Analysis of Customer Characteristic towards Private Label
in Retail Industry using Data Mining Isti Surjandari 1 , Afrizal Juansyah 2
Industrial Engineering Department, University of Indonesia, Depok 16424 Tel: (021) 78888805. Fax: (021) 78885656 Email: 1 isti@ie.ui.ac.id, 2 afrizal_tiui09@yahoo.com
Abstract Private label demand will be likely to get higher as the consumers become more and more selective in buying the products of high quality with an optimal price. However, in Indonesia, for some people, the traditional perceptions of private label were once of low quality and just attracting the most cost- conscious consumers. Considering this, retailers need to pay more attention to private label products in facing the competition. The current study discusses consumers attitude and satisfaction with respect to each category of private label products. Frequencies of purchase and consumer characteristics are also discussed. Using Data Mining with C5.0 algorithm, the research resulting decision tree model that contains a classification of PL customers based on their frequency of purchase.
1. INTRODUCTION The development of more advanced retail industry lately has caused type, quality, and cost of goods sold to be more varied. Often the consumers find it difficult to distinguish goods of medium quality and high quality only from his physical appearance only, especially if the product is packaged in an attractive packaging. To avoid disappointment due to misinterpretation in choosing items, many consumers prefer goods produced by the producer who has been known and has brand popularity, which often appear in the advertisement to assure the quality of the product. Viewing such phenomena, some retailers in Indonesia try to pack their product in their own brand packaging which is commonly referred as PL (Private Label). This is certainly done with the considerations based on the good company image that has been sustainable enough and known by the customers. The majority of consumers stated that the decision made on buying products relies on cheap prices, good quality, display position, the retailer brand, and the packaging at last. In addition, the prospect of the retail industry is looking good in the future ahead. The estimation until the end of the year reaches approximately 16-17%. Modern retail is now becoming the driver of the economic growth. It has been constantly developed in improving the quality based on the consumers appreciation. One big obstacle that the private label in Indonesia has been facing is the misinformation of secondary quality of the product they produce, which refers to second class consumers who want to save money in buying goods. This leads to the customers unwilling to buy and use the private label products. In addition, the phenomena could be related to the lack of education and understanding in the market society about the private label in developing country. As for instance, Malaysia and Taiwan have half of their respondents agreed that they know only a few about retail brands to try the products. To face the challenge, the retail company requires a precise and strong marketing strategy while still paying attention to the characteristics and behaviors of the developing market of private label segmentation. Based on the previous discussion, this research is purposed to understand the characteristic and behavior of private label customer and the level of satisfaction in each category that influence the frequency of PL purchasing using data mining method. The data 2
used in this research is the primary data collected from direct and online questionnaire. The steps done in the research are divided into three basic stages; initial identification, data collecting and processing, analysis and conclusion. At the first stage, the initial identification is to sort out the basics of the research which involve the topic, purpose, and the research theory. The collecting and processing data is the second stage begins with determining the data needed up to processing it to its final data form. The last stage will require analysis to the processed data and to give the final conclusion, with the suggestion added at the end of the research. 2. THEORY
2.1 Definition of Data Mining Data mining is the method used in analyzing excessive number of data to understand the hidden connections inside and summarize them into a new understandable and useful data for the reader [2]. Basic data applications had been used in several fields including management, trade industry data management, education, government administration and others. The problem is often found in processing the data is the difficulty in finding the useful information within the huge number. Therefore, this method has become the answer to the problem. 2.2 Classification Classification is a data mining function that assigns items in a collection to target categories or classes. The goal of classification is to accurately predict the target class for each case in the data. A classification task begins with a data set in which the class assignments are known. Predictor variable is a variable on which to base consideration in classifying the data into classes. To achieve its objectives, the classification has algorithms. The algorithm itself is a set of commands or instructions to solve a problem. 2.3 Decision Tree Decision tree model is a method of classification in data mining that uses a tree structure representation where each node represents an attribute, branches represent the value of the attribute, and the leaves represent the class. Decision tree is the most popular method of classification used. In addition to the relatively rapid formation, the result of the model that was built was easy to understand [3]. In the decision tree, there are three types of nodes that have their respective functions. The following is an explanation of the three nodes: a. Root Node
As the topmost node in decision tree, this node has no input and zero or more output.
b. Internal Node
A branching node, this node has only one input and has an output of at least two attributes.
c. Leaf Node
An end node, at this node, there is only one input and has no output. This node represents class of classification.
