Академический Документы
Профессиональный Документы
Культура Документы
MARKETING
RESEARCH
JAGROOK DAWRA
What does a manager do?
What is the most important skill he should have?
Why MR?
Why not to take decisions in absence of information?
Shooting from the Hip – Lee Iacocca
WHAT IS MR?
Information
Information processing
Tools and techniques
SWOT
Observation
One-on-one interviews
Complex statistical tools and techniques – Regression, Factor,
Cluster, etc.
FROM MARKETING RESEARCH TO MARKETING ANALYTICS…
MR MA
Focus is on analysis of survey-based carefully Focus is on analysis of existing data obtained from
collected data. past behavior.
Data size typically in a few hundreds. Data size enormous.
Focus on prediction from consumer’s psychological Focus is on prediction from past occurrences.
aspects – perceptions/ attitudes.
Data collection efforts are directed towards
Data collection is automatic and often mechanical.
obtaining a representative sample
4
Evolution of an organization and its information needs.
As an organization grows, its information needs also grow.
Is there a real need for marketing research?
Research takes time and costs money.
Value of information vs. Cost of information?
DEFINITION – AMA
Marketing research: the function that links the consumer, customer, and
public to the marketer through information – information used to identify
and define marketing opportunities and problems; generate, refine, and
evaluate marketing actions; monitor marketing performance; and improve
the understanding of marketing as a process.
Business problem/
Business objective Research objective Research method Interpretation Recommendations
symptom
PROCESS
Research Method Interpretation Recommendation
Alice: Oh, no, no. I was just wondering if you could help me find
my way.
Cheshire Cat: Well that depends on where you want to get to.
Alice: Oh, it really doesn't matter, as long as...
Cheshire Cat:Then it really doesn't matter which way you go.
Alice in Wonderland (Lewis Carroll)
Research objectives, when achieved, provide the information necessary to solve the problem
identified.
Research objectives state what the researchers must do.
IT ALL STARTS WITH DETERMINING OBJECTIVES
Types of Research
Exploratory Research: collecting information in an unstructured and informal manner.
Descriptive Research refers to a set of methods and procedures describing marketing
variables
Causal Research (experiments): allows isolation of causes and effects.
SOURCES OF DATA
Survey instrument
Asking the right question in the right way
DESIGN DATA
COLLECTION QUESTIONNAIRE
Sample plan refers to the process used to select units from the population to be included
in the sample. Gives you representativeness!
Sample size refers to determining how many elements of the population should be
included in the sample. Gives you accuracy!
DATA COLLECTION
Data collection is very important because, regardless of the data analysis methods used,
data analysis cannot fix bad data.
Data collection errors may be attributed to field workers or respondents.
Researchers must know the sources of these errors and the controls to minimize them.
DATA ANALYSIS
Data analysis involves entering data into computer files, inspecting data for errors, and
running tabulations and various statistical tests.
SPSS
Excel
Any other software
Unethical – sugging
Selling under the guise of research
What is frugging?
Fund-raising under the guise of research
The cover letter of a mail questionnaire says that it will "only take a few minutes to fill out." But
pretests have shown that at least fifteen minutes are needed to fill it out.
Read the case “The Coop” for the next class on Thursday/ Wednesday.
Open up a free account on qualtrics.com and familiarize yourselves with its features.
CR to form groups and mail the list to me by Wednesday.
CR to form a group id for students of MR, coordinate with the PGP/ IT office and
communicate the same to all of us.
EXERCISE 1 [IN-CLASS]
Please write the names of five people in the class you would like to form a group with.
EXERCISE 2 [OPTIONAL]
STUDY 1 OBJECTIVE
Objectives:
Measure consumer preference, price-feature trade-offs and willingness to pay.
STUDY 2 OBJECTIVES
Objectives:
This study explored the size and type of customer segments.
The study shall also profile the customers in these segments
based on various demographic factors, media habits,
psychographics, etc.
Key deliverables:
Consumer segments, size of segments, profiles of segments,
create top target groups (who they are, what they’re like, how to
reach them)
STUDY 3 OBJECTIVES
Objectives: Competitor analysis. Who are
the major competitors in Goa? How are they
positioned in consumer minds? What are the
possibilities of differentiating a new offering?
