Description
This research is to focus on consumer buying behaviour or customer preferences of different brands of shampoos from various segments prevailing in market and thus finding out the opportunities for new entrants.
MARKET RESEARCH PROJECT
[Final Report]
[Usage pattern of Shampoo among female students at SIC campus]
Executive Summary:
The main objective behind this research was to focus on consumer buying behaviour or customer preferences of different brands of shampoos from various segments prevailing in market and thus finding out the opportunities for new entrants. The research was conducted at Symbiosis Infotech Campus, Hinjewadi, Pune to determine the shampoo purchase behaviours of female students in the age group of 20 to 30. The research was conducted on a sample size of 100 respondents. The population was deviden into brand loyals nad normal customers and hence any new onsumer can be put in the corresponding category based on his characteristics.After finding out the main factors influencing the buying behaviour of female students and segmenting this market into two segments i.e. experiencers and makers, we found out gap in the current market and concluded that the new product should be targeted at experiencers. It should be medium-priced (Rs. 50-60 for 100 ml.), should be mild on hair and should give shine to the hair as other attributes are already present in the shampoos available in the market.
Contents
Executive Summry: ......................................................................................................................... 2 toc.................................................................................................................................................... 4 Introduction: .................................................................................................................................... 5 METHODOLOGY AND LIMITATIONS:-................................................................................... 7
toc
Introduction:
(Source: Internet and AC Neilson Report) Background: The Indian shampoo industry is estimated at Rs. 2141 cr and is growing at an average rate of 20% per annum. According to AC Nielsen, shampoo is one of the fastest growing categories within FMCG sector and is expected to grow at 25% per annum in the coming years. From a penetration level of 13 per cent in 2000 to 31.9 per cent in 2005, to now almost a third of the country's rural population uses shampoo with penetration levels zooming at 36 per cent in 2008. Urban markets account for 80% of the total shampoo market.
?
Segmentation:
The shampoo industry is segmented on benefit platforms: - Cosmetic (shine, health and strength) - Anti Dandruff (AD) and - Herbal 20% of the total shampoo market is accounted by the AD shampoos. The AD segment is the fastest growing segment, growing at 10% to 12 % every year. Shampoo Market Size in India Size of shampoo market – Rs. 2141 cr Anti - Dandruff Shampoo - 20 % of above Sachet Sales - 70 % of above
?
Usage Rate:
The frequency of shampoo usage is very low in India. Most consumers use shampoo only once or twice in a week. About 50% of consumers use ordinary toilet soaps to wash their hair. About 15% of consumers use toilet soaps as well as shampoo for cleaning their hair. Also 70% of the total shampoo sales are through sachet sales. HLL has higher stakes in the rural market with an 80% share.
Usage of Shampoos
Use Toilet Soaps + Shampoo Both 15%
Use Shampoo 35%
Use Toilet Soaps 50%
?
Major Market Players and their Brands
HUL Market Leader (Sunsilk, Clinic All Clear, Clinic Plus, clinic Active, Sunsilk Nutracare) P & G (Pantene, Head &Shoulders, Rejoice) CavinKare (Chik, Chik Satin Nyle Herbal) Dabur and Ayur.
Major Players (%Share)
Others 17% CavinCare 12% HUL 47%
P&G 24%
Research Objectives: Primary Research Objective: To determine the usage pattern of hair shampoos prevalent among the female students at the SIC campus.
Secondary Research Objectives: ? To determine the various market segments of shampoo users at SIC campus. ? To determine the various factors which are considered while buying shampoo. ? To determine the various factors that influence consumer buying behaviour. ? To determine the perception of consumers about the different attributes of shampoos. ? To determine the perception of consumers about the different brands of shampoos. ? To determine the attributes of a new, potentially successful shampoo in the existing market. ? To determine the discriminating criterion to segregate the population into loyal and normal customers.
METHODOLOGY AND LIMITATIONS:Methodology: ? Data Collection Methods:
After collecting secondary data from internet reports and syndicated services, data was collected using exploratory research and primary data collection methods.
?
Measurement Techniques:
The following measurement techniques were used: 1) Questionnaire - Exploratory Questionnaire - Open ended - Recruitment Questionnaire - Main Questionnaire - Close ended (Multiple choice / Dichotomous) 2) Attitude Scales - Direct Response Attitude Scales (Rating Scales and Attitude Scales)
- Indirect Derived Attitude Scales (Perceptual Mapping using Factor Analysis, Cluster Analysis, Discriminant Analysis and Attribute Based Perceptual Mapping)
? Sampling Techniques: The sampling technique which was used is Non-probability Quota Sampling with a sample size of 100 respondents.
?
Analysis Techniques:
The research was conducted at SIC campus. The sample size was 100 respondents and sample frame would consist of the female students from all three SCMHRD, SIIB and SCIT colleges. The data was collected through questionnaire. The analysis was carried out using SPSS 15.0 windows evaluation version using following techniques: Cross tabulation Graphs Regression & Correlation ANOVA
Limitations: 1. The scope is limited as it is applicable to only Symbiosis Info tech Campus female students which more or less proved to be very similar to each other in terms of purchasing behaviour in shampoos. Hence this sample can’t be extrapolated to whole population outside. 2. There is a further scope of getting a clearer picture by doing a conjoint analysis.
FINDINGS:
CHI-SQUARE WITH CROSS TABS: CHI- SQUARE was done to find out if there was any relation between monthly expenses and expenditure on shampoo of consumers. The chi-square analysis proves that there exists a significant relation between monthly expenditure and the expenditure on shampoos by the consumers.
RANGE * MONTHLYEXP Crosstabulation Count MONTHLYEXP 20003000<2000 (A) 3000(B) 4000(C) 0 0 0 10 4 4 4 22 7 4 10 9 30 7 8 14 3 32
RANGE High Low Mediu m VHigh Total
21 0 0 0 0 21
>4000(D) 0 0 2 14 0 16
Total 21 24 18 42 16 121
The graphical representation of the same output is shown below:
Cluster Analysis is a multivariate procedure ideally suited to segmentation applications in marketing research. Segmentation involves identifying groups of target customers who are similar in buying habits, demographics, or psychographics. The methods used here are:a) Hierarchical clustering or Linkage methods (average)and b) Non-hierarchical clustering or Nodal methods.
From our agglomeration schedule in hierarchical clustering, we have identified two clusters present in the population. Further in K-means clustering after seeing the group means of each variable of each of the cluster , based on the 16 variables in questionnaire the main features of the two clusters are as follows:CLUSTER 1
eople belonging to this cluster like to be leaders of the group. Career is very important to them. They take their purchasing decisions on their own. In times of trouble they look for friends rather than family. There is no bias towards low cost products in this cluster. They moderately prefer to dress for fashion rather than comfort. They do not at all buy things on impulse rather diligently think before purchasing. They don’t mind learning new things even if they are not useful to them. They are the lead users of newly launched products. They like a lot of variety in life. They would love to go for party at the expense of their work. They mildly prefer staying abroad if given a chance. Advertisement plays a major role in their purchasing decisions. They believe that quality comes at a price. They feel that they are more capable than others. Attractive packaging doesn’t influence their buying behavior. CLUSTER 2:- People belonging to this cluster don’t like to be leaders of the group. Career is important to them. They are influenced by others in their purchasing decisions. In times of trouble they look for family rather than friends. Their probability of choosing low cost and expensive products is same. They prefer to dress for comfort rather than fashion. They do not at all buy things on impulse rather diligently think before purchasing. They don’t want to learn new things if they are not useful to them. They come under the category of laggards of newly launched products. They are neutral over seeking variety in life. They would work rather than going out for a party. They won’t prefer staying abroad if given a chance. Advertisement doesn’t affect their purchasing decisions. They strongly disagree that quality comes at a price. They feel that there are people who are more capable than them. Attractive packaging doesn’t influence their buying behavior at all. People in Cluster 1 can be taken as “Experiencers” of VALS -2 framework who are young, enthusiastic, impulsive who seek variety and excitement in life. They spend a comparatively high proportion of income on fashion, entertainment and socializing. People in Cluster 2 can be taken as “Makers” of VALS-2 framework who are practical, down-toearth , self-sufficient family oriented people who love to work. They seek products with a practical or functional purpose. Cluster 1 can be targeted with new products (shampoos) as it represents early users and opinion leaders by catchy advertisement and quality product. Cluster 2 represents late users and laggards and hence will follow cluster 1 in purchasing new products. SEMANTIC DIFFERENTIAL SCALE TO COMPARE BRANDS L’Oreal (L) AND DOVE(D) ON THE FOLLOWING PARAMETERS :
The research conducted revealed that the following pattern is prevalent amongst the population. It shows that Dove is the shampoo which has been tried by most of the consumers (21%). Next comes L’Oreal (16%) and Sunsilk (13%) which has been used. Clinic plus (10% ) and Head & Shoulders and Ayur (9% each) are less used. Clinic All Clear (8%) forms only a portion of the usage.