The advantage of decision tree model is its capability to break down process of complex decision-making becomes much simpler so that decision-makers will be easier to interpret the solution of the problem.
2.4 C5.0 Algorithm
Algorithm is a set of commands or instructions to solve a problem. C5.0 algorithm is one of the decision tree algorithms. Decision tree formed by this algorithm has non-binary branches. This algorithm was developed by Ross Quinlan. C5.0 algorithms compute information gain of each predictor variable. Predictor variables that have the greatest information gain were 3
chosen to be the first branch level. At the first level of branches, predictor variables that have biggest information gain are chosen to split into two branches at the level below the first level branches. The step is repeated continuously until all predictor variables split into branches. C5.0 uses classification model in the form of if-then to classify the data items that have numeric and categorical attributes. In general, the stages in the building process of the algorithm C5.0 decision tree is as follows: 1) At the initial stage, the decision tree is described as a single node representing the training set. 2) If the sample contains all of the same class, then the node is a leaf and is labeled with the name of the class. 3) If not, the algorithm using information gain- based measure will choose predictor variables that will separate the records into individual classes. Variable will be the test or decision variable at the node. 4) Branch developed for each known value of the test variable, and the samples were separated based on the branch. 5) The algorithm uses the same process recursively to form a decision tree
3. Data Processing 3.1 Data Collecting This section will explain the stages in the process of data collection that consists of the variable data determination, the preparation of the contents, and method of distributing questionnaires. 3.1.1 Research Variable Determinating Below are the 28 data variable used in the research: Table 3.1 Research Variable Category Item Choices Basic Information Gender Male; Female Age <=25; 26-35; 36-45; 46-55; >55 Occupation Students; Housewives; Entrepreneurs; Public sector employees; Private sector employees; Other Education level SMA; D3; S1; S2; S3; Other Marriage status Single; Married Monthly Income <= Rp 2.000.000; Rp 2.000.001 - Rp 5.000.000; Rp 5.000.001 - Rp 10.000.000; Rp 10.000.001 - Rp 15.000.000; Rp 15.000.001-Rp 25.000.000; > Rp 25.000.000 PL Purchasing Behavior Frequency of PL purchasing Never buy PL; Ever only once bought PL; Sometimes buy specific PL products, Always buy specific PL products Most frequently bought PL brand Carrefour; Hypermart; Giant; Alfamart; Indomaret; Never; Other Reason of not buying PL Not interested in PL, Less attractive packaging, Lack of trust in quality of PL, Not aware of PL Reason of buying PL Low and competitive price, Good quality, Promotion, catalog, etc, Good packaging, Other product planned to purchase is not in stock Satisfaction rate towards each category of PL Grocery products Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Snack Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Toiletries and household products Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Houseware products Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Cottons and wipes Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Clothes and Footwares Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Home appliances and electronic products Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Stationery Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Satisfaction rate towards PL Price Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Quality Dont know; Very unsatisfied; Unsatisfied; Satisfied; Very satisfied Purchasing Behavior I think that buying the product with the lowest price is important Extremely disagree; Disagree; Neutral; Agree; Extremely agree I am willingly spending more time to find the cheapest price Extremely disagree; Disagree; Neutral; Agree; Extremely agree I always look and compare the price of the products first before buying Extremely disagree; Disagree; Neutral; Agree; Extremely agree I chose price as the primary consideration in choosing products Extremely disagree; Disagree; Neutral; Agree; Extremely agree I think that getting the best quality products is essential Extremely disagree; Disagree; Neutral; Agree; Extremely agree I don't want to sacrifice expected quality just to get cheap prices Extremely disagree; Disagree; Neutral; Agree; Extremely agree I think that the quality of PL and NB is not significantly different Extremely disagree; Disagree; Neutral; Agree; Extremely agree I chose quality as the primary consideration in choosing products Extremely disagree; Disagree; Neutral; Agree; Extremely agree