Key deliverables: competitor analysis,
positioning strategies.
STUDY 4 OBJECTIVES
STUDY 5 OBJECTIVES
Study of enablers in the real estate business in Goa. Different government/ legal bodies, Banks, etc., act as enablers
in the real estate business
Key deliverables: Detail activities of enablers, Interdependence of enablers on each other and their influence on
real estate business, Preference structure of property dealers.
Study was done and completed by students of MR course of PGP 12-14 batch, guided by two professors.
FUNDAMENTAL RESEARCH PARADIGMS
Problems
Converting
this P to K Instruments
The world
Knowledge
CONVERTING P TO K
Convert P to K Convert P to K
JAANE BHI DO
YAARON…
Interpret
Analyze data to verify if the
predictive statement is true
Convert P to K Convert P to K
WHAT IS A BETTER PROMOTION STRATEGY?
EDLP
6
5.5
Price in $
4.5
MRP: $5.50
4
3.5
Price HiLo
Time
6
HiLo: High-Low 5.5
Price in $ 4.5
3.5
3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Trip
Research question: What effect does HiLo promotion strategy have on trust on the store?
APPROACH 1: INTERPRETIVISM
Conduct an experiment:
Subject one group of people to EDLP strategy over a period of time.
Subject another group of people to HiLo strategy
Collect data on how much they trust each store.
Conclude
APPROACH 2: POSITIVISM
Positivism Interpretivism
Governed through explicitly stated theories and Focuses on understanding what is happening in a
hypotheses. particular context.
Researchers try to be emotionally neutral and make Let the data ‘lead the way’
a clear distinction between reason and feeling,
science and personal experience.
Positivists maintain a clear distinction between facts
and value judgments.
QUALITATIVE RESEARCH
FORMS OF DATA
60
“Every animate and inanimate object on earth will soon be generating data, including our homes, our car, and yes,
even our bodies.”
Anthony D. Williams, Author – ‘Wickinomics’
An average person today processes more data in a single day than a person in 1500s did in an entire lifetime.
Much of this data is qualitative, though the analysis done on it may sometimes be quantitative.
61
PROCESS OF CONDUCTING QUALITATIVE RESEARCH
cases
SAMPLING METHOD
TRIANGULATION
68
COMMON SOFTWARE THAT HELP
IN QUALITATIVE DATA ANALYSIS
NviVo
Atlas Ti
RQDA – Freeware uses R
69
USING NVIVO
70
USING NVIVO
71
72
Coding Memoing
Visuals Queries
73
WHERE DOES THE DATA COME FROM?
Interviews
FGDs
Secondary sources – company files, memos, minutes of meetings, etc.
Social media like Facebook, Twitter, Linkedin, web pages,YouTube, Picasa, etc.
Field notes
Etc.
74
Helps you appropriately code.
Coded text can be easily organized into themes
Emergent themes consistently appearing are often called propositions.
Propositions are generalizable and have very few exceptions.
Descriptive statistics – frequency of word occurrence, between and within group
analysis, cross tabulations, cluster analysis of word coding and similarity.
Kappa coefficient [triangulation statistic between researchers]
Does not establish cause and effect relationships.
But data can be exported to Excel/ SPSS and analyzed.
WHAT KIND OF ANALYSIS DOES NVIVO 75
DO?
WHY IS CODING IMPORTANT?
76
HERMENEUTICS
77
HISTORICITY AND HERMENEUTICS
78
THE HERMENEUTIC CIRCLE
It refers to the idea that one's understanding of the text as a whole is established by reference to the individual
parts and one's understanding of each individual part by reference to the whole.
79
Interpretation of a text needs the researcher to have some prior
knowledge. For instance, she needs to know the language, its
vocabulary and social conventions.
Social conventions change – SMS lingo: LOL, LMAO, TY, , , :-/ …
The more poetic a person is, more cryptic is his/ her language and
interpretation is less mechanical.
Mark Twain: “I don’t let my school interfere with my education”
80
81
Attitude towards the brand
Whose attitude – is there a
Attitude towards the product difference between groups?
Attitude towards the Men vs. women
performance Classification as per other
demographics.