USAGE PATTERN
4% 8% 9% 9% 10% 13% 7% 3% 16% 21% L'OREAL DOVE SUNSILK CLINIC PLUS H&S AYUR CAC HIMALAYA
D - Dove 1 EXPENSIVE ATTRACTIVE PAKAGING GOOD FRAGRANCE HIGHER MILDNESS D D 2
and
S – Sunsilk 3 4 S S 5 CHEAP ORDINARY PACKAGING ORDINARY FRAGRANCE MEDIUM MILDNESS
S D
D S
Using this scale(5-point),the average ranking of responses of the respondents of two close medium priced competitors, dove and sunsilk, were taken. The result shows that Dove is perceived to be an expensive shampoo with attractive packaging, mild fragrance and high mildness. Sunsilk, on the other hand is perceived as a cheaper brand with normal packaging but good fragrance and medium mildness.
FACTOR ANALYSIS: Factor analysis is a very useful method of reducing data complexity by reducing the no. of variables into few latent or underlying factors which explain most of the consumer buying behaviour. In this analysis using principle component analysis and varimax rotation we could find out that 5 factors extracted accounted for 80.922% of the total variance (information content in the original 14 variables). This meant that we could economize on the number of variables from 14 to 5 underlying factors while we lose only approximately 20% of the information content of the variables. The factor analysis shows that there are 5 factors that affect the purchase decision. The percentage of total variance explained by these 5 factors cumulatively equals to 80.82% These are: Factor 1 = fn(economical fragrance, silky & smooth, prevents hair fall, bounce) Factor 2 = fn(cleansing power, oil control properties) Factor 3 = fn(herbal content, medicinal value) Factor 4 = fn(brand image) Factor 5 = fn(availability)
Hence we conclude, Factor 1 = Value for money Factor 2 = Favourable effects on hair Factor 3 = Medicinal Properties Factor 4 = Availability Factor 5 = Brand Image
The following perceptual maps were drawn out of the output wrt the factors extracted. Hence we can see the positioning of various attributes vis-à-vis various factors. An arrow longer in size and closer to the axis shows that the attribute has strong correlation with that dimension. PERCEPTUAL MAPS
1) VFM VS BASIC FEATURES VFM 0.78 0.06 0.95 0.94 0.04 BASIC FEATURES -0.18 0.96 0.11 0.09 0.96
FRAGRANCE CLEANSING POWER SILKY & SMOOTH BOUNCE OIL PROPERTIES
VFM Vs BASIC FEATURES
1.20 1.00 0.80 0.60 0.40 0.20 0.00 -0.20 -0.40 0.00 0.20 0.40 0.60 0.80 BASIC FEATURES
OIL CLEANSING
BOUNCE SILKY
1.00
2) VFM vs MEDICINAL VALUE MEDICINAL VFM PROPERTIES FRAGRANCE 0.78 0.00 SILKY & SMOOTH 0.95 0.05 HERBAL 0.04 0.97 ECONOMICAL 0.80 -0.04 PREVENTS HAIR FALL 0.96 0.02 MEDICINAL VALUE 0.02 0.97 BOUNCE 0.94 -0.02
5) BASIC FEATURES vs MEDICINAL PROPERTIES BASIC MEDICINAL FEATURES PROPERTIES CLEANSING POWER HERBAL MEDICINAL VALUE OIL PROPERTIES 0.96 0.05 0.05 0.96 0.06 0.97 0.97 0.09
6) BASIC FEATURES VS AVAILABILITY BASIC FEATURES AVAILABILITY CLEANSING POWER AVAILABILITY OIL PROPERTIES 0.96 -0.04 0.96 0.07 -0.79 0.08
DISCRIMINANT ANALYSIS:
Out of the population surveyed, 33% of the respondents are ready to spend around Rs.60-80 on a bottle of shampoo. 28% feel that price poses no bar on their decision. 22% of them are ready to spend more than Rs.80, whereas 15 are willing to spend around Rs. 40-60.
PRICE RANGE PEOPLE LOOK FOR WHILE BUYING
2% 22% 15% 28% 33% 60-80 PRICE NO BAR 40-60 >80 <40
Discriminant Function:
Wilks' Lambda Test of Function(s) 1 Wilks' Lambda .460 Chisquare 55.629
df 4
Sig. .000
Canonical Discriminant Function Coefficients Function 1 .184 .018 1.884 .053 -4.237
TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS (Constant)
Unstandardized coefficients
Standardized Canonical Discriminant Function Coefficients Function 1 .320 .361 .744 .064
TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS
Functions at Group Centroids SWITCH_TO_ANOT Function HER_BRAND 1 LOYAL CUSTOMER .793 NORMAL -.970 CUSTOMER Unstandardized canonical discriminant functions evaluated at group means
Classification Results(a) Predicted Group Membership Total LOYAL NORMAL LOYAL CUSTOMER CUSTOMER CUSTOMER 43 12 55 7 78.2 38 21.8 45 100.0
Origina Coun l t %
SWITCH_TO_ANOT HER_BRAND LOYAL CUSTOMER NORMAL CUSTOMER LOYAL CUSTOMER
NORMAL 15.6 CUSTOMER a 81.0% of original grouped cases correctly classified.
84.4
100.0
From the discriminant analysis, the canonical distribution function derived is: Z= -0.018 (since how long are you using this shampoo ) + 0.184 ( would you like to try new shampoo if introduced ) + 1.884 (are you satisfied with the shampoo currently being used ) + 0.053(no of shampoos you have used sice past 2 yrs) - 4.237
The classification accuracy of the discriminant function = 81% (19 cases misclassified) The wilks lambda = 0.46 which is <= 0.5 , hence is an acceptable solution Sig= 0.000 and the confidence level = 100% = ( 1-0.000) The best discriminator is = the satisfaction with the shampoo they are currently using (.744). Discriminant Analysis is a technique which enabled us to distinguish between brand loal customers and normal customers. The attributes which helped us to classify people were:1. Long term usage of the shampoo 2. Tendency to try newly launched shampoo 3. No. of brands of shampoos used in the past 2 years 4. Satisfaction with respect to current shampoo The dependent variable was taken to be the tendency to switxh between shampoo brands currently. The most important attribute for classification had been proved to be satisfaction of the shampoo that the customer is presently using. In future we can estimate if a particular customer is going to be brand loyal to our shampoo or not using the following discriminant equation and putting the corresponding values of independent variables for the new customer. Z= -0.018 (since how long are you using this shampoo ) + 0.184 ( would you like to try new shampoo if introduced ) + 1.884 (are you satisfied with the shampoo currently being used ) + 0.053(no of shampoos you have used sice past 2 yrs) - 4.237 The discriminate score cutoff came out to be -2.235. The new customers having a discriminant score greater than this value will be considered to be brand loyals.