4
3.1.2 Drafting the Questionairre Content The content of the questionnaire contains 5 different subdivision based on the customer characteristics such as; personal information, behavior in purchasing private label, satisfaction level towards specific private label brand, satisfaction level on price and quality of private label product, and the customer behavior in shopping. The data was collected by distributing questionnaires either in person or online to respondents. Direct data collection is done in a minimarket outlets and hypermarkets around Jakarta and Depok on May 20 th to June 17 th 2013. After the questionnaires have been filled out completely and correctly, they will then be used in the data processing stage. There are 518 questionnaires that have been collected at the time of this data collection process is completed. After going through a stage of elimination in missing datas, there were 508 questionnaires that were subsequently entered into the data processing stage. The data is converted into a format Ms. Excel, and then imported into the program RapidMiner 5.0. 3.2 Data Processing 3.2.1 Customer classification with Algorithm C5.0 in Data Mining Before the customer classification is done by forming a decision tree, it is first required to define variables and objective predictor variables adapted to the purpose of this study. The purpose of this study was to determine the classification of the characteristics and behavior of customers shopping the retail market and the level of satisfaction in each product category based on PL intensity of product purchase. So in this study, the target variable is the intensity of PL purchases. Meanwhile, the predictor variables are variables which serve as a reference to classify the data into classes. It can be seen in Table 3.1, the predictor variables in this study where 27 variables were classified into five major groups of customer characteristics. After determining the target variable and the predictor variables in this study, the decision tree model can be formed using the software RapidMiner 5.0. RapidMiner is open source software for knowledge discovery and data mining. RapidMiner has approximately 400 procedures (operator) data mining, including the operator to input-output, and visualization of data preprocessing. Thousands of data mining applications that have been developed using RapidMiner is widely used in business and research. In RapidMiner, there are elements of a collection called operator command to create a model that consists of Process Control, Utility, Repository Access, Import, Export, Data Transformation, Modeling, and Evaluation. To make a decision tree model, the operator used is Retrieve, Validation, Decision Tree, Apply Model, and Performance. In the Main Process column, there are operators Retrieve and Validation. Retrieve operator is an operator that is used to read the data by a collection of objects from the document storage. This operator is a collection of data obtained from the questionnaire which further serve as input. While the operator separates Validation randomly input the data which come out as the training data and the testing data, and then eventually these operators evaluate the model. In the training field, there is also the operator Decision Tree is used to make the classification decision tree model of nominal data. While on testing field, there is an Apply Model operator used to apply the input data set into the model and the Performance operator displays a list of performance values of a collection of attributes. On the following page there is a display of the data processing classification of Private Label consumer in the form of a decision tree model.