Attitude towards the attributes
Amateur vs. professional
Attitude towards the aftersales photographer.
service
Usage – heavy vs. light user.
Attitude towards the ease of use
Usage – purpose (wildlife/ friends
Etc. and family/ scenery/ etc.)
EXAMPLE – ATTITUDE 82
83
POTENTIAL EXERCISES
POTENTIAL EXERCISES
Brand failures: using the cases in the book on brand failures (and also perhaps the
same author’s book on brand successes) build a generalized theory around what
causes brands to succeed/ fail. See if your emerging theory conforms to the ones that
already exist in the theory.
What type of comments/ content diffuse faster over social media?
May need to control for size of the network.
84
BLUEMOUNTAIN RESORTS
WHAT DID WE LEARN FROM BLUE MOUNTAIN RESORT CASE?
Objectivity: Definition of the task at hand – “Objective of playing darts is to hit the bulls eye”
Have defined a ‘rubrics’ for your projects – sent it over email.
“Siddarth’s ability to hit the bulls eye”
What is being measured? Siddarth? Ability?
Practicability: Online questionnaire may not be the right instrument for a study on Indian Kirana store shoppers.
Reliability
Validity
TRUE SCORE TEST THEORY Random sources
of error
X o XT X S X R
Observed Score
True Score
Systematic sources
of error
TYPES OF RELIABILITY…
Test-Retest: stability of the instrument – test it 2-3 times after a span of time.
Alternate form reliability: in test-retest reliability, the same instrument is given the second time. In alternate form,
the questions may change.
Split half reliability: if an instrument has six questions that measure ‘loyalty’, compare the results of first three with
the next three.
… TYPES OF RELIABILITY…
Internal consistency – Several questions are often used to measure one construct.
E.g. ‘Price consciousness’ is measured using the statements below:
… TYPES OF RELIABILITY
Internal consistency is said to be achieved when scores of items are highly correlated.
Cronbach’s alpha is used to calculate internal consistency.
Internal consistency is also called inter-item reliability.
Inter rater reliability: more than two judges judging a boxing match.
Three faculty members judging your SIP presentations.
TYPES OF VALIDITY…
Face validity: There is prima facie evidence to say that the results are accurate. E.g. the study predicts a market
share of 45%. Secondary sources also confirm this.
…TYPES OF VALIDITY…
Attitudinal angle: A customer is loyal to a brand if (s)he has a positive disposition towards it.
Please rate the following brands on a scale 1 to 5 (I = dislike, 5= like)
Colgate
Pepsodent
Close-up
Etc.
…TYPES OF VALIDITY…
Concurrent validity
Measuring attitudinal loyalty and behavioral loyalty should be highly correlated.
Convergent validity: Two or more methods of measurement leading to similar results. E.g. market share being predicted by
analysis of scanner panel data and a survey method.
Discriminant validity: can a set of questions discriminate between two constructs:
…TYPES OF VALIDITY
What do the following questions measure?
…TYPES OF VALIDITY
What do the following questions measure?
CLASSROOM EXERCISE…1
Rishi measures attitudinal loyalty using a few statements constructed by him and observes low effect on
repurchase intention.
Predictive validity?
Construct validity?
What needs to be done to rectify the situation?
CLASSROOM EXERCISE…2
Chandan measures trust as an overall construct and also measures it on its dimensions of ability, benevolence and
integrity. He observes a difference between the mean scores of both measurements.
Content validity?
How can this be corrected?
CLASSROOM EXERCISE…3
Jaysun measures trust using its three dimensions and gets the following Cronbach
alphas:
How important are the following attributes in an insecticide. Please give a rating
between 1 (=not important at all) and 10 (=Extremely important)
Brand
Safety
Ease of application
Action
Effectiveness
Smell
Price
As a manager in an insecticide company, can you use this information to design a product?
Sir Humphrey: "You know what happens: nice young lady comes up to you.
