Beruflich Dokumente
Kultur Dokumente
Positioning in
Detergent Industry
Factor Analysis Approach
Analytical Marketing
Under guidance of
Prof. Srinivas Prakhya
Prepared by:
Dharmesh Gandhi
Alok Shukla
Kaveri Ingale
Kumar Ashutosh
Venkata Phani Prasad Chavali
Apurva Pushpen Thanawala
Contents
1
Introduction...........................................................................................................................................3
Approach ...............................................................................................................................................3
No rotation ............................................................................................................................................11
4.1.2
Varimax rotation....................................................................................................................................12
4.1.3
4.2
4.2.1
No rotation ............................................................................................................................................15
4.2.2
Varimax rotation....................................................................................................................................16
4.2.3
Promax rotation.....................................................................................................................................19
5.2
5.2.1
No rotation ............................................................................................................................................26
5.2.2
Varimax rotation....................................................................................................................................27
5.2.3
Promax rotation.....................................................................................................................................28
2|Page
1 Introduction
The case we have taken here is of the detergent industry. The motive was to select a brand that has not been
doing well and do a repositioning exercise. From the newspapers sources, we found that approximately share
for the HULs Surf Excel is 14%, P&Gs Ariel is 3% and Henkels Henko is 1%. Considering the fact that Henkel is
operating in India since early 90s, and still has not make any significant progress in terms market share, we
selected this brand for repositioning.
The idea is to find out the core factors driving the consumer choice. The approach was to do a dip stick survey
to find out the observable attributes that are important to the consumer (decision maker). From this we try to
find the unobservable latent factors that are driving these attributes. Then we map the chosen brands on these
factors to get a perceptual map.
We also reduce the consideration set to just the competing brands of Henko to see if the results make more
sense. This perceptual map is then used to get an insight of where these different brands figure in the minds of
the consumer. Hopefully, then it can become very clear where Henko is going wrong and an approach to correct
that can be figured out.
2 Approach
Firstly, we do a qualitative analysis to get a feel of the situation on the ground. To do that we analyze the
advertisements for the different brands in the industry and try to understand what the brand managers are trying
to portray and what is the positioning stance taken by the different brands.
To be able to figure out the factors, we needed to have data at our disposal which would quantify the various
perceptions of different users of detergents.
To design our survey attributes, we first went ahead with a dip stick survey to understand the various factors that
are considered by a consumer while buying his/her choice of detergent powder.
The factors thus distilled out of the dip-stick survey were as follows.
Wash Quality
Price
Fragrance
Packaging
Appropriateness for washing machine
Brand Recall
Indian Institute of Management, Bangalore
3|Page
Gentle on Hands
We gave equal weight age to these attributes. Probably, the weight ages of these attributes would be different in
a consumers mind. However, we shall start with this and see whether the results are intuitive enough.
Once these variables were decided, the next task was to define the questionnaire. The idea of the end result of the
survey was to be able to create a perception map of the brands in question and figure where Henko figures among
those.
To meet the goal we decided to pick up well known brands (along with Henko) to figure in our questionnaire. The
brands thus picked up were Surf Excel, Ariel, Rin and Nirma.
Sample set of questions (that were put for each of these brands) is as follows.
The survey was a web survey and was floated to decision makers mainly in the IT segment. We got around 80
responses which we used for the analysis.
3 Qualitative Analysis
Brand Positioning and Imagery as conveyed through the advertisements
Brand
S.No.
Proposition/Positioning
Advertising Imagery*
Name
Specific emphasis on school kids
.Mothers now have the freedom to
Surf Excel
4|Page
Wheel
Tide
Henko
portay country of origin (Germany)
Middle and premium segment.
consumers.
Mr White
Rin
5|Page
Ariel
Nirma
6|Page
The scree plot indicates the fact that most of the variance is explained by the first factor.
