factor Groups 127.793 4 31.948 49.667 0.00
score 4 Within
Groups 223.207 347 0.643
Total 351 351
REGR Between
factor Groups 56.512 4 14.128 16.647 0.00
score 5 Within
Groups 294.488 347 0.849
Total 351 351

17
The statistical significance has been tested for all the five factors individually. The
Anova table indicates that for all the five factors clusters are different from each other at
5% level of significance. So, the Anova table confirms that the clusters are statistically
significant from each other and hence, it can be concluded that there exist 5 different
shopper clusters or segments that are different in terms of their buying behavior.

6. Managerial Implications

The research findings can be used by organized retailers for target marketing strategy.
Retailers especially super market owners can reorient their strategy as per the cluster
targeted. The cluster characteristics can be summarized as below:

Table: Cluster Characteristics

Clusters Import- Percenta- Age Income Shopping Monthly

ant ge of the group Frequency expense
Factor sample majority
Cluster 1 Mercha 22% 18-34 15,000 to Twice or 2001 to
ndising 35,000 more than 3000
twice a
week
Cluster 2 Service 25% 25-44 10,000 to Once a 2001-
25,000 week 5000
Cluster 3 Mercha 26% 18-44 35,000 to twice a 3001-5000
ndising 60,000 week
Cluster 4 Conven 10% 35-60 25,000 to Once a 3001-
ience 45,000 week 5000
Cluster 5 Status 17% 25-60 45,000 to 2-3 times a 3001 to
100,000 month 10,000

Apart from factor analysis and cluster analysis, other relevant analysis was done using the
data collected. Majority of shoppers are of the opinion that lighting and cleanliness,

18
product display, well stocked store and display attract a shopper to enter the outlet.
Retailers can‟t afford to compromise on these attributes. Shopper preference in terms of
display of merchandise was a part of the questionnaire. A big majority of the shoppers
(78%) are of the opinion that they prefer products displayed product-wise/category-wise
rather than brand wise. But it also reflects that there are 22% customers who prefer to see
the products displayed brand wise. These 22% may be the customers who are more brand
loyal and would like to stick to their brand choice in multiple product segments. In such a
research on shopper behavior, it is necessary to understand shoppers who are loyal to the
store compared to the ones loyal to brand. In this case only 31% of the shoppers say that
they will look for the product in a different store if it is not available in the existing store.
However, 69% of them don‟t mind switching the brand and picking up the next preferred
brand. So, store loyalty is quite strong compared to brand loyalty in low involvement
products like staple merchandise in supermarkets. As a vast majority of the shoppers are
of the opinion that they switch brands in case of non availability, it is imperative to
understand what makes them switch the brand. Shoppers were asked to choose what
makes them opt for a different brand. The outcomes suggest that non availability of the
choice brand, promotional schemes and good display are the reasons that may induce
brand switch among shoppers. Shoppers were asked what keeps a store in negative light.
The response of the shoppers of the features that puts the store in negative light can be
bad layout having less space between rows that creates inconvenience in shopping,
improper lighting and absence of directional messages. So retailers in addition to
develop segment specific target market strategy should be careful in deciding the various
offerings that attract shoppers to the outlet so as to increase the footfall to induce
purchase.

7. Conclusion

The analysis of the shopper response indicates five factors that influence shopper buying
behavior. These factors can be named as Merchandising, Status/Shopper Image,
Location/Convenience, Service and Availability. These five factors or constructs measure
the influence of all the 23 identified attributes or variables. These five factors are
statistically significant and explain more than 83% of the total variance. The five

19
constructs or factors justify the acceptance of the 1st Hypothesis. The constructs explain
that shoppers look at five specific factors while evaluating an outlet and making their
shopping decision. The Factor Merchandising consists of attributes like Variety and
assortment, Ambience, Cleanliness, Display, Lighting, Ease of shopping and Signages.
This factor explained more than 28% of variance and is the most critical among the five
factors justifying the acceptance of 2nd hypothesis. As the factor analysis explained the
basic constructs in shopper buying behavior, Cluster analysis explained that shoppers are
divided among themselves in terms their behavior around those five factors. The
shoppers got divided into five clusters as per the importance they attach to different
factors. The anova table justified the presence of five distinct clusters and hence it divides
the shoppers into five segments. These segments or shopper clusters behave in unique
ways; the shoppers within a cluster are similar in their behavior while shoppers in
different clusters are different from each other. So apart from justifying the acceptance
of the third Hypotheses i.e. Shoppers are different in their buying behavior in selection of
an outlet, the results bring out that retailers can no longer assume that all shoppers are the
same. Rather different and unique retailing strategy should be used to appease different
clusters of shoppers.

So based on the above results and discussions it can be concluded that, “Shoppers go
through a complex evaluation process before selecting a retail outlet”. The evaluation
process is influenced by five factors namely Merchandising, Status/Shopper Image,
Convenience/Location, Service and Availability. Out of these five factors, Merchandising
is the most important criterion grouping seven attributes. Hence retailers must keep this
in mind while designing the layout and merchandise presentation inside the outlet. When
shopper behaviour was analysed using factor scores, the SPSS output indicated presence
of five distinct clusters and hence can be concluded that “Shoppers are different in their
buying behavior in selection of an outlet”. Retailers must understand the nature and
behavior of these clusters or shopper segments and design market targeting strategy.

###

20