Attribute Based Perceptual Mapping Using Discriminant Analysis: A perceptual map provides the practical information necessary to diagnose and address brand issues and opportunities. It helps us to identify our strengths and weaknesses relative to our competitors and what steps we need to take to improve our position. This tool does three things: 1. It determines how a customer simplifies factors or aspects to differentiate competitors. These are used to create the axis in the map. 2. It shows position of competitors relative to each other, and particularly the distance from one competitor to another. 3. It determines specifically what factors or aspects contribute to the simplified decisionmaking We have used this tool to understand the positioning of various brands used by the target customers. When the important features of a shampoo are rated by respondents on a seven point Likert scale and fed to the SPSS software for Attribute based perceptual mapping analysis the attributes were clubbed under two functions. Function1 = Function of (Smooth hair on usage, mild shampoo, fragrance) Function2 = Function of (Bounce to hair, Rich Lather, Shiny hair on usage) We name the function F1 as Shampoo features and F2 as Usage effects. An ABPM map of the brands Dove, Ultradoux, Sunsilk, Pantene and Head and shoulders is shown in the below map.
While dove has been positioned as a mild shampoo with nice fragrance, Ultradoux does well on bounce and smoothness. Sunsilk has been positioned well as a hair smoothening shampoo with nice fragrance. Pantene, as its positioning statement reads “Shine, Pantene” does well on Shine and bounce features. Head and Shoulders is mild and gives shine on usage. There is a big gap between Head and shoulders and the rest of the shampoos. Similarly there is a gap between Sunsilk range of shampoos and Dove where one can look for the launch of a new product. Hence we can work for a new shampoo with the attributes of a milder shampoo that gives a nice shine on usage with a price similar to that of Head and Shoulders. Alternately we can develop a shampoo which is mild, fragrant and which gives smoothness to hair on usage. Recommendations: 1) Shampoo industry is fairly saturated with many market players and many brands in India. In order to compete with the present brands, a newly launched product should have unique positioning on the attributes that customer’s value. One such proposition is to develop a shampoo which is highly associated with shine and mildness. 2) A second gap in the market can be found as there is no shampoo which is currently highly associated with Smoothness and fragrance.
APPENDIXES:CHI-SQUARE: ANALYSIS:
H0:- There is no significant relationship (at 90% confidence level) between monthly expenses and expenditure on shampoo of customers. H1:- There is a significant relationship (at 90% confidence level) between monthly expenses and expenditure on shampoo of customers. SPSS OUTPUT
Chi-Square Tests Asymp. Sig. (2-sided)
Value Pearson Chi-Square Likelihood Ratio N of Valid Cases
df
155.152(a) 16 .000 143.740 16 .000 121 a 16 cells (64.0%) have expected count less than 5. The minimum expected count is 2.12.
RANGE * MONTHLYEXP Crosstabulation Count MONTHLYEXP A RANGE H L M VH Total 21 0 0 0 0 21 0 10 4 4 4 22 B 0 7 4 10 9 30 C 0 7 8 14 3 32 D 0 0 2 14 0 16 Total 21 24 18 42 16 121
Since Pearson’s coefficient significance is < 0.05, we reject the null hypothesis and this proves that there exists a significant relation between monthly expenditure and the expenditure on shampoos by the consumers.
Factor Analysis: Extraction Method: Principal Component Analysis. Total Variance Explained Initial Eigenvalues Extraction Sums of Squared Loadings % of Cumulative % of Cumulative Variance % Total Variance % 28.779 28.779 4.029 28.779 28.779 17.851 46.630 2.499 17.851 46.630 16.944 63.574 2.372 16.944 63.574 9.783 73.357 1.370 9.783 73.357 7.565 80.922 1.059 7.565 80.922 6.880 87.803
Compone nt 1 2 3 4 5 6
Total 4.029 2.499 2.372 1.370 1.059 .963
Rotation Sums % Total Var 4.008 2.377 2.357 1.494 1.092
.800 5.713 93.516 .486 3.472 96.988 .323 2.304 99.293 .099 .707 100.000 3.90E2.79E-015 100.000 016 12 1.95E1.40E-015 100.000 016 13 -1.97E-1.41E-015 100.000 016 14 -6.63E-4.73E-015 100.000 016 Extraction Method: Principal Component Analysis.
7 8 9 10 11
Rotated Component Matrix(a) Component 2 3 -.257 -.296 -.179 -.003 .955 .062 .113 .052 .048 .965
LATHER FRAGRANCE CLEANSING SILKY HERBAL BRANDIMAG -.082 .270 -.111 .158 E FRIENDS -.074 -.010 .213 .594 ADS -.001 -.552 .487 .457 ECONOMICA .798 -.183 -.044 -.238 L AVAILABILIT -.142 -.044 .293 -.787 Y HAIRFALL .962 .105 .017 .077 MEDVALUE .015 .052 .970 .038 BOUNCE .944 .095 -.017 .058 OIL .043 .957 .086 .077 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a Rotation converged in 7 iterations.
1 .083 .783 .062 .946 .038
4 -.388 -.224 .066 .094 .025
5 -.171 .433 .130 -.163 -.077 .717 .107 .185 .440 .107 -.148 -.086 -.129 .120
Component Transformation Matrix
Compone nt 1 2 3 4 1 .996 .048 -.011 -.043 2 -.009 .585 .743 .324 3 .054 -.786 .604 .049 4 -.042 .164 .284 -.942 5 -.061 -.107 .056 -.060 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.
5 .065 .041 -.113 -.058 .989
LATHER FRAGRANCE CLEANSING POWER SILKY & SMOOTH HERBAL BRAND IMAGE FRIENDS ADS ECONOMICAL AVAILABILITY PREVENTS HAIR FALL MEDICINAL VALUE BOUNCE OIL PROPERTIES
Rotated Component Matrix Component BASIC MEDICINAL BRAND VFM FEATURES PROPERTIES AVAILABILITY IMAGE 0.08 -0.26 -0.30 -0.39 -0.17 0.78 -0.18 0.00 -0.22 0.43 0.06 0.95 0.04 -0.08 -0.07 0.00 0.80 -0.14 0.96 0.02 0.94 0.04 0.96 0.11 0.05 0.27 -0.01 -0.55 -0.18 -0.04 0.11 0.05 0.09 0.96 0.06 0.05 0.97 -0.11 0.21 0.49 -0.04 0.29 0.02 0.97 -0.02 0.09 0.07 0.09 0.02 0.16 0.59 0.46 -0.24 -0.79 0.08 0.04 0.06 0.08 0.13 -0.16 -0.08 0.72 0.11 0.18 0.44 0.11 -0.15 -0.09 -0.13 0.12
ANALYSIS: The factor analysis shows that there are 5 factors that affect the purchase decision. The percentage of total variance explained by these 5 factors cumulatively equals to 80.82% These are: Factor 1 = fn(economical fragrance, silky & smooth, prevents hair fall, bounce) Factor 2 = fn(cleansing power, oil control properties) Factor 3 = fn(herbal content, medicinal value) Factor 4 = fn(brand image) Factor 5 = fn(availability)
Hence we conclude, Factor 1 = Value for money Factor 2 = Favourable effects on hair Factor 3 = Medicinal Properties Factor 4 = Availability Factor 5 = Brand Image After the factors are extracted, the communalities are: Factor 1 = 4.30 Factor 2 = 1.78 Factor 3 = 1.80 Factor 4 = 0.55 Factor 5 = 0.40
Cluster Analysis: SPSS Output: Agglomeration Schedule Coefficient s Stage Cluster First Appears Next Stage
Cluster Combined