Obviously you want to create a good impression, you don't want to look a
fool, do you? So she starts asking you some questions: Mr. Woolley, are you
worried about the number of young people without jobs?“
Bernard Woolley: "Yes”
SH: "Are you worried about the rise in crime among teenagers?“
BW: "Yes”
SH: "Do you think there is a lack of discipline in our Comprehensive
schools?“
BW: "Yes”
SH: "Do you think young people welcome some authority and leadership in
their lives?“
BW: "Yes”
SH: "Do you think they respond to a challenge?“
BW: "Yes”
SH: "Would you be in favor of reintroducing Compulsory army Service?“
BW: "Oh...well, I suppose I might be.”
SH: "Yes or no?“
BW: "Yes"
Sir Humphrey: Alternatively the young lady can get the opposite result.“
Sir Humphrey: "Mr. Woolley, are you worried about the danger of war?“
Sir Humphrey: "Do you think there is a danger in giving young people guns
and teaching them how to kill?“
Sir Humphrey: "Do you think it is wrong to force people to take up arms
against their will?“
Disadvantages:
Difficult to code and interpret
Respondents may not give complete answers
Closed-ended question provides options on the questionnaire that can be answered quickly and
easily.
SCALE CHARACTERISTICS/ LEVELS
Likert scale,
interval
scale
Scale items
MULTI-DIMENSIONAL SCALE
Ability
Benevolence
Integrity
HOW DO YOU DECIDE ‘WHAT YOU ARE MEASURING
WAS THE CONSTRUCT OF INTEREST?
Collect data
Purify measure
Collect data
Adopted from
Churchill GA Assess reliability • Coefficient alpha, split-half, etc.
(1979), “A Paradigm
for Developing a
Better Measure of
Marketing
Assess Validity • Criterion Validity, Discriminant Validity, etc.
Constructs”, JMR,
pp. 64-73
Develop norms • Methods of summarizing distribution of scores
FACTOR ANALYSIS
THE MATH BEHIND FACTOR ANALYSIS METHOD
FACTOR ANALYSIS
A simple example
3 courses – Finance, Marketing and Business Policy
If 𝑋1 , 𝑋2 , 𝑋3 are marks (out of 10) obtained by the students,
1 3 6 5
2 5 3 3
3 9 4 8
4 4 5 7
5 7 9 5
These grades are functions of two underlying factors 𝐹1 [quantitative ability] and
𝐹2 [Verbal ability]
𝑋𝐹𝑖𝑛 = 𝑙𝐹𝑖𝑛,𝑄𝐴 𝐹𝑄𝐴 + 𝑙𝐹𝑖𝑛,𝑉𝐴 𝐹𝑉𝐴 + 𝜀1
Variable 𝑋𝒊 Loading on
𝐹𝑄𝐴 , 𝑙𝑋,𝑄𝐴 𝐹𝑉𝐴 , 𝑙𝑋,𝑉𝐴
Finance 𝑋𝐹𝐼𝑁 + 0
Marketing 𝑋𝑀𝑘𝑡 0 +
Business Pol 𝑋𝐵𝑃 0 +
ROTATION
SOME TERMS
Varimax – searches for a rotation of the original factors such that the variance of the
loadings is maximized.
Most popularly used method of rotation.
Qurtimax – minimizes the number of factors needed to explain each variable.
Equimax – compromise between Varimax and Quartimax
While all the above are orthogonal rotations, sometimes, oblique rotation provides a
better explanation of the factors. But this method is rarely used. They may be used
when one expects the factors to be slightly correlated.
BEHAVIORAL RESEARCH: THROUGH EXPERIMENTATION
SESSION 7
Experiments marked the emergence of modern science in the 16th and
17th century.
‘Bodies That Stay Atop Water or Move Within It.’ (Galileo, 1612)
‘On loadstone and magnetic bodies.’ (William Gilbert, 1600)
An experiment is a way of systematically observing a phenomenon.
AN EXPERIMENT IS DIFFERENT FROM MERE OBSERVATION
It is not possible to simultaneously give and not give discounts to the same consumers.
Cause probing research tries to create reasonable approximations to physically
impossible counterfactual.
Make probabilistic inferences.
1st video: Versace
CAUSAL RELATIONSHIP
The relationship may not be causal at all, but due to a third variable (often called a Confound).
Wisdom causes bot income and education to be high.
