Professional Documents
Culture Documents
Analysis
Analysis of the
relationship between
Bi Variate two, three or more
analysis variables
simultaneously.
Factor and
Uni Variate Regression Analyses
Looks at variables (questions) are examples
analysis one at a time. Frequencies and
averages are examples
Factor analysis Used to: examine inter-relationships S I ON variables, with the aim
between
of data reduction, or to identifySES
underlying themes (eQ and WBs);
HIS from survey data (eQ and WBs)
build Key performance indicators
T
D I N
ERE
V
Correspondence CO graphical summary of brands’ positioning – in relative or
Provide
Analysis/Biplots and absolute terms – across a range of perceptions/images (Used in
Mapping WBs and ad hoc studies)
-1 1
Product price & market share 0 Perfect positive correlation
Total cost=fixed + variable costs
• Correlation is a number between -1 and 1 that measures the linear association between
two variables (questions often attitudinal statements in MR)
• Correlation does not imply causation
• Zero or low correlation does not imply that there is no association at all, just no linear
association
10
Correlation of 0.17 is low, but there is visible
8 association between visitation and commitment
Commitment to Company X
6
4
2
0
0 2 4 6 8 10
y end (Dependent)
12
• gradient=2, intercept=1 6
4
2
• Not seen in Market Research eg electricity bill. Total costs 0
0 1 2 3 4 5 6 7 8 9
25
• Approximate linear relationship 20
y = 3.5x - 3.3
y independent
• 15
• all points lie close to the line 10
• gradient=3.5, intercept=-3.3 5
• Line is a good fit (97%) 0
0 1 2 3 4 5 6 7 8 9
x independent
• Uses
– Use FA to reduce number of questions asked in future research
waves
– Use factors with other techniques (eg regression and cluster
analyses) to analyse data more successfully with uncorrelated
data
Q7 – Uncertain 0.767
Q3 – Nervous 0.697
customers
Has expertise in providing insurance solutions
X
Line of best fit: Y = 1.8 + 2.15*X
8
Brand Equity Index
7
6
5
4
3
2
y = 0.118x + 0.485
1 2
R = 0.80
0
0 10 20 30 40 50
Brand Share (val)
1 Very dissatisfied 1
• Satisfaction = the 2 2
dependent variable 3 3
4 4
5 Very Satisfied 5
DK/Can't say (Do not read out) 6
Factor 3: Distribution/Established
Satisfaction
Satisfaction 40%
Factor 2: Reputation*
12%
Awareness
0.09
I * Customer Focus H
M I
* Overall Quality
P G
O* Product hard to use * Delivery time H
R
T
A * Emergency Responsive Rep *
N ordering L
C O
E W
P ER FO R M A N C E
[Topic of Presentation] December 6, 2010 Page 40 Confidential & Proprietary
Copyright © 2007 The Nielsen Company
ML
Multiple Lineear Regression Summary R
• Linear Regression
– eg Key Driver analysis Multiple Regression Model:
– usually based on attitudinal data
– The relationship is linear (ie a
Y = c + b1x1 + b2x2 + b3x3 + ..+ e
straight line can describe the
relationship) and is additive in
nature
– Based on correlation
– Use model fit R2 (adjusted) 30
20
y independent
15
COCO POPS FRUITY BIX KELLOGGS CORN KELLOGGS RICE NUTRI-GRAIN VITA BRITS WEET-BIX WEETBIX MILO
FLAKES BUBBLES CRUNCH
Good source of energy 11% 12% 26% 12% 46% 34% 63% 2% 6%
Children like the taste 69% 12% 23% 34% 32% 8% 22% 1% 8%
Good for kids 10% 12% 31% 19% 22% 33% 66% 2% 3%
Like the taste 37% 12% 43% 26% 38% 19% 47% 2% 5%
High in fibre
Good source of energy
[Topic of Presentation] December 6, 2010 Page 47 Confidential & Proprietary
Copyright © 2007 The Nielsen Company
Correspondence Map: Example 1 Ma
pp
ing
C o n v e n ie n t to g e t to
S ta ff p ro v id e g o o d s e r
F o o d a n d G ro c e rie s a
[Topic of Presentation] December 6, 2010 Page 51 Confidential & Proprietary
Copyright © 2007 The Nielsen Company
Ma
Correspondence Map: UK ST Data Example 2 pp
ing
COCO POPS FRUITY BIX KELLOGGS CORN KELLOGGS RICE NUTRI-GRAIN VITA BRITS WEET-BIX WEETBIX MILO
FLAKES BUBBLES CRUNCH
Good source of energy 11% 12% 26% 12% 46% 34% 63% 2% 6%
Children like the taste 69% 12% 23% 34% 32% 8% 22% 1% 8%
Good for kids 10% 12% 31% 19% 22% 33% 66% 2% 3%
Like the taste 37% 12% 43% 26% 38% 19% 47% 2% 5%
Biplots use
absolute
data values
CA.... CA can...