Stage Cluster 1 Cluster 2 Cluster 1 Cluster 2 Cluster 1 Cluster 2 1 72 97 6.000 0 0 7 2 26 52 6.000 0 0 20 3 55 87 8.000 0 0 35 4 37 56 8.000 0 0 17 5 9 41 8.000 0 0 23 6 76 86 12.000 0 0 53 7 20 72 12.000 0 1 55 8 64 70 12.000 0 0 54 9 19 61 12.000 0 0 56 10 30 40 12.000 0 0 69 11 35 39 12.000 0 0 23 12 18 24 12.000 0 0 65 13 57 85 13.000 0 0 52 14 53 83 13.000 0 0 57 15 67 78 13.000 0 0 54 16 65 66 13.000 0 0 70 17 7 37 13.000 0 4 56 18 21 32 13.000 0 0 53 19 10 23 13.000 0 0 22 20 1 26 15.000 0 2 51 21 2 11 15.000 0 0 58 22 10 69 15.500 19 0 59 23 9 35 15.500 5 11 61
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70
50 49 48 47 46 45 44 43 42 38 36 55 34 33 31 29 28 27 25 22 17 16 14 13 12 8 5 1 55 21 64 4 7 53 2 1 4 2 21 17 64 18 36 2 4 2 4
96 95 94 93 92 91 90 89 88 84 82 81 80 79 77 75 74 73 71 68 63 62 60 59 58 54 51 6 57 76 67 20 19 55 7 10 46 9 53 48 98 21 45 15 25 30 65
16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.333 19.833 20.000 20.000 21.000 21.000 22.300 22.300 22.750 24.500 24.500 25.250 26.000 26.250 26.545 27.000 28.455 28.667 29.667 29.875
0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 20 35 18 8 0 17 14 21 51 55 58 53 44 54 12 34 61 60 67 68
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 6 15 7 9 52 56 22 28 23 57 26 0 62 29 0 42 10 16
92 75 63 97 60 66 73 80 82 76 66 52 89 83 79 80 88 84 68 74 63 92 95 81 85 78 74 59 57 62 64 60 58 62 61 72 68 67 65 78 71 71 81 69 70 72 77
71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
18 1 18 5 3 1 4 8 1 29 13 5 4 13 4 4 8 4 4 4 1 16 1 1 3 1 1
64 2 44 22 49 38 18 17 31 43 36 42 33 27 12 5 13 28 34 29 4 50 8 16 14 3 47
33.877 34.296 37.778 39.000 40.000 40.190 40.740 41.000 41.478 42.000 42.500 43.500 45.333 47.333 52.438 54.725 55.667 56.950 57.619 59.114 59.368 72.000 77.006 87.287 90.667 103.763 221.396
65 59 71 50 0 72 70 49 76 39 47 74 77 81 83 85 78 86 88 89 79 45 91 93 75 94 96
64 69 30 43 25 33 73 63 38 31 66 32 37 41 48 82 84 40 36 80 90 24 87 92 46 95 27
73 76 77 82 95 79 83 87 91 90 84 86 85 87 86 88 93 89 90 91 93 94 94 96 96 97 0
Analysis: Jump between 96th and 97th step is maximum and hence we conclude that there are two clusters in our sample. Further going to K-Means clustering, final cluster centers are as follows:
Final Cluster Centers Cluster LIKE_LEADER CAREER_IMPORTANT GO_OUT LOOK_FOR_FRIENDS ADULT_STUFF FASHION IMPULSE_PURCHASE LEARN_NEW_THINGS EVERYDAY_GOODS VARIETY 1 2.74 2.10 2.69 3.50 3.36 2.88 5.09 3.57 2.79 2.33 2 4.33 3.23 4.43 4.40 4.18 4.15 6.00 5.13 3.95 3.80
FOLLOW_HEAD STAY_ABROAD ADVERTISEMENT QUALITY_PRICE CAPABLE ATTRACTIVE_PACKAGI NG
2.78 2.86 2.52 1.64 3.29 4.10
4.08 4.78 4.40 3.03 4.95 5.40
ANOVA Cluster Mean Square df 59.370 1 29.778 1 71.291 1 19.176 1 15.645 1 38.224 1 19.768 1 57.319 1 31.685 1 51.324 1 39.955 1 86.628 1 83.917 1 45.547 1 64.991 1 Error Mean Square 1.499 1.254 1.439 2.522 1.783 1.846 2.256 1.756 1.202 1.491 2.030 2.811 1.543 1.254 1.437 df F Mean Square 39.608 23.753 49.526 7.604 8.774 20.702 8.763 32.637 26.354 34.413 19.684 30.816 54.402 36.325 45.238 Sig. df .000 .000 .000 .007 .004 .000 .004 .000 .000 .000 .000 .000 .000 .000 .000
LIKE_LEADER 96 CAREER_IMPORTANT 96 GO_OUT 96 LOOK_FOR_FRIENDS 96 ADULT_STUFF 96 FASHION 96 IMPULSE_PURCHASE 96 LEARN_NEW_THINGS 96 EVERYDAY_GOODS 96 VARIETY 96 FOLLOW_HEAD 96 STAY_ABROAD 96 ADVERTISEMENT 96 QUALITY_PRICE 96 CAPABLE 96 ATTRACTIVE_PACKAGI 39.796 1 1.906 96 20.879 .000 NG The F tests should be used only for descriptive purposes because the clusters have been chosen to maximize the differences among cases in different clusters. The observed significance levels are not corrected for this and thus cannot be interpreted as tests of the hypothesis that the cluster means are equal. ANALYSIS: - Since p-value for all the factors is less than 0.05, we conclude that all the factors have a significant influence in clustering the sample into segments.
Discriminant Analysis: SPSS Output: Group Statistics
SWITCH_TO_ANO THER_BRAND 1.00
2.00
Total
TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS
Std. Mean Deviation Valid N (listwise) Unweighte Weighte Unweighte Weighte d d d d 5.8545 1.45829 55 55.000 27.8364 1.7818 2.0545 4.4667 7.4667 1.1556 2.6000 5.2300 18.6700 1.5000 2.3000 26.52376 .41682 1.26810 2.02933 9.05438 .36653 1.15601 1.86328 22.88886 .50252 1.24316 55 55 55 45 45 45 45 100 100 100 100 55.000 55.000 55.000 45.000 45.000 45.000 45.000 100.000 100.000 100.000 100.000
Eigenvalues Functio Eigenvalu % of Cumulative Canonical n e Variance % Correlation 1 1.15(a) 100.0 100.0 .663 a First 1 canonical discriminant functions were used in the analysis. Canonical Discriminant Function Coefficients Function 1 .184 .018 1.884 .053 -4.237
TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS (Constant)
Unstandardized coefficients Functions at Group Centroids SWITCH_TO_ANOT Function HER_BRAND 1 LOYAL CUSTOMER .793 NORMAL -.970 CUSTOMER Unstandardized canonical discriminant functions evaluated at group means ANALYSIS
iscriminant score cutoff for classifying new consumers into the two groups. = (0.793* 55 – 0.970 * 45 )/100 =-2.235 Classification Results(a) Predicted Group Membership Total LOYAL NORMAL LOYAL CUSTOMER CUSTOMER CUSTOMER 43 12 55 38 21.8 84.4 45 100.0 100.0
SWITCH_TO_ANOT HER_BRAND Origina Coun LOYAL CUSTOMER l t NORMAL 7 CUSTOMER % LOYAL CUSTOMER 78.2 NORMAL 15.6 CUSTOMER a 81.0% of original grouped cases correctly classified.
The classification accuracy came out to be 81% which implies that the model is fairly acceptable. The eigen values are also greater than 1 and hence the factors are significant Attibute based perceptual mapping using Discriminant Analysis: Eigenvalues % of Cumulative Canonical Function Eigenvalue Variance % Correlation 1 4.749(a) 81.4 81.4 0.909 2 1.083(a) 18.6 100 0.721 a First 2 canonical discriminant functions were used in the analysis.
Structure Matrix Function 1 2 Mild .846(*) 0.278 smooth .782(*) -0.33 Fragrance .748(*) 0.076 Bounce 654 .727(*) Shine 0.541 .654(*) Lather -0.263 .926(*) Pooled within-groups correlations between discriminating variables and standardized canonical discriminant functions Variables ordered by absolute size of correlation within function. * Largest absolute correlation between each variable and any discriminant function a This variable not used in the analysis.
Functions at Group Centroids Function BRAND 1 2 1 0.935 -0.201 2 0.254 0.204 3 -0.343 0.127 4 -1.506 -0.123 5 0.659 -0.008 Unstandardized canonical discriminant functions evaluated at group means Analysis: When the important features of a shampoo are rated by respondents on a seven point Likert scale and fed to the SPSS software for Attribute based perceptual mapping analysis the attributes were clubbed under two functions. Function1 = Function of (Smooth hair on usage, mild shampoo, fragrance) Function2 = Function of (Bounce to hair, Rich Lather, Shiny hair on usage) We name the function F1 as Shampoo features and F2 as Usage effects.
doc_629176754.docx
This research is to focus on consumer buying behaviour or customer preferences of different brands of shampoos from various segments prevailing in market and thus finding out the opportunities for new entrants.
