Conjoint Analysis

• A method that examines a product as a

whole. The notion being that the whole is
greater than the sum of the parts.

• Examining attributes in isolation may lead

to a limited understanding of how people
will react to a product.
• CA is a multivariate technique used specifically to understand
how respondents develop preferences for products or services.
It is based on the simple premise that consumers evaluate the
value or utility of a product / service / concept / idea (real or
hypothetical) by combining the utility provided by each attribute
characterizing the product / service / concept / idea
• CA is a decompositional method. Respondents provide overall
evaluations of products that are presented to them as combos of
attributes. These evaluations are then used to infer the utilities
of the individual attributes comprising the products. In many
situations, this is preferable to asking respondents how
important certain attributes are, or to rate how well a product
performs on each of a number of attributes
C A permits:

 examination of tradeoffs people make;

 determination of interactions; and,
 predict market shares.
 Managerial uses
1. Find the product with the optimum set of
features
2. Determine the relative importance of each
feature in consumer choices
3. Estimate market share among products
4. Identify market segments
5. Evaluate the impact of price changes or
other marketing mix decisions.
 Commercial Applications
• Technique is widely used by consumer and industrial product
companies, service companies, marketing research, advertising and
consulting firms
• Over 400 commercial applications per year even in the mid 80s
• Types of applications include
– Consumer durables: automobiles, refrigerators, car stereos, condos, food
processors, HDTV
– Industrial products: copy machines, forklift trucks, computer software, aircraft
– Consumer nondurables: bar soaps, hair shampoos, disposable diapers
– Services: car rentals, credit cards, hotels, performance art series, rural health
care systems, BART
– Other: MBA job choice
 A Survey

• Familiarity & usage of value assessment methods

• 58 industrial firms in the top 125 of the Fortune 500 list
• 16 market research firms from the top 40
 Survey Results
Method Industrial Market Research
Familiarity % Usage % Familiarity % Usage %
Internal Engg. 61.3 42.5 - -
Assessment

Field value-in-use 63.8 36.3 25 5

Focus group 92.5 60 90 60

Direct survey 91.3 48.8 85 55

Benchmarks 83.8 27.5 80 25

Conjoint 75 28.8 90 60

Compositional 45 10 40 5
methods
 Steps in CA
• Identification of respondents
• Identification and definition of attributes in customer
language
• Specification of attribute variation and levels
• Creation of objects (experimental design)
• Creation of instrument, including socioeconomic,
demographic and usage questions
• Sampling plan
• Data collection
• Data analysis: Typically, regression analysis separately by
respondent
• Market simulation: exploration of “what-if” questions
 Terminology
 Attributes
 Features that describe a product. For example, frozen
conentrate orange juice might be described by: price,
pulp, blend and sweetness.
 Attribute Levels
 Price may be described on a continuous scale; pulp
might be described as low or high; blend might be
described as pure or mixed (e.g. orange-guava); and,
sweetness might be described as high or low.
 Trade-offs
 How much would you pay for a new car?
 Would it matter whether it was a Lada or a Rolls Royce?
 How much more would you pay for a Rolls Royce?
 Part of the impact may come from a main-effect.
 Main Effects
 A main effect refers to the impact of changing
levels of an attribute.
 What is the impact of decreasing price by $5.00?
 What would happen if we changed the color from
blue to red?
 Interactions
 What would be the impact of increasing the price of a 6-
pack of beer by $2.00?
 Will the impact of this price increase be the same if it is
Wildcat or Heineken?
 “Kool-Aid” example.
 “Evenings and Weekends Free” example.
 The interaction between brand and attribute values may be
examined.
Rating Task

Price $1.00

Pulp Low

Fruit Pure

Sweet High

Not At All Attractive 1 2 3 4 5 6 7 Very Attractive

Choice Task

SET 1 Alternative A Alternative B Alternative C

Price $1.00 $1.00

Pulp Low High

Neither
Fruit Pure Pure A Nor B

Sweet High Low

Choose only
  
one option.
 Conjoint Analysis

Ratings Based Choice Based

 Conjoint Analysis

Ratings Based Choice Based

 Rating-Based

Step 1: Determine Relevant Attributes

Step 2: Determine Relevant Attribute Levels

 Rating-Based

Step 1: Determine Relevant Attributes

Step 2: Determine Relevant Attribute Levels

You want enough attributes and attribute levels to make the task
realistic, but not enough to make the task explode. Exploratory
research is particularly valuable in aiding this decision.
 Rating-Based

Step 3: Develop Experimental Design

Step 4: Obtain Evaluations/Ratings.