5
4. ANALYSIS AND DISCUSSION OF FINDINGS Analysis of the retail customer characteristics classification also refers to the results of the decision trees using RapidMiner 5.0. Based on leaf/terminal node previously mentioned, there are 15 nodes that are characteristic of customers who ever only once buy PL, 23 nodes for customers who sometimes buy specific PL, 23 nodes for customers who always buy specific PL, and the last 4 nodes for customers who never buy PL. These nodes were chosen not only because it serves as a leaf/terminal node but also because it has a high market share. Out of the total 518 questionnaires distributed 508 were usable. 60% of the respondents were women due to the fact that mostly women buy groceries at the minimarket or hypermarket. In terms of the age and marriage status, the majority of the respondents (77.5%) was under 35 years old and was single (60.2%). It is caused by culture or habit (collectivist Culture) of Indonesian young/productive people who do activities in groups, as well as shopping. In terms of education level, the majority of respondents have S1 (Strata Satu) level (52%) which mostly they works as private sector employees (34.6%) with monthly income between Rp 2.000.000 and Rp 5.000.000 (41.5%). Education level indicates amount of knowledge owned by the respondents that could affect buying decision. In the other hand, the proportion of housewives in the sample was relatively low (8.5%) which the majority of them stated that they sometimes buy specific PL (53.5%) and always buy specific PL (25.6%). Based on decision tree analysis (Figure 4.1), majority of consumers who always buy specific PL stated that good quality of PL is the reason of purchasing those products. As the highest market share of this class (0.98%), they stated that they are fully satisfied toward grocery product and home appliances category of PL. Therefore, to maintain this class of consumers to be loyal, the retailer have to do a strategy that focuses on that both category, by adding variant of products and doing product innovation as well as guarantee the best quality. Consumer class "sometimes buy a specific PL" is the class with the most respondents (66.7%). With the highest market share, the majority of this class (42.72%) chose low and competitive price of PL as a key element that encourage them to buy PL. The magnitude of this class segment indicates the amount of opportunities for the retail market to make the private label consumers into loyal customers by assuring the quality as the key value of PL. Consists of 23 nodes, this class has quite complex characteristics. There is a 3.74% market share with the characteristics of customers who are single and stated that they are satisfied toward grocery product of PL because of its good quality. In consumer class of ever only once buy PL, majority of them (29.5%) chose Indomaret as most frequently bought PL brand. Most of them (48.4%) stated that they bought PL because other brand product planned to purchase was not in stock. Highest market share of this class is worth 4.13% of the total respondents is consumer who is single and no stock of other product planned to purchase as the main reason that encourage them to buy PL. They stated that they are not willingly to spend more time just to find the cheapest products. To anticipate that, retailers ought to ensure that display of their product could attract customer spontantly. Efforts that can be implemented is designing the packaging as attractive as possible and put the PL display in golden eye level area which is an area of 40-60 degrees from the human vision (assuming height of Indonesian adult consumers are between 150-170 cm) so it tends to first awared by consumers when standing in front of the rack. Consumer class of never buy PL that has proportion of 15.5% of total respondents has highest market share worth of 5.71% that stated that the reason of not buying PL was lack of trust in quality of that product. 6
A B C A B C Figure 4.1 Decision Tree of Retail Consumer Classification
7
T a b l e
4 . 1
C u s t o m e r
S e g m e n t a t i o n
b a s e d
o n
F r e q u e n c y
o f
P r i v a t e
L a b e l
P u r c h a s i n g
F r e q u e n c y
o f
P L