• Summarises large amount of • Be misinterpreted - map
information from tables presented visually, highlights
succinctly and visually relative strengths of brands
• Identifies relationships between • mean numbers from analysis
statements, between brands & difficult to interpret
between statements and brands • Be hard to compare different
• Removes halo effects of brands different maps - how different
as it is a relative analysis they are?
• Probably need to show absolute • Should be described in
scores as well qualitative, or passive
language...eg ’brands tends to
But.... be’ or ‘near to’
Agenda
• Review BEI Calculation & Interpretation
• Review BEM
• Image Analyses
8
Brand Equity Index
7
6
5
4
3
2 y = 0.118x + 0.485
1 2
R = 0.80
0
0 10 20 30 40 50
Brand Share (val)
Age Count BEI Std Dev Count BEI Std Dev Count BEI Std Dev
1.00 16-19 years 165 6.5 3.407 165 3.7 3.159 165 1.8 1.828
2.00 20-24 years 155 6.0 3.452 155 3.3 3.002 155 1.4 1.653
3.00 25-29 years 122 4.2 3.499 122 3.9 3.559 122 1.3 1.535
4.00 30-39 years 130 5.5 3.501 130 3.6 3.235 130 1.2 1.354
Normative Database
5.0 and
5% • Only about 15% of brands
above
command a brand equity
Strong score of more than 3.0
3.1 - 5.0 10% brands • About 35% are in the
range 1.0 - 3.0
1.0 - 3.0 35% • Majority of brands have
an equity score of less
than 1.0
Less than
50%
1.0
0% 10% 20% 30% 40% 50% 60%
Ultimate Objective:
Ensure Success of Brand and Company Profitability
Brand A
10
Brand C
8
Brand B
Brand Equity Index Score
6 Brand D
4
Aheadof3.6
all other Significantlylower
brands than BrandA,
2
1.8
0.9 aheadofBrandsB 0.9
0 &D
Brand A Brand B Brand C Brand D
Whatconsumersknow Whatconsumersdoorfeel
Awareness ConsumerLoyalty
Consideration
Attributes BrandEquityIndex
Benefits
Attitudes PricePremium
Known Brand/Image
(53%)
Brand
Equity Index
Awareness
(16%)
R2 =55%
Nutrition/Health
(14%)
• Image Factor
– A brand for me Health
– Tastes good
– A brand that makes me feel good, Image
etc
• Health Factor Consider
– Made from whole soy beans
– No cholesterol Awareness
– No lactose, etc
0 20 40 60
% Contribution to BEI of
Attribute
D is t ic t iv e n e s s S c o r e s
C o n v e n ie n t t o g e t t o
S t a f f p r o v id e g o o d s e r
F o o d a n d G r o c e r ie s a r
• Green better than average
• Red worse than average
E v e r y t h in g I n e e d in t h
[Topic of Presentation] December 6, 2010 Page 80 Confidential & Proprietary
Copyright © 2007 The Nielsen Company
MVA Summary
Conclusions and final obervations
Factor analysis Used to: examine inter-relationships between variables, with the
aim of data reduction, or to identify underlying themes (eQ and
WBs); build Key performance indicators from survey data (eQ and
WBs)
Interdependence Dependence
Identify structure of Prediction of Dependent
interrelationships variables by Other
independent variables
Is the structure of
relationships How many
among.? variables are being
predicted or
explained?
Variable Cases/ Objects
s Respondent
s One dep. Several Multiple
variable in a dep. relationship
single Variables in s of dep.
relationship single and indep.
ir, Anderson Tatham, Black: Multivariate Data Analysis Prentice Hall relationship variables
Cases/
Variables Objects
Respondent
s
How are the
attributes
Factor Cluster measured?
analysis analysis
Metric Nonmetric
Multidimensional
scaling Correspondence
analysis
ir, Anderson Tatham, Black: Multivariate Data Analysis Prentice Hall