7|Page
Correlation Matrix
Correlation
Wash
Quality
1
Wash Quality
Fragrance
0.708
Packaging
0.631
Washing
Machine
Appropriateness
0.693
Gentle
on
hands
0.589
Brand
Recall
0.595
Price
0.526
Fragrance
0.708
0.672
0.681
0.612
0.481
0.506
Packaging
0.631
0.672
0.631
0.521
0.497
0.497
Washing
Machine
Appropriateness
0.693
0.681
0.631
0.591
0.467
0.651
Gentle on
hands
0.589
0.612
0.521
0.591
0.427
0.445
Brand Recall
0.595
0.481
0.497
0.467
0.427
0.335
Price
0.526
0.506
0.497
0.651
0.445
0.335
There is a significant amount of correlation across the factors underscoring the usefulness of factor analysis. (They
have to be correlated if they are representing the same underlying set of latent variables).
Determinant=.02 => there is no singularity in the data.
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure
.908
of Sampling Adequacy.
Bartlett's Test of Sphericity
Approx. Chi-Square
df
Sig.
1549.372
21
.000
KMO value of 0.908 indicates suitability of factor analysis. Bartletts test is also significant.
8|Page
Communalities
Extraction
Wash Quality
.739
Fragrance
.689
Packaging
.586
Washing Machine
.793
Appropriateness
Gentle on hands
.491
Brand Recall
.445
Price
.544
However if we see the communalities matrix, the Gentle on hands and Brand Recall attributes have less than
50% of their variance explained. Even price and packaging have very less variance explained by the 2 factors.
Hence we might need to move onto 3 factors because of this significant data loss.
9|Page
Total
% of Variance
Cumulative %
4.023
57.478
57.478
.264
3.772
61.250
10 | P a g e
4.1.1 No rotation
The factor loadings for two factor analysis shows that the first factor describes all the variables and the second
factor does not give any specific information about the loadings of the attributes.
Factor 1
Factor 2
Wash Quality
.842
.171
Fragrance
.823
.107
Packaging
.761
.085
.862
-.223
Gentle on hands
.699
.056
Brand Recall
.609
.273
Price
.675
-.298
2 Factor No Rotation
11 | P a g e
4.1.2
Varimax rotation
Factor 1
Factor 2
Wash Quality
.729
.455
Fragrance
.671
.488
Packaging
.610
.462
.473
.755
Gentle on hands
.545
.440
Brand Recall
.630
.221
Price
.285
.681
Varimax rotation provides two factors which explain two choices. First factor explains the brand, wash quality,
packaging etc. The second factor explains washing machine appropriateness and price.
Total Variance Explained
Extraction Sums of Squared Loadings
Factor
Total
% of Variance
Cumulative %
% of Variance
Cumulative %
4.023
57.478
57.478
2.353
33.620
33.620
.264
3.772
61.250
1.934
27.630
61.250
Here, the variance is distributed across the 2 factors in a much more even manner and a perceptual map makes
more sense here.
12 | P a g e
4.1.3
Pattern Matrix
Factor 1
Factor 2
Wash Quality
.752
.128
Fragrance
.638
.223
Packaging
.567
.230
.129
.783
Gentle on hands
.485
.248
Brand Recall
.784
-.151
Price
-.099
.816
1.000
.811
.811
1.000
Here you can see that the factors are highly correlated
13 | P a g e
Communalities
Extraction
Wash Quality
.708
Fragrance
.867
Packaging
.580
Washing Machine
.805
Appropriateness
Gentle on hands
.485
Brand Recall
.707
Price
.534
The communalities show a better result compared to the 2 factor solution. However the Gentle on Hands is
still not adequately explained.
Extraction Sums of Squared Loadings
Total
% of Variance
Cumulative %
4.056
57.945
57.945
.366
5.234
63.179
.262
3.742
66.921
14 | P a g e
The 3 factor solution explains around 5% more variance than the 2 factor solution and the communalities also
suggest a better explanation of the attributes.
4.2.1
No rotation
The factor loadings for three factor analysis for all the five products shows that the first factor describes all the
variables and the second & third factor does not give a lot of information about the loadings of attributes.