MARKET RESEARCH PROJECT
[Final Report]
[Usage pattern of Shampoo among female students at SIC campus]
Executive Summary:
The main objective behind this research was to focus on consumer buying behaviour or customer preferences of different brands of shampoos from various segments prevailing in market and thus finding out the opportunities for new entrants. The research was conducted at Symbiosis Infotech Campus, Hinjewadi, Pune to determine the shampoo purchase behaviours of female students in the age group of 20 to 30. The research was conducted on a sample size of 100 respondents. The population was deviden into brand loyals nad normal customers and hence any new onsumer can be put in the corresponding category based on his characteristics.After finding out the main factors influencing the buying behaviour of female students and segmenting this market into two segments i.e. experiencers and makers, we found out gap in the current market and concluded that the new product should be targeted at experiencers. It should be medium-priced (Rs. 50-60 for 100 ml.), should be mild on hair and should give shine to the hair as other attributes are already present in the shampoos available in the market.
Contents
Executive Summry: ......................................................................................................................... 2 toc.................................................................................................................................................... 4 Introduction: .................................................................................................................................... 5 METHODOLOGY AND LIMITATIONS:-................................................................................... 7
toc
Introduction:
(Source: Internet and AC Neilson Report) Background: The Indian shampoo industry is estimated at Rs. 2141 cr and is growing at an average rate of 20% per annum. According to AC Nielsen, shampoo is one of the fastest growing categories within FMCG sector and is expected to grow at 25% per annum in the coming years. From a penetration level of 13 per cent in 2000 to 31.9 per cent in 2005, to now almost a third of the country's rural population uses shampoo with penetration levels zooming at 36 per cent in 2008. Urban markets account for 80% of the total shampoo market.
?
Segmentation:
The shampoo industry is segmented on benefit platforms: - Cosmetic (shine, health and strength) - Anti Dandruff (AD) and - Herbal 20% of the total shampoo market is accounted by the AD shampoos. The AD segment is the fastest growing segment, growing at 10% to 12 % every year. Shampoo Market Size in India Size of shampoo market – Rs. 2141 cr Anti - Dandruff Shampoo - 20 % of above Sachet Sales - 70 % of above
?
Usage Rate:
The frequency of shampoo usage is very low in India. Most consumers use shampoo only once or twice in a week. About 50% of consumers use ordinary toilet soaps to wash their hair. About 15% of consumers use toilet soaps as well as shampoo for cleaning their hair. Also 70% of the total shampoo sales are through sachet sales. HLL has higher stakes in the rural market with an 80% share.
Usage of Shampoos
Use Toilet Soaps + Shampoo Both 15%
Use Shampoo 35%
Use Toilet Soaps 50%
?
Major Market Players and their Brands
HUL Market Leader (Sunsilk, Clinic All Clear, Clinic Plus, clinic Active, Sunsilk Nutracare) P & G (Pantene, Head &Shoulders, Rejoice) CavinKare (Chik, Chik Satin Nyle Herbal) Dabur and Ayur.
Major Players (%Share)
Others 17% CavinCare 12% HUL 47%
P&G 24%
Research Objectives: Primary Research Objective: To determine the usage pattern of hair shampoos prevalent among the female students at the SIC campus.
Secondary Research Objectives: ? To determine the various market segments of shampoo users at SIC campus. ? To determine the various factors which are considered while buying shampoo. ? To determine the various factors that influence consumer buying behaviour. ? To determine the perception of consumers about the different attributes of shampoos. ? To determine the perception of consumers about the different brands of shampoos. ? To determine the attributes of a new, potentially successful shampoo in the existing market. ? To determine the discriminating criterion to segregate the population into loyal and normal customers.
METHODOLOGY AND LIMITATIONS:Methodology: ? Data Collection Methods:
After collecting secondary data from internet reports and syndicated services, data was collected using exploratory research and primary data collection methods.
?
Measurement Techniques:
The following measurement techniques were used: 1) Questionnaire - Exploratory Questionnaire - Open ended - Recruitment Questionnaire - Main Questionnaire - Close ended (Multiple choice / Dichotomous) 2) Attitude Scales - Direct Response Attitude Scales (Rating Scales and Attitude Scales)
- Indirect Derived Attitude Scales (Perceptual Mapping using Factor Analysis, Cluster Analysis, Discriminant Analysis and Attribute Based Perceptual Mapping)
? Sampling Techniques: The sampling technique which was used is Non-probability Quota Sampling with a sample size of 100 respondents.
?
Analysis Techniques:
The research was conducted at SIC campus. The sample size was 100 respondents and sample frame would consist of the female students from all three SCMHRD, SIIB and SCIT colleges. The data was collected through questionnaire. The analysis was carried out using SPSS 15.0 windows evaluation version using following techniques: Cross tabulation Graphs Regression & Correlation ANOVA
Limitations: 1. The scope is limited as it is applicable to only Symbiosis Info tech Campus female students which more or less proved to be very similar to each other in terms of purchasing behaviour in shampoos. Hence this sample can’t be extrapolated to whole population outside. 2. There is a further scope of getting a clearer picture by doing a conjoint analysis.
FINDINGS:
CHI-SQUARE WITH CROSS TABS: CHI- SQUARE was done to find out if there was any relation between monthly expenses and expenditure on shampoo of consumers. The chi-square analysis proves that there exists a significant relation between monthly expenditure and the expenditure on shampoos by the consumers.
RANGE * MONTHLYEXP Crosstabulation Count MONTHLYEXP 20003000<2000 (A) 3000(B) 4000(C) 0 0 0 10 4 4 4 22 7 4 10 9 30 7 8 14 3 32
RANGE High Low Mediu m VHigh Total
21 0 0 0 0 21
>4000(D) 0 0 2 14 0 16
Total 21 24 18 42 16 121
The graphical representation of the same output is shown below:
Cluster Analysis is a multivariate procedure ideally suited to segmentation applications in marketing research. Segmentation involves identifying groups of target customers who are similar in buying habits, demographics, or psychographics. The methods used here are:a) Hierarchical clustering or Linkage methods (average)and b) Non-hierarchical clustering or Nodal methods.
From our agglomeration schedule in hierarchical clustering, we have identified two clusters present in the population. Further in K-means clustering after seeing the group means of each variable of each of the cluster , based on the 16 variables in questionnaire the main features of the two clusters are as follows:CLUSTER 1
eople belonging to this cluster like to be leaders of the group. Career is very important to them. They take their purchasing decisions on their own. In times of trouble they look for friends rather than family. There is no bias towards low cost products in this cluster. They moderately prefer to dress for fashion rather than comfort. They do not at all buy things on impulse rather diligently think before purchasing. They don’t mind learning new things even if they are not useful to them. They are the lead users of newly launched products. They like a lot of variety in life. They would love to go for party at the expense of their work. They mildly prefer staying abroad if given a chance. Advertisement plays a major role in their purchasing decisions. They believe that quality comes at a price. They feel that they are more capable than others. Attractive packaging doesn’t influence their buying behavior. CLUSTER 2:- People belonging to this cluster don’t like to be leaders of the group. Career is important to them. They are influenced by others in their purchasing decisions. In times of trouble they look for family rather than friends. Their probability of choosing low cost and expensive products is same. They prefer to dress for comfort rather than fashion. They do not at all buy things on impulse rather diligently think before purchasing. They don’t want to learn new things if they are not useful to them. They come under the category of laggards of newly launched products. They are neutral over seeking variety in life. They would work rather than going out for a party. They won’t prefer staying abroad if given a chance. Advertisement doesn’t affect their purchasing decisions. They strongly disagree that quality comes at a price. They feel that there are people who are more capable than them. Attractive packaging doesn’t influence their buying behavior at all. People in Cluster 1 can be taken as “Experiencers” of VALS -2 framework who are young, enthusiastic, impulsive who seek variety and excitement in life. They spend a comparatively high proportion of income on fashion, entertainment and socializing. People in Cluster 2 can be taken as “Makers” of VALS-2 framework who are practical, down-toearth , self-sufficient family oriented people who love to work. They seek products with a practical or functional purpose. Cluster 1 can be targeted with new products (shampoos) as it represents early users and opinion leaders by catchy advertisement and quality product. Cluster 2 represents late users and laggards and hence will follow cluster 1 in purchasing new products. SEMANTIC DIFFERENTIAL SCALE TO COMPARE BRANDS L’Oreal (L) AND DOVE(D) ON THE FOLLOWING PARAMETERS :The research conducted revealed that the following pattern is prevalent amongst the population. It shows that Dove is the shampoo which has been tried by most of the consumers (21%). Next comes L’Oreal (16%) and Sunsilk (13%) which has been used. Clinic plus (10% ) and Head & Shoulders and Ayur (9% each) are less used. Clinic All Clear (8%) forms only a portion of the usage.