MANIPULABLE AND NON- MANIPULABLE CAUSES
Manipulable
Dosage, extent of discount, format of discount, color and type of packaging, etc.
Non – manipulable
Age, gender, number of times they have bought a particular brand, etc.
AFTER AN EXPERIMENT
B
A C
Casual Description
A affects C.
B also affects C.
A affects C by a larger amount as compared to B.
B moderates the relationship between A and C [A*B have a significant effect on C]
Or B mediates the relation between A and C
Casual explanation
SOME TERMS AND DEFINITIONS
Units
Treatments
Observations
Settings in which the study is conducted.
GENERALIZATIONS
Construct validity generalizations - inferences about constructs that research operations presents.
External validity generalizations - inferences about whether casual relationship holds over variations in persons,
settings, treatment and measurement variables.
VIDEOS
Hedonic Utilitarian
Fun, sensorial, spontaneous Functional. Sensible and useful
Eating out for pleasure Eating out to better understand local culture
Visiting a museum for fun Visiting the museum for collecting material for class
project.
LOCUS OF CHOICE
In study 1 the no choice condition involved an expert making the choice on behalf of the consumer. The
mitigation in satisfaction observed in the utilitarian condition could then be explained with participants’ belief that
the curator was at least as qualified as they were in selecting the most pedagogically effective alternative.
The aim of the study was to tie the results to personal causation and not to expertise.
STUDY 1: 2 BY 2 DESIGN
Imagine that you have been working suffering mild
Choice, Choice,
Utilitarian Hedonic really
back pains
hard and
during
a general
the semester
soreness.andYour
doctor tells
achieved important
you thatacademic
these aresuccesses.
just
You
symptoms
think thatof overall
a professional
body fatigue
massageand at a
recommends
local spa renowned
a professional
for the quality
massage ofat a
No No local spa
their massage
renowned
treatments
for thewould
quality of
choice, choice, represent
their massagea great
treatments.
way to reward
Imagineyourself.
having
Utilitarian Hedonic a gift certificate
Imagine having ato giftspend
certificate
at theto spaspend
that
at
offered
the spafourthat
types of massage
offered treatment.
four types of
massage
The are
You treatment
certificate
treatment.
allowedisto valid onlyany
select forof
The
massage
the
You certificate
four
are #3 (of is
massages
allowed the valid
to four only
select forof
massages)
any
massage
the #3 (of the four massages)
four massages.
STUDY 2 RESULTS
SETTING UP AN EXPERIMENT
PRESTUDY
A prestudy is done before the study when the researcher is not sure of the extent to which the stimulus’
manipulations.
E.g. high vs. low discount – how ‘high’ is ‘high’?
Treatment: The differential stimulus applied to different groups.
Pretest: information gathered before the treatment.
Posttest: information gathered after the treatment.
Control: the group not given a treatment or the group against which the dependent variable is studied.
EXPERIMENTAL DESIGNS
Random assignment of subjects in different treatment groups ascertains that only the treatment affects the
dependent variable and nothing else.
MANIPULATION CHECKS
Some questions included in the questionnaire that measure the ability of the treatment given to manipulate the
independent variable.
Manipulation checks are defined as a process that is used to verify if the experiment worked or works.
Where should the manipulation check questions be placed in an experiment?
ANALYSIS OF EXPERIMENTAL DATA
T-test
ANOVA
TWO SAMPLE COMPARISON
An analyst at a department store wants to evaluate a recent credit card promotion. To this end,
500 cardholders were randomly selected. Half received an ad promoting a reduced interest rate on
purchases made over the next three months, and half received a standard seasonal ad.
Compare the spending of the two groups.
H0: mprom = mno_prom
PAIRED SAMPLE COMPARISON
A physician is evaluating a new diet for her patients with a family history of heart disease.To test
the effectiveness of this diet, 16 patients are placed on the diet for 6 months.Their weights and
triglyceride levels are measured before and after the study, and the physician wants to know if
either set of measurements has changed.
Determine whether there is a statistically significant difference between the pre- and post-diet
weights and triglyceride levels of these patients.
H0W: mold_wt = mnew_wt
H0T: mold_tri = mnew_tri
ONE-WAY ANOVA EXERCISE-1
A sales manager wishes to determine the optimal number of product training days needed for new
employees. He has performance scores for three groups: employees with one, two, or three days of
training.