The world according to Fred.
Equal to the “Price” column
multiplied by the “Pulp” column.
 Note that this is due to the type of
coding we used. If dummy coding
had been used, a different method
of coding for interactions would
Equal to the “Price” column have to be employed. Only 2-way
multiplied by the “Pulp” column. interactions are presented here.
The same method would apply for
the 3-way interactions and the 4-
way interactions.
 Rating-Based

Step 7: Run Regression!

The “magic” p < 0.05.
 Rating-Based

Step 8: Plot coefficients and/or calculate market share.

 Price Main Effect

0
$1.5 $1
 Pulp Main Effect

0
Low High
 Fruit Main Effect

4.5

3.5

2.5

1.5

0.5

0
Pure Mixed
 Sweetness Main Effect

4.6

4.5

4.4

4.3

4.2

4.1

3.9

3.8

3.7
High Low
 Price by Pulp Interaction

$1.5 (-1)
3
$1 (1)

0
LoPulp(-1) HiPulp(1)
 Rating-Based

• Shares must be calculated by determining

the choice that each person would have
made in a given situation.

• Do people typically make judgments in

purchase situations?
 Conjoint Analysis

Ratings Based Choice Based

 Choice-Based

Step 1: Determine Relevant Attributes

Step 2: Determine Relevant Attribute Levels

 Choice-Based

Step 3: Develop Experimental Design

Step 4: Obtain Evaluations/Ratings.

Step 7: Put Into Survey

SET 1 Alternative A Alternative B Alternative C

Price $1.00 $1.00

Pulp Low High

Neither
Fruit Pure Pure A Nor B

Sweet High Low

Choose only
  
one option.
SET 2 Alternative A Alternative B Alternative C
Price $1.00 $1.50

Pulp Low Low

Neither
Fruit Pure Pure A Nor B

Sweet Low High

Choose only
  
one option.
Step 8: Another Regression!

Multi-Nominal Logistic Regression (MNL)