USAGE PATTERN
4% 8% 9% 9% 10% 13% 7% 3% 16% 21% L'OREAL DOVE SUNSILK CLINIC PLUS H&S AYUR CAC HIMALAYA
D - Dove 1 EXPENSIVE ATTRACTIVE PAKAGING GOOD FRAGRANCE HIGHER MILDNESS D D 2
and
S – Sunsilk 3 4 S S 5 CHEAP ORDINARY PACKAGING ORDINARY FRAGRANCE MEDIUM MILDNESS
S D
D S
Using this scale(5-point),the average ranking of responses of the respondents of two close medium priced competitors, dove and sunsilk, were taken. The result shows that Dove is perceived to be an expensive shampoo with attractive packaging, mild fragrance and high mildness. Sunsilk, on the other hand is perceived as a cheaper brand with normal packaging but good fragrance and medium mildness.
FACTOR ANALYSIS: Factor analysis is a very useful method of reducing data complexity by reducing the no. of variables into few latent or underlying factors which explain most of the consumer buying behaviour. In this analysis using principle component analysis and varimax rotation we could find out that 5 factors extracted accounted for 80.922% of the total variance (information content in the original 14 variables). This meant that we could economize on the number of variables from 14 to 5 underlying factors while we lose only approximately 20% of the information content of the variables. The factor analysis shows that there are 5 factors that affect the purchase decision. The percentage of total variance explained by these 5 factors cumulatively equals to 80.82% These are: Factor 1 = fn(economical fragrance, silky & smooth, prevents hair fall, bounce) Factor 2 = fn(cleansing power, oil control properties) Factor 3 = fn(herbal content, medicinal value) Factor 4 = fn(brand image) Factor 5 = fn(availability)
Hence we conclude, Factor 1 = Value for money Factor 2 = Favourable effects on hair Factor 3 = Medicinal Properties Factor 4 = Availability Factor 5 = Brand Image
The following perceptual maps were drawn out of the output wrt the factors extracted. Hence we can see the positioning of various attributes vis-à-vis various factors. An arrow longer in size and closer to the axis shows that the attribute has strong correlation with that dimension. PERCEPTUAL MAPS
1) VFM VS BASIC FEATURES VFM 0.78 0.06 0.95 0.94 0.04 BASIC FEATURES -0.18 0.96 0.11 0.09 0.96
FRAGRANCE CLEANSING POWER SILKY & SMOOTH BOUNCE OIL PROPERTIES
VFM Vs BASIC FEATURES
1.20 1.00 0.80 0.60 0.40 0.20 0.00 -0.20 -0.40 0.00 0.20 0.40 0.60 0.80 BASIC FEATURES
OIL CLEANSING
BOUNCE SILKY
1.00
2) VFM vs MEDICINAL VALUE MEDICINAL VFM PROPERTIES FRAGRANCE 0.78 0.00 SILKY & SMOOTH 0.95 0.05 HERBAL 0.04 0.97 ECONOMICAL 0.80 -0.04 PREVENTS HAIR FALL 0.96 0.02 MEDICINAL VALUE 0.02 0.97 BOUNCE 0.94 -0.02
5) BASIC FEATURES vs MEDICINAL PROPERTIES BASIC MEDICINAL FEATURES PROPERTIES CLEANSING POWER HERBAL MEDICINAL VALUE OIL PROPERTIES 0.96 0.05 0.05 0.96 0.06 0.97 0.97 0.09
6) BASIC FEATURES VS AVAILABILITY BASIC FEATURES AVAILABILITY CLEANSING POWER AVAILABILITY OIL PROPERTIES 0.96 -0.04 0.96 0.07 -0.79 0.08
DISCRIMINANT ANALYSIS:
Out of the population surveyed, 33% of the respondents are ready to spend around Rs.60-80 on a bottle of shampoo. 28% feel that price poses no bar on their decision. 22% of them are ready to spend more than Rs.80, whereas 15 are willing to spend around Rs. 40-60.
PRICE RANGE PEOPLE LOOK FOR WHILE BUYING
2% 22% 15% 28% 33% 60-80 PRICE NO BAR 40-60 >80 <40
Discriminant Function:
Wilks' Lambda Test of Function(s) 1 Wilks' Lambda .460 Chisquare 55.629
df 4
Sig. .000
Canonical Discriminant Function Coefficients Function 1 .184 .018 1.884 .053 -4.237
TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS (Constant)
Unstandardized coefficients
Standardized Canonical Discriminant Function Coefficients Function 1 .320 .361 .744 .064
TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS
Functions at Group Centroids SWITCH_TO_ANOT Function HER_BRAND 1 LOYAL CUSTOMER .793 NORMAL -.970 CUSTOMER Unstandardized canonical discriminant functions evaluated at group means
Classification Results(a) Predicted Group Membership Total LOYAL NORMAL LOYAL CUSTOMER CUSTOMER CUSTOMER 43 12 55 7 78.2 38 21.8 45 100.0
Origina Coun l t %
SWITCH_TO_ANOT HER_BRAND LOYAL CUSTOMER NORMAL CUSTOMER LOYAL CUSTOMER
NORMAL 15.6 CUSTOMER a 81.0% of original grouped cases correctly classified.
84.4
100.0
From the discriminant analysis, the canonical distribution function derived is: Z= -0.018 (since how long are you using this shampoo ) + 0.184 ( would you like to try new shampoo if introduced ) + 1.884 (are you satisfied with the shampoo currently being used ) + 0.053(no of shampoos you have used sice past 2 yrs) - 4.237
The classification accuracy of the discriminant function = 81% (19 cases misclassified) The wilks lambda = 0.46 which is <= 0.5 , hence is an acceptable solution Sig= 0.000 and the confidence level = 100% = ( 1-0.000) The best discriminator is = the satisfaction with the shampoo they are currently using (.744). Discriminant Analysis is a technique which enabled us to distinguish between brand loal customers and normal customers. The attributes which helped us to classify people were:1. Long term usage of the shampoo 2. Tendency to try newly launched shampoo 3. No. of brands of shampoos used in the past 2 years 4. Satisfaction with respect to current shampoo The dependent variable was taken to be the tendency to switxh between shampoo brands currently. The most important attribute for classification had been proved to be satisfaction of the shampoo that the customer is presently using. In future we can estimate if a particular customer is going to be brand loyal to our shampoo or not using the following discriminant equation and putting the corresponding values of independent variables for the new customer. Z= -0.018 (since how long are you using this shampoo ) + 0.184 ( would you like to try new shampoo if introduced ) + 1.884 (are you satisfied with the shampoo currently being used ) + 0.053(no of shampoos you have used sice past 2 yrs) - 4.237 The discriminate score cutoff came out to be -2.235. The new customers having a discriminant score greater than this value will be considered to be brand loyals.
Attribute Based Perceptual Mapping Using Discriminant Analysis: A perceptual map provides the practical information necessary to diagnose and address brand issues and opportunities. It helps us to identify our strengths and weaknesses relative to our competitors and what steps we need to take to improve our position. This tool does three things: 1. It determines how a customer simplifies factors or aspects to differentiate competitors. These are used to create the axis in the map. 2. It shows position of competitors relative to each other, and particularly the distance from one competitor to another. 3. It determines specifically what factors or aspects contribute to the simplified decisionmaking We have used this tool to understand the positioning of various brands used by the target customers. When the important features of a shampoo are rated by respondents on a seven point Likert scale and fed to the SPSS software for Attribute based perceptual mapping analysis the attributes were clubbed under two functions. Function1 = Function of (Smooth hair on usage, mild shampoo, fragrance) Function2 = Function of (Bounce to hair, Rich Lather, Shiny hair on usage) We name the function F1 as Shampoo features and F2 as Usage effects. An ABPM map of the brands Dove, Ultradoux, Sunsilk, Pantene and Head and shoulders is shown in the below map.