H0: m1 = m2 = m3
H0: s1 = s2 = s3
In response to customer requests, an electronics firm is
developing a new DVD player. Using a prototype, the marketing
team has collected data. ANOVA is being used to discover if
consumers of various ages rated the design differently.
Looking at the DVD data, market researchers ask:
Are the groups different from each other by age?
H0: m18-24 = m25-31 = …=m53-59
Are the two groups between the ages of 32 and 45 really different?
H0: m32-38 = m39-45
Y1 1X 1 2 X 2
LINE FITTING IN REGRESSION
We believe/ hypothesize: Revenues are related to
economy
=> Revenues (R ) are related to income (Y), interest
rates (I), prices (P), time (T)
=> R = a + b*Y + c*I + d*P + e*T
Assumptions:
All influences that are not included, have no effect
All influences that are included have a precise effect
DATA TYPES
Variable definitions
Grade: Grade obtained
Books: number of books owned
Attend: Number of classes attended
Model 1: Regress grade against books
Model 2: Regress grade against books with intercept term=0.
What happens to R-sq? is this a better model in terms of prediction?
σ 𝑌𝑖 −𝑌 2
Calculate mean square error for both the models 𝑀𝑆𝐸 = 𝑁
Which is the better model in terms of prediction?
DATA SET BOOKS.TXT: WHAT IS THE R-SQ IN EACH MODEL.
Grade
Grade Grade
Variable definitions:
Car: number of minutes per week spent on taking care of one’s car
Sex: Gender, 0=female 1=male
Age: age in years
Extro: projection of self-image through objects (in this case the car) they own.
Model 1: Run regression with all the variables.
Model 2: Run regression with all the variables but use the ‘stepwise’ feature.
CATEGORICAL INDEPENDENT VARIABLES
05 stress.txt,Variables:
Stress: measured on a scale
Job: 1 = primary school teacher; 2 = Secondary school teacher; 3 = College lecturer; 4 = old university lecturer; 5 = new
university lecturer.
Dummy variable coding
Effect coding
MODERATION (INTERACTION)
06 test.txt;Variables:
LEARN: learning environment
TEST: test environment
0 = dry; 1 = wet
07 stress1.txt:Variable definitions:
STRESS: the amount of stress in a person’s life.
EVENTS: the number and severity of events that cause stress (e.g. changing job, shifting house, etc.)
STATUS: 0 = single, 1 = married
Explore if:
Higher life event scores lead to higher levels of stress.
Married people feel less stress.
Having a partner reduces the impact of stressful events. Events cause lesser stress in case of married people.
Standardize the continuous variables. (why?)
MODERATION WHEN BOTH INDEPENDENT VARIABLES ARE
CONTINUOUS
03 books.txt
Grade: Grade obtained
Books: number of books owned
Attend: Number of classes attended
Is it important to read a number of books and attend number of classes together to score better?
Do not forget to normalize the two continuous variables.
MEDIATION - COMPLETE
Variable definition
enjoy: degree of joy derived
Buy
from reading a book.
buy: number of books
bought
read: number of books Enjoy Read
read.
Does buying books mediate
(also tell fully or partially)
between propensity to
enjoy books and the
number of books read?
PILGRIM BANK
CASE 1
Research Question:
Would encouraging transaction migration to lower cost channels improve customer profitability?
Research Objective:
Is there a difference between the profitability of on-line and off-line customers?
Secondary objective:
Is profitability different for different types of people?
DUMMY VARIABLES
Income (Rs.) D1 D2 D3
10000 and less 0 0 0
10000-20000 1 0 0
20000-30000 0 1 0
30000 and more 0 0 1
DUMMY VARIABLES
Positioning studies
Positioning- repositioning decisions
Product attribute finalization
Advertising decisions
MDS MAP (MENTAL MAP)
Outside Oriented
Going to church
Doing voluntary
service
Going to a beauty
parlor
Family Sunbathing
Powdering the baby Personal
Baking a cake
Relaxing in a bath
Using a room
Smoothening on a
freshener Wearing soft clothes hand lotion
Home Oriented
TWO PROCESSES
Proxcal
1= similar, 10 = dissimilar. ‘dissimilar’
Alscal
0= same pile, 1 = different piles
TYPICAL MDS QUESTIONNAIRE - PROXCAL
Using fabric Smelling a Going to a Etc.
softener flower parlor
Using fabric Small rating large rating
softener for small for large
perceived perceived
distance distance
(dissimilar)
Smelling a
flower
Going to a
parlor
Etc.