Step 8: Direct Estimation of Market Share
• Why Use Conjoint?
• Conjoint analysis evaluates product/service
attributes in a way that no other method
can. Traditional survey approaches ask
respondents to estimate how much value
they place on each attribute. This is a very
difficult task for any person to complete,
much less someone who doesn't spend
every waking moment thinking about the
most important features of a product.
• Conjoint analysis, on the other hand,
attempts to break the task into a series of
choices or ratings. These choices or ratings,
when taken together, allow us to compute
the relative importance of each of the
attributes studied. Instead of "stated
importance," conjoint analysis uses "derived
importance" values for each attribute or
feature.
• Another of the advantages of conjoint
analysis is the ability to use the results to
develop market simulation models that can
be used into the future. Markets continue to
change as new competitors enter, new
products are introduced, price wars erupt
and marketers develop new advertising
programs. With traditional research
approaches, every time a major change
takes place in the market, a new survey
needs to be conducted to find out how
people feel about the changes and how it
will affect their purchases.
• With conjoint analysis, the changes in the
product or market can be incorporated into
the simulation model to generate predictions
of how buyers will respond to the changes.
In most markets, these models are accurate
for two or three years before you need to
conduct a mini version of the original study
to determine if any adjustments must be
made to the key parameters of the model.
• METHODOLOGY
• There are several basic steps taken in
applying the conjoint analysis to a problem.
To begin with, my first job was to find out
the most important attributes/features of the
hi-fi speakers from the point of view of the
target customers. Also, I had to consider
only those attributes on which the company
could easily work on and didn't add to the
cost of product.
• The next step was to determine the data
collection methodology. Since the product
is a high end product, random sampling
would have never given me the inputs in
regard of our objective. Taking this
limitation into consideration, I restricted
myself to the global theatre, showroom
name for PROFX, and hence followed
stratified sampling technique. I also had a
list of few audiophiles of the city as my
respondents.
• The third step was the selection of type of
conjoint analysis to be used: Preference
based or Choice based. In preference based
conjoint analysis the attributes are broken up
in smaller sets to reduce the burden on the
respondent. The limitation of preference
based sampling is that it rules out price
attribute and brand attribute. As we all
know, price attribute can not be left out for
any product category in India. Hence, I
preferred Choice-based conjoint analysis to
preference-based conjoint analysis.
• Also, choice based conjoint analysis asks
target customers to make simulated
purchase decisions rather than indications
of preference in terms of rating of the
speaker designs. The next step involved in
the exercise was the collection of data. This
wasn't a tough task as all that I could have
done was - wait for the customers at GT.
Though I took some pains in collecting
the data from Chennai and Mumbai.
Also, to get the responses from the
audiophiles was never an easy job.
• After the data collection, the utilities of
various products were found with the help
of software known as "Minitab". The final
task of the whole study was to develop a
simulation model and to predict the impact
of changes in the existing products and
introduction of new products in the market
• LIMITATIONS
• The limitation of the study could be
summarized as follows: (explanation)
• Error in selecting respondents: I approached
persons who looked amiable and friendly, so
there is a possibility that there may be an
element of biasness in it.
• Constraint of time: All the research-based
studies are under the constraint of time. Also
the duration for the study was only few weeks
which made me to restrict myself to 171
respondents.
• Missing of the attributes: In the study I
considered only seven attributes. There are
many more attributes that may influence a
customer decision. I did not incorporate the
entire possible attribute due to the nature of
our study, which did not permit me to have
all the attributes.
• Relative rating by the respondents: Some of
the respondents rated the cards after
considering all the cards, which was not
desirable.
• Non response errors: Some of the
customers did not fill up all the required
questions that hampered my analysis to an
extent and thus there could be small error
due to this.
• Lengthy response time: To fill up all the
cards a customer had to spend a fair amount
of time and some of the respondents were
found to have been a little laid back towards
the end.
• SAMPLING FRAME
• Definition of the population: The company's
target segment is the cities AI and A
• I have considered city, age, profession and
household income as the sampling units
beside all the attributes of a speaker.
• CITY: I had three cities for the survey:
• Bangalore, Chennai, Mumbai
• PROFESSION: This was done to find
which profession background does my
respondents comes from. I had four
categories:
• Professional, Businessman, Student
• Others(pl. specify)
• AGE: This was also done to find which age
segment(s) does my respondents comes
from. I had five categories:
• < 18 years, 18-25 years, 25-35 years, 35-50
years ,> 50 years
• ANNUAL HOUSEHOLD INCOME: This
was also done to find which income segment
does my respondents comes from. I had four
categories
• < 3 lacs
• 3-6 lacs
• 6-10 lacs
• > 10 lacs
 Attributes
• Looks : Simple, Trendy
• Size : Big, Small
• Brand : Indian, Foreign
• Sound : Bass, Vocal
• Price : 16k, 10k
• Warranty : 1 yr, 3 yrs
• Free-home-demo : Yes , No
Simple Big Indian Bass 16k 1 yr Yes
Simple Big Indian Vocal 10k 3 yrs No
Simple Small Foreign Bass 16k 3 yrs No
Simple Small Foreign Vocal 10k 1 yr Yes
Trendy Big Foreign Bass 10k 1 yr No
Trendy Big Foreign Vocal 16k 3yrs Yes
Trendy Small Indian Bass 10k 3yrs Yes
Trendy Small Indian Vocal 16k 1yr No
• FINDINGS:
• People give maximum importance to sound
followed by price, brand, size, warranty,
looks and free-home-demo.
• People prefer more of bass than vocals.
• They give almost similar importance to
brand and price. It is a very sensitive issue.
It means that the company should have a
competitive pricing according to the brand.
• People of Bangalore prefer big speakers to
small speakers and size does play an
important role in the decision process that we
can say from the weightage they put on
size.
• People are not much bothered about the
services given by the company such as
extended warranty or free-home-demo