While dove has been positioned as a mild shampoo with nice fragrance, Ultradoux does well on bounce and smoothness. Sunsilk has been positioned well as a hair smoothening shampoo with nice fragrance. Pantene, as its positioning statement reads “Shine, Pantene” does well on Shine and bounce features. Head and Shoulders is mild and gives shine on usage. There is a big gap between Head and shoulders and the rest of the shampoos. Similarly there is a gap between Sunsilk range of shampoos and Dove where one can look for the launch of a new product. Hence we can work for a new shampoo with the attributes of a milder shampoo that gives a nice shine on usage with a price similar to that of Head and Shoulders. Alternately we can develop a shampoo which is mild, fragrant and which gives smoothness to hair on usage. Recommendations: 1) Shampoo industry is fairly saturated with many market players and many brands in India. In order to compete with the present brands, a newly launched product should have unique positioning on the attributes that customer’s value. One such proposition is to develop a shampoo which is highly associated with shine and mildness. 2) A second gap in the market can be found as there is no shampoo which is currently highly associated with Smoothness and fragrance.
APPENDIXES:CHI-SQUARE: ANALYSIS:
H0:- There is no significant relationship (at 90% confidence level) between monthly expenses and expenditure on shampoo of customers. H1:- There is a significant relationship (at 90% confidence level) between monthly expenses and expenditure on shampoo of customers. SPSS OUTPUT
Chi-Square Tests Asymp. Sig. (2-sided)
Value Pearson Chi-Square Likelihood Ratio N of Valid Cases
df
155.152(a) 16 .000 143.740 16 .000 121 a 16 cells (64.0%) have expected count less than 5. The minimum expected count is 2.12.
RANGE * MONTHLYEXP Crosstabulation Count MONTHLYEXP A RANGE H L M VH Total 21 0 0 0 0 21 0 10 4 4 4 22 B 0 7 4 10 9 30 C 0 7 8 14 3 32 D 0 0 2 14 0 16 Total 21 24 18 42 16 121
Since Pearson’s coefficient significance is < 0.05, we reject the null hypothesis and this proves that there exists a significant relation between monthly expenditure and the expenditure on shampoos by the consumers.
Factor Analysis: Extraction Method: Principal Component Analysis. Total Variance Explained Initial Eigenvalues Extraction Sums of Squared Loadings % of Cumulative % of Cumulative Variance % Total Variance % 28.779 28.779 4.029 28.779 28.779 17.851 46.630 2.499 17.851 46.630 16.944 63.574 2.372 16.944 63.574 9.783 73.357 1.370 9.783 73.357 7.565 80.922 1.059 7.565 80.922 6.880 87.803
Compone nt 1 2 3 4 5 6
Total 4.029 2.499 2.372 1.370 1.059 .963
Rotation Sums % Total Var 4.008 2.377 2.357 1.494 1.092
.800 5.713 93.516 .486 3.472 96.988 .323 2.304 99.293 .099 .707 100.000 3.90E2.79E-015 100.000 016 12 1.95E1.40E-015 100.000 016 13 -1.97E-1.41E-015 100.000 016 14 -6.63E-4.73E-015 100.000 016 Extraction Method: Principal Component Analysis.
7 8 9 10 11
Rotated Component Matrix(a) Component 2 3 -.257 -.296 -.179 -.003 .955 .062 .113 .052 .048 .965
LATHER FRAGRANCE CLEANSING SILKY HERBAL BRANDIMAG -.082 .270 -.111 .158 E FRIENDS -.074 -.010 .213 .594 ADS -.001 -.552 .487 .457 ECONOMICA .798 -.183 -.044 -.238 L AVAILABILIT -.142 -.044 .293 -.787 Y HAIRFALL .962 .105 .017 .077 MEDVALUE .015 .052 .970 .038 BOUNCE .944 .095 -.017 .058 OIL .043 .957 .086 .077 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a Rotation converged in 7 iterations.
1 .083 .783 .062 .946 .038
4 -.388 -.224 .066 .094 .025
5 -.171 .433 .130 -.163 -.077 .717 .107 .185 .440 .107 -.148 -.086 -.129 .120
Component Transformation Matrix
Compone nt 1 2 3 4 1 .996 .048 -.011 -.043 2 -.009 .585 .743 .324 3 .054 -.786 .604 .049 4 -.042 .164 .284 -.942 5 -.061 -.107 .056 -.060 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.
5 .065 .041 -.113 -.058 .989
LATHER FRAGRANCE CLEANSING POWER SILKY & SMOOTH HERBAL BRAND IMAGE FRIENDS ADS ECONOMICAL AVAILABILITY PREVENTS HAIR FALL MEDICINAL VALUE BOUNCE OIL PROPERTIES
Rotated Component Matrix Component BASIC MEDICINAL BRAND VFM FEATURES PROPERTIES AVAILABILITY IMAGE 0.08 -0.26 -0.30 -0.39 -0.17 0.78 -0.18 0.00 -0.22 0.43 0.06 0.95 0.04 -0.08 -0.07 0.00 0.80 -0.14 0.96 0.02 0.94 0.04 0.96 0.11 0.05 0.27 -0.01 -0.55 -0.18 -0.04 0.11 0.05 0.09 0.96 0.06 0.05 0.97 -0.11 0.21 0.49 -0.04 0.29 0.02 0.97 -0.02 0.09 0.07 0.09 0.02 0.16 0.59 0.46 -0.24 -0.79 0.08 0.04 0.06 0.08 0.13 -0.16 -0.08 0.72 0.11 0.18 0.44 0.11 -0.15 -0.09 -0.13 0.12
ANALYSIS: The factor analysis shows that there are 5 factors that affect the purchase decision. The percentage of total variance explained by these 5 factors cumulatively equals to 80.82% These are: Factor 1 = fn(economical fragrance, silky & smooth, prevents hair fall, bounce) Factor 2 = fn(cleansing power, oil control properties) Factor 3 = fn(herbal content, medicinal value) Factor 4 = fn(brand image) Factor 5 = fn(availability)
Hence we conclude, Factor 1 = Value for money Factor 2 = Favourable effects on hair Factor 3 = Medicinal Properties Factor 4 = Availability Factor 5 = Brand Image After the factors are extracted, the communalities are: Factor 1 = 4.30 Factor 2 = 1.78 Factor 3 = 1.80 Factor 4 = 0.55 Factor 5 = 0.40
Cluster Analysis: SPSS Output: Agglomeration Schedule Coefficient s Stage Cluster First Appears Next Stage
Cluster Combined
Stage Cluster 1 Cluster 2 Cluster 1 Cluster 2 Cluster 1 Cluster 2 1 72 97 6.000 0 0 7 2 26 52 6.000 0 0 20 3 55 87 8.000 0 0 35 4 37 56 8.000 0 0 17 5 9 41 8.000 0 0 23 6 76 86 12.000 0 0 53 7 20 72 12.000 0 1 55 8 64 70 12.000 0 0 54 9 19 61 12.000 0 0 56 10 30 40 12.000 0 0 69 11 35 39 12.000 0 0 23 12 18 24 12.000 0 0 65 13 57 85 13.000 0 0 52 14 53 83 13.000 0 0 57 15 67 78 13.000 0 0 54 16 65 66 13.000 0 0 70 17 7 37 13.000 0 4 56 18 21 32 13.000 0 0 53 19 10 23 13.000 0 0 22 20 1 26 15.000 0 2 51 21 2 11 15.000 0 0 58 22 10 69 15.500 19 0 59 23 9 35 15.500 5 11 61
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70
50 49 48 47 46 45 44 43 42 38 36 55 34 33 31 29 28 27 25 22 17 16 14 13 12 8 5 1 55 21 64 4 7 53 2 1 4 2 21 17 64 18 36 2 4 2 4
96 95 94 93 92 91 90 89 88 84 82 81 80 79 77 75 74 73 71 68 63 62 60 59 58 54 51 6 57 76 67 20 19 55 7 10 46 9 53 48 98 21 45 15 25 30 65