Factor 1
Factor 2
Factor 3
Wash Quality
.831
.130
.012
Fragrance
.882
-.165
-.248
Packaging
.761
.023
-.014
Washing Machine
.842
-.083
.300
Gentle on hands
.696
-.014
.009
Brand Recall
.631
.552
-.061
Price
.647
-.098
.325
Appropriateness
15 | P a g e
4.2.2
Varimax rotation
Factor 1
Factor 2
Factor 3
Wash Quality
.481
.488
.488
Fragrance
.380
.805
.273
Packaging
.443
.500
.364
.756
.395
.279
Gentle on hands
.430
.458
.300
Brand Recall
.208
.235
.780
Price
.661
.257
.175
Price,
Washing
machine
appropriateness
Wash Quality, Packaging, Gentle to hand
Brand Recall
Three-factor perceptual map has been broken into three two-dimensional perceptual map for better
understanding. The three maps are drawn against three factor solutions achieved in the previous section.
Basic Features
Washing Machine
Appropriateness (Premium-ness)
16 | P a g e
The first perceptual map shows that Henko is placed poorly with factor 2, which means wash quality and
packaging is not so good. Also it is not perceived as gentle to hand. It scores positive on the washing machine
appropriateness and price, but again it scores low with respect to other brands.
Brand
Recall
Brand
Recall
Basic Features
This perceptual map shows brand positions with respect to Factor-2 and Factor-3. Henko is very poorly placed if
these two factors are considered.
17 | P a g e
Total
% of Variance
Cumulative %
% of Variance
Cumulative %
4.056
57.945
57.945
1.810
25.856
25.856
.366
5.234
63.179
1.622
23.169
49.024
.262
3.742
66.921
1.253
17.897
66.921
The rotation spreads the variance across these factors which might make the perceptual map more relevant.
18 | P a g e
4.2.3
Promax rotation
1.000
.818
.681
.818
1.000
.730
.681
.730
1.000
Factor 2
Factor 3
Wash Quality
.269
.337
.318
Fragrance
-.059
1.046
-.095
Packaging
.235
.435
.151
Washing Machine
.837
.085
-.016
Gentle on hands
.264
.396
.084
Brand Recall
-.070
-.051
.922
Price
.817
-.051
-.070
Appropriateness
The factor 1 is the clearly the washing machine appropriateness factor which commands a premium in the
market probably because of the target segment differentiation.
19 | P a g e
The axes are shown orthogonal just for the sake of representation
20 | P a g e
The analysis till now has been over the five brands that have been assumed to constitute the industry.
However, Henko effectively competes only with Surf Excel and Ariel. So it might make sense to only look at the
data for Surf Excel, Henko and Ariel.
21 | P a g e
Packaging
0.623
Washing
Machine
Appropriateness
0.666
Gentle
on
hands
0.578
Brand
Recall
0.753
Price
0.511
1
0.618
0.618
1
0.608
0.603
0.535
0.459
0.547
0.69
0.343
0.504
0.666
0.608
0.603
0.492
0.631
0.483
Gentle on hands
0.578
0.535
0.459
0.492
0.523
0.235
Brand Recall
0.753
0.547
0.69
0.631
0.523
0.544
Price
0.511
0.343
0.504
0.483
0.235
0.544
Wash
Quality
1
Fragrance
0.653
Fragrance
Packaging
0.653
0.623
Washing
Machine
Appropriateness
Wash Quality
There is a significant amount of correlation across the factors underscoring the usefulness of factor analysis.
(They have to be correlated if they are representing the same underlying set of latent variables).
22 | P a g e
As was the case with the all-brands case, most of the variance can be explained by one factor itself.
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure
.891
of Sampling Adequacy.
Bartlett's Test of Sphericity Approx. Chi-Square
df
Sig.
936.039
21
.000
KMO value of 0.891 indicates suitability of factor analysis. Bartletts test is also significant.