p u r c h a s i n g M a r k e t
S h a r e G e n d e r A g e O c c u p a t i o n M a r r i a g e
S t a t u s R e a s o n
o f
n o t
b u y i n g
P L R e a s o n
o f
b u y i n g
P L M o s t
f r e q u e n t l y
b o u g h t
P L
b r a n d S a t i s f a c t i o n
t o w a r d s
g r o c e r y
p r o d u c t s
( P L ) S a t i s f a c t i o n
t o w a r d s
s n a c k
c a t e g o r y
( P L ) S a t i s f a c t i o n
t o w a r d s
h o u s e w a r e
p r o d u c t s
( P L ) S a t i s f a c t i o n
t o w a r d s
h o m e
a p p l i a n c e s
a n d
e l e c t r o n i c
p r o d u c t s
( P L ) S a t i s f a c t i o n
t o w a r d s
s t a t i o n e r y
( P L ) S a t i s f a c t i o n
t o w a r d s
p r i c e
o f
P L S a t i s f a c t i o n
t o w a r d s
q u a l i t y
o f
P L I
a m
w i l l i n g l y
s p e n d i n g
m o r e
t i m e
t o
f i n d
t h e
c h e a p e s t
p r i c e I
a l w a y s
l o o k
a n d
c o m p a r e
t h e
p r i c e
o f
t h e
p r o d u c t s
f i r s t
b e f o r e
b u y i n g I
c h o s e
p r i c e
a s
t h e
p r i m a r y
c o n s i d e r a t i o n
i n
c h o o s i n g
p r o d u c t s I
t h i n k
t h a t
g e t t i n g
t h e
b e s t
q u a l i t y
p r o d u c t s
i s
e s s e n t i a l I
d o n 't
w a n t
t o
s a c r i f i c e
e x p e c t e d
q u a l i t y
j u s t
t o
g e t
c h e a p
p r i c e s I
t h i n k
t h a t
t h e
q u a l i t y
o f
P L
a n d
N B
i s
n o t
s i g n i f i c a n t l y
d i f f e r e n t 4 , 1 3 % M a r r ie d P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k 1 , 7 7 % P r o m o t io n ,
c a t a lo g ,
e t c S a t is f ie d D is a g r e e 1 , 3 8 % S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k U n s a t is f ie d A g r e e 1 , 1 8 % S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k S a t is f ie d S a t is f ie d A g r e e A g r e e 1 , 1 8 % S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k U n s a t is f ie d S a t is f ie d U n s a t is f ie d A g r e e A g r e e 0 , 9 8 % M a le G o o d
p a c k a g in g U n s a t is f ie d 0 , 7 9 % F e m a le G o o d
q u a lit y D o n 't
k n o w S a t is f ie d 0 , 7 9 % S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k N e u t r a l 0 , 5 9 % G o o d
q u a lit y D o n 't
k n o w A g r e e 0 , 5 9 % P r o m o t io n ,
c a t a lo g ,
e t c U n s a t is f ie d U n s a t is f ie d 0 , 3 9 % P r iv a t e
s e c t o r
S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k S a t is f ie d N e u t r a l A g r e e 0 , 3 9 % S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k D o n 't
k n o w S a t is f ie d U n s a t is f ie d A g r e e A g r e e 0 , 3 9 % M a le S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k S a t is f ie d D is a g r e e A g r e e 0 , 3 9 % P r o m o t io n ,
c a t a lo g ,
e t c D o n 't
k n o w U n s a t is f ie d N e u t r a l 0 , 3 9 % < =
2 5
P r o m o t io n ,
c a t a lo g ,
e t c U n s a t is f ie d A g r e e N e u t r a l 4 2 , 7 2 % L o w
a n d
c o m p e t it iv e
p r ic e 3 , 7 4 % S in g le G o o d
q u a lit y S a t is f ie d 2 , 7 5 % G o o d
q u a lit y U n s a t is f ie d N e u t r a l 2 , 5 6 % P r o m o t io n ,
c a t a lo g ,
e t c U n s a t is f ie d A g r e e A g r e e 2 , 1 6 % M a r r ie d G o o d
q u a lit y S a t is f ie d D o n 't
k n o w 1 , 7 7 % F e m a le G o o d
p a c k a g in g 1 , 5 7 % G o o d
q u a lit y D o n 't
k n o w D o n 't
k n o w 1 , 5 7 % G o o d
q u a lit y U n s a t is f ie d D is a g r e e 0 , 9 8 % M a r r ie d G o o d
q u a lit y S a t is f ie d S a t is f ie d N e u t r a l 0 , 5 9 % M a r r ie d G o o d
q u a lit y C a r r e f o u r S a t is f ie d S a t is f ie d A g r e e N e u t r a l 0 , 5 9 % S t u d e n t s S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k S a t is f ie d N e u t r a l A g r e e 0 , 5 9 % F e m a le S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k S a t is f ie d D is a g r e e A g r e e 0 , 5 9 % S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k D is a g r e e 0 , 5 9 % P r o m o t io n ,
c a t a lo g ,
e t c S a t is f ie d N e u t r a l 0 , 5 9 % P r o m o t io n ,
c a t a lo g ,
e t c D o n 't
k n o w U n s a t is f ie d U n s a t is f ie d N e u t r a l 0 , 5 9 % 2 6 - 3 5
P r o m o t io n ,
c a t a lo g ,
e t c U n s a t is f ie d A g r e e N e u t r a l 0 , 3 9 % M a le G o o d