16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.000 16.333 19.833 20.000 20.000 21.000 21.000 22.300 22.300 22.750 24.500 24.500 25.250 26.000 26.250 26.545 27.000 28.455 28.667 29.667 29.875
0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 20 35 18 8 0 17 14 21 51 55 58 53 44 54 12 34 61 60 67 68
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 6 15 7 9 52 56 22 28 23 57 26 0 62 29 0 42 10 16
92 75 63 97 60 66 73 80 82 76 66 52 89 83 79 80 88 84 68 74 63 92 95 81 85 78 74 59 57 62 64 60 58 62 61 72 68 67 65 78 71 71 81 69 70 72 77
71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
18 1 18 5 3 1 4 8 1 29 13 5 4 13 4 4 8 4 4 4 1 16 1 1 3 1 1
64 2 44 22 49 38 18 17 31 43 36 42 33 27 12 5 13 28 34 29 4 50 8 16 14 3 47
33.877 34.296 37.778 39.000 40.000 40.190 40.740 41.000 41.478 42.000 42.500 43.500 45.333 47.333 52.438 54.725 55.667 56.950 57.619 59.114 59.368 72.000 77.006 87.287 90.667 103.763 221.396
65 59 71 50 0 72 70 49 76 39 47 74 77 81 83 85 78 86 88 89 79 45 91 93 75 94 96
64 69 30 43 25 33 73 63 38 31 66 32 37 41 48 82 84 40 36 80 90 24 87 92 46 95 27
73 76 77 82 95 79 83 87 91 90 84 86 85 87 86 88 93 89 90 91 93 94 94 96 96 97 0
Analysis: Jump between 96th and 97th step is maximum and hence we conclude that there are two clusters in our sample. Further going to K-Means clustering, final cluster centers are as follows:
Final Cluster Centers Cluster LIKE_LEADER CAREER_IMPORTANT GO_OUT LOOK_FOR_FRIENDS ADULT_STUFF FASHION IMPULSE_PURCHASE LEARN_NEW_THINGS EVERYDAY_GOODS VARIETY 1 2.74 2.10 2.69 3.50 3.36 2.88 5.09 3.57 2.79 2.33 2 4.33 3.23 4.43 4.40 4.18 4.15 6.00 5.13 3.95 3.80
FOLLOW_HEAD STAY_ABROAD ADVERTISEMENT QUALITY_PRICE CAPABLE ATTRACTIVE_PACKAGI NG
2.78 2.86 2.52 1.64 3.29 4.10
4.08 4.78 4.40 3.03 4.95 5.40
ANOVA Cluster Mean Square df 59.370 1 29.778 1 71.291 1 19.176 1 15.645 1 38.224 1 19.768 1 57.319 1 31.685 1 51.324 1 39.955 1 86.628 1 83.917 1 45.547 1 64.991 1 Error Mean Square 1.499 1.254 1.439 2.522 1.783 1.846 2.256 1.756 1.202 1.491 2.030 2.811 1.543 1.254 1.437 df F Mean Square 39.608 23.753 49.526 7.604 8.774 20.702 8.763 32.637 26.354 34.413 19.684 30.816 54.402 36.325 45.238 Sig. df .000 .000 .000 .007 .004 .000 .004 .000 .000 .000 .000 .000 .000 .000 .000
LIKE_LEADER 96 CAREER_IMPORTANT 96 GO_OUT 96 LOOK_FOR_FRIENDS 96 ADULT_STUFF 96 FASHION 96 IMPULSE_PURCHASE 96 LEARN_NEW_THINGS 96 EVERYDAY_GOODS 96 VARIETY 96 FOLLOW_HEAD 96 STAY_ABROAD 96 ADVERTISEMENT 96 QUALITY_PRICE 96 CAPABLE 96 ATTRACTIVE_PACKAGI 39.796 1 1.906 96 20.879 .000 NG The F tests should be used only for descriptive purposes because the clusters have been chosen to maximize the differences among cases in different clusters. The observed significance levels are not corrected for this and thus cannot be interpreted as tests of the hypothesis that the cluster means are equal. ANALYSIS: - Since p-value for all the factors is less than 0.05, we conclude that all the factors have a significant influence in clustering the sample into segments.
Discriminant Analysis: SPSS Output: Group Statistics
SWITCH_TO_ANO THER_BRAND 1.00
2.00
Total
TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS
Std. Mean Deviation Valid N (listwise) Unweighte Weighte Unweighte Weighte d d d d 5.8545 1.45829 55 55.000 27.8364 1.7818 2.0545 4.4667 7.4667 1.1556 2.6000 5.2300 18.6700 1.5000 2.3000 26.52376 .41682 1.26810 2.02933 9.05438 .36653 1.15601 1.86328 22.88886 .50252 1.24316 55 55 55 45 45 45 45 100 100 100 100 55.000 55.000 55.000 45.000 45.000 45.000 45.000 100.000 100.000 100.000 100.000
Eigenvalues Functio Eigenvalu % of Cumulative Canonical n e Variance % Correlation 1 1.15(a) 100.0 100.0 .663 a First 1 canonical discriminant functions were used in the analysis. Canonical Discriminant Function Coefficients Function 1 .184 .018 1.884 .053 -4.237
TRY_NEW_BRAND MONTHS_CURRENT_B RAND_USED SATISFIED_CURRENT_ OFFERING NO_OF_SHAMPOOS_L AST_TWO_YEARS (Constant)
Unstandardized coefficients Functions at Group Centroids SWITCH_TO_ANOT Function HER_BRAND 1 LOYAL CUSTOMER .793 NORMAL -.970 CUSTOMER Unstandardized canonical discriminant functions evaluated at group means ANALYSIS
iscriminant score cutoff for classifying new consumers into the two groups. = (0.793* 55 – 0.970 * 45 )/100 =-2.235 Classification Results(a) Predicted Group Membership Total LOYAL NORMAL LOYAL CUSTOMER CUSTOMER CUSTOMER 43 12 55 38 21.8 84.4 45 100.0 100.0SWITCH_TO_ANOT HER_BRAND Origina Coun LOYAL CUSTOMER l t NORMAL 7 CUSTOMER % LOYAL CUSTOMER 78.2 NORMAL 15.6 CUSTOMER a 81.0% of original grouped cases correctly classified.
The classification accuracy came out to be 81% which implies that the model is fairly acceptable. The eigen values are also greater than 1 and hence the factors are significant Attibute based perceptual mapping using Discriminant Analysis: Eigenvalues % of Cumulative Canonical Function Eigenvalue Variance % Correlation 1 4.749(a) 81.4 81.4 0.909 2 1.083(a) 18.6 100 0.721 a First 2 canonical discriminant functions were used in the analysis.
Structure Matrix Function 1 2 Mild .846(*) 0.278 smooth .782(*) -0.33 Fragrance .748(*) 0.076 Bounce 654 .727(*) Shine 0.541 .654(*) Lather -0.263 .926(*) Pooled within-groups correlations between discriminating variables and standardized canonical discriminant functions Variables ordered by absolute size of correlation within function. * Largest absolute correlation between each variable and any discriminant function a This variable not used in the analysis.
Functions at Group Centroids Function BRAND 1 2 1 0.935 -0.201 2 0.254 0.204 3 -0.343 0.127 4 -1.506 -0.123 5 0.659 -0.008 Unstandardized canonical discriminant functions evaluated at group means Analysis: When the important features of a shampoo are rated by respondents on a seven point Likert scale and fed to the SPSS software for Attribute based perceptual mapping analysis the attributes were clubbed under two functions. Function1 = Function of (Smooth hair on usage, mild shampoo, fragrance) Function2 = Function of (Bounce to hair, Rich Lather, Shiny hair on usage) We name the function F1 as Shampoo features and F2 as Usage effects.
doc_629176754.docx