Indian Institute of Management, Bangalore
23 | P a g e
Communalities
Extraction
Wash Quality
.737
Fragrance
.881
Packaging
.610
Washing Machine
.583
Appropriateness
Gentle on hands
.402
Brand Recall
.804
Price
.390
As can be seen in the communalities output, the Gentle on Hands and Price attributes are not explained
significantly
Extraction Sums of Squared Loadings
Factor
Total
% of Variance
Cumulative %
3.964
56.629
56.629
.441
6.305
62.934
The 2 factors explain around 63% of the variance which is about the same as when compared with the all
brands case.
The factor loadings for two factor analysis shows that the first factor describes all the variables and the second
factor does not give any specific information about the loadings of the attributes.
Indian Institute of Management, Bangalore
24 | P a g e
Factor 1
Factor 2
Wash Quality
.842
.168
Fragrance
.854
-.390
Packaging
.771
.124
.758
.094
Gentle on hands
.633
.001
Brand Recall
.814
.376
Price
.543
.309
2 Factor No Rotation
So looking at the above scenario, it makes sense to look at the 3 factor solution.
.755
Fragrance
.999
Packaging
.624
Washing Machine
.583
Appropriateness
Gentle on hands
.549
Brand Recall
.776
Price
.514
All the variables have more than 50% of their variance explained which can be termed satisfactory.
Indian Institute of Management, Bangalore
25 | P a g e
5.2.1 No rotation
The factor loadings for three factor analysis shows that the first factor is strongly loaded with most of the variables
but mostly with fragrance, wash quality and packaging. The second factor is strongly loaded with brand recall ,
wash quality and price. The 3rd factor has relatively weaker factor loadings.
Factor 1
Factor 2
Factor 3
Wash Quality
.656
.560
.101
Fragrance
.999
-.005
.000
Packaging
.621
.472
-.124
Washing Machine
.611
.458
-.013
Gentle on hands
.537
.337
.383
Brand Recall
.551
.687
.003
Price
.346
.537
-.324
Appropriateness
So this makes a strong case for rotating the solution and getting a cleaner solution that is more intuitive.
26 | P a g e
5.2.2
Varimax rotation
Wash Quality
Factor 1
.552
Factor 2
.582
Factor 3
.333
Fragrance
.267
.355
.895
Packaging
.605
.346
.373
Washing Machine
Appropriateness
.527
.425
.353
Gentle on hands
.186
.660
.280
Brand Recall
.679
.525
.201
Price
.695
.117
.130
In contrast with the all brands case, the washing machine appropriateness is not that strong an attribute in driving
the factor loadings.
Factor-1: Price, brand recall, packaging
Factor-2: Gentle on hands, brand recall, wash quality
Factor-3: Fragrance
Total
% of Variance
Cumulative %
Total
% of Variance
Cumulative %
2.900
41.430
41.430
1.998
28.540
28.540
1.622
23.172
64.602
1.491
21.294
49.834
.278
3.968
68.570
1.312
18.736
68.570
Around 6% more variance is explained by a 3 factor solution compared to a 2 factor solution. The variance is
much more equitably distributed
27 | P a g e
5.2.3
Promax rotation
Pattern Matrix
The significant factors have been highlighted.
Factor 1
Factor 2
Factor 3
Wash Quality
.351
.519
.058
Fragrance
-.028
.030
.995
Packaging
.589
.062
.207
Washing Machine
.420
.254
.163
Gentle on hands
-.244
.895
.036
Brand Recall
.607
.409
-.127
Price
.935
-.270
-.041
Appropriateness
As can be seen from the above factor loadings, the promax gives the cleanest solution which is expected given the
extent of correlation amongst the factors.
Factor Correlation Matrix
Factor
1.000
.798
.628
.798
1.000
.722
.628
.722
1.000
28 | P a g e
29 | P a g e
The axes are shown orthogonal just for the sake of representation
30 | P a g e
It is very clear where Henko is lagging. It does not have the brand strength to compete with either Surf Excel or
Ariel in its category. It is interesting to note that as soon as we remove the brands Nirma and Rin, the washing
machine appropriateness as a driving factor is no longer there. Probably, that is what defines this particular
segment.
So, the obvious thing as expected would be to build its brand through better advertising, marketing reach and
channel reach to reach its target segment.
31 | P a g e