EXERCISE
Eg. Toothpastes
Germ killing
Fresh breath
Calcium
Strong teeth
White teeth
Colgate
Close up
Pepsodent
Brands may also form a part of the attributes.
QUESTIONNAIRE FOR ALSCAL
IIM- IIM -
Cards Raipur Ranchi
Instructions given to
respondent:
Put the B-schools in IIM IIM
different piles, keeping
similar ones in one pile and
Kashipur Rohtak
dissimilar ones in different.
You can make as many piles
IIM
as you want.
IIM Trichy
Udaipur
HOW DOES ROTATION OF AXIS HELP IN MDS
MDS by itself is not enough. Why?
What are the pros and cons of both the methods – Proxcal and Alscal?
Using factor analysis (factor scores) for positioning studies
CONJOINT ANALYSIS
SOME QUESTIONS THAT CONJOINT ANSWERS
Rational
Bundle Preference
of due to
Attributes Attributes
Preference
Brand due to
Brand
Is the basis of differentiation (rational vs. emotional) dependent on the type of product category?
What products are more likely to be sold on
Rational basis (Attributes)?
Emotional basis (Brand)?
CUSTOMER FOCUSED PRODUCT PLANNING
Our marketing actions
Product, price
-10
-15
-20
WHICH OF THESE PRODUCTS WOULD BE MORE
PREFERRED?
Product A Product B A UA B UB
More preferred
Typically many factors or attributes are included.
Factors/ Attributes often include price and brand name.
DECIDING ATTRIBUTES
Related to consumer choice (based on qualitative market research like FGDs or Interviews)
Distinguish different product offerings.
Managerially relevant.
DECIDING LEVELS OF ATTRIBUTES
Cards
NEXT CLASS
Case: Sunbeam
CONSUMER CHOICE
Solibloc Airstick
Solibloc Airstick
Fits in a Fits in a
pocket book bag
0.5 min 10 min
Easy to Medium to
inflate inflate
Rs. 2000 Rs. 500
WHAT WOULD HAPPEN IF A NEW COMPANY
RELISTICK WERE TO INTRODUCE A PRODUCT…
Solibloc Airstick Relistick
Depth
Hamam Clinic Plus Ponds
Clinic All
Clear
Clinic Plus
Ayurvedic
Lirl
Product Line Brand Family Brand Portfolio
REDESIGNING OF PRODUCT LINES
• Mature
• Redesign the product lines to make them more responsive to needs and dictates of its markets.
• To optimize market share of each product category.
• Goals:
– What models should be in the line?
– What should their physical appearance be?
– What should their performance characteristics be?
PROCESS
Related to consumer choice (based on qualitative market research like FGDs or Interviews)
Distinguish different product offerings.
Managerially relevant.
DECIDING LEVELS OF ATTRIBUTES
design.exe
Reorder.exe
ESTIMATING PART-WORTH UTILITIES
Using Linmap
Estimated for each respondent
Enter data in a notepad in the following format
RES1 6 4 15 9 5 8 16 10 13 2 3 11 7 14 12 1 20 19 18 17
Change the extension from .txt to .dat
Pump.dat Checkdat.exe
ASSIGNMENT 2
Adding (or subtracting) the same constant to all partworths of an attribute does not
alter the relative values of overall preference.
The fact that part-worth for 0.5 min is positive and 10 min is negative does not mean
that the respondent ‘likes’ 0.5 min and ‘dislikes’ 10 min. zero point on the partworth
scale is arbitrary.
It is not meaningful to compare the part-worth of the level of one factor with some
partworth of some other factor.
Units of measurements are arbitrary. Transformations can be done. But they should
be done to all partworths of all attributes.
OTHER SIMULATIONS AND ANALYSES