p a c k a g in g D o n 't
k n o w 0 , 3 9 % M a le G o o d
q u a lit y D o n 't
k n o w S a t is f ie d 0 , 3 9 % M a r r ie d G o o d
q u a lit y G ia n t S a t is f ie d S a t is f ie d A g r e e 0 , 3 9 % S in g le P r o d u c t
o f
o t h e r
b r a n d
p la n n e d
t o
p u r c h a s e
is
n o t
in
s t o c k S a t is f ie d S a t is f ie d U n s a t is f ie d A g r e e A g r e e 0 , 3 9 % 4 6 - 5 5
P r o m o t io n ,
c a t a lo g ,
e t c U n s a t is f ie d A g r e e N e u t r a l 0 , 3 9 % P r o m o t io n ,
c a t a lo g ,
e t c U n s a t is f ie d A g r e e D is a g r e e 0 , 3 9 % P r o m o t io n ,
c a t a lo g ,
e t c U n s a t is f ie d D is a g r e e 0 , 9 8 % M a r r ie d G o o d
q u a lit y C a r r e f o u r S a t is f ie d S a t is f ie d A g r e e A g r e e 0 , 5 9 % M a r r ie d G o o d
q u a lit y S a t is f ie d U n s a t is f ie d 0 , 3 9 % M a r r ie d G o o d
q u a lit y A lf a m a r t S a t is f ie d S a t is f ie d A g r e e 0 , 3 9 % P r o m o t io n ,
c a t a lo g ,
e t c S a t is f ie d A g r e e 5 , 7 1 % L a c k
o f
t r u s t
in
q u a lit y
o f
P L 5 , 7 1 % N o t
in t e r e s t e d
in
P L 3 , 1 5 % N o t
a w a r e
o f
P L 0 , 9 8 % L e s s
a t t r a c t iv e
p a c k a g in g S o m e t im e s
b u y
s p e c if ic
P L
p r o d u c t s A lw a y s
b u y
s p e c if ic
P L
p r o d u c t s N e v e r
b u y
P L E v e r
o n ly
o n c e
b o u g h t
P L 8
With the same number of market share, cutomers also stated that they are not interested in private label. It is because for most consumers, brand is a guarantee of good quality, which makes hard for them to trust other products not familiar with like this PL. And also the perception stated that the PL is only for consumers who are budget-conscious and can not afford to buy a similar product that is much more expensive (National Brand). Economic conditions of someone really affect the consideration in product selection and purchasing decision on a particular product [3]. After analyzing the relationship between consumer demographics and the level of satisfaction towards each category of PL, obtained an indication that there is a pattern of inconsistent levels of satisfaction in each of several categories of PL consumer groups. For example, PL category of home appliances and electronic product has a tendency to depend on age. Consumers over the age of 35 years old tend to be satisfied towards that category, while consumers aged less than 35 years old tend to be unsatisfied. Another example, the PL category products of hoousewares seems to depend on the marriage status and monthly income. Consumers who are married are less likely to be satisfied in the category of PL products, while consumers single fully satisfied. Likewise, the quality of the housewares category tend to meet the expectations of consumers who have an income less than or equal to Rp 2,000,000, but consumers who have more than Rp 2,000,000 was unsatisfied on this category. 5. CONCLUSION At the end of the study, it is obtained a decision tree model which represents the classification based on the frequency of PL purchasing, ranging from never to always buy certain PL products. Additionally, there is a pattern of inconsistent level of satisfaction in each of several categories of PL consumer groups. For example, PL product category of home appliances and electronic products tends to depend on age, while housewares category depends on marital status and the monthly income. Results of the questionnaire showed that the category with the highest level of satisfaction and awareness is grocery products and cottons & wipes products. This is useful if retailers plan to launch the new PL products then, they should prioritize that both categories. REFERENCES [1] Boutsouki, Christina. (2008). Consumer Behaviour towards Own Label: Monitoring the Greek Experience. Agricultural Economics Review Vol. 9 No. 1. [2] Hand, D., Mannila, H., & Smyth, P. (2001). Principles of data mining. Cambridge, MA: MIT Press. [3] Kotler, Philip. (2005). Marketing Management The Millenium Edition. New Jersey: Prentice Hall Internation, Inc. [4] Lacrose, Daniel T. (2005). Discovering knowledge in data : an introduction to data mining. New Jersey : Jhon Wiley & Sons, Inc.