Research and Discovery Tools for Experimentation - 17 Apr 2024 - v 2.3 (1).pdf
Conjoint Analysis - Part 1/3
1. Class Outline
• Measuring Consumer Preference
– Problem with intentions-to-buy and self-reported methods
– Conjoint Analysis
2. Preferences
• Intention to buy
– too hard a question
– overconfident about buying probability
stated
preferences
• Self-reported importance weights
– lack of discrimination, everything is important
• Trade-offs
– benefits and costs
– what is desired most
• Actual purchase data
– counts
– sales force reports
revealed
preferences
Consumer
preferences
4. How to Learn What Customers Want?
• Ask Direct Questions about preference :
– What brand do you prefer?
– How much RAM do you prefer?
– How much hard disk space do you prefer?
– What type of monitor do you prefer?
– What price do you prefer?
• What is the problem with this?
5. Problems with Direct Questioning
• Answers are often trivial and unenlightening
– “I prefer more processor speed to less”
– “I prefer more RAM to less”
– “I prefer more hard disk space to less”
– “I prefer a lower price to a higher price”
6. More Discrimination from
Trade-off Questions
Traditional Approach
Conjoint Study
Average Importance
Ratings
Average Importance
Ratings
Brand
Brand
Scan Time
Scan Time
Resolution
Resolution
Price
Price
0
2
4
6
8
10
0
10 20 30 40 50
8. The conjoint world: terms and concepts
•
•
•
•
•
•
Attributes
Attribute levels
Profiles
Utility models
Part-worths
Types of conjoint:
ranking, rating, choice
conjoint world
9. iPhone 5
Black & Slate
White & Silver
16GB
32GB
64GB
$199
$299
$399
• Attribute:_________________________
• Attribute level:_____________________
10. iPhone 5
Black & Slate
White & Silver
16GB
32GB
64GB
$199
$299
$399
• Attribute: Brand, Memory, Price, Color
• Attribute level: Apple/Samsung,
16GB/32GB/64GB, $199/$299,$399, Black/White
11. Conjoint Analysis
• Conjoint Analysis: A technique that enables a
researcher to estimate consumers’ valuations of
different attributes
– Allows us to understand how consumers make
trade-offs among attributes of products and
services
– How much are consumers willing to pay for
different attributes?
13. Conjoint Reveals What People Really Want
What MBA's
Said...
Rank Based on
Conjoint
Salary
6th
1st
Region of US
2nd
2nd
Job Location
5th
3rd
People/Culture
1st
4th
Functional Area
3rd
5th
Firm Growth
7th
6th
Business Travel
8th
7th
Opp. to Advance
4th
8th
Job Attribute
Consumer
preferences
14. Conjoint Analysis - Intuition
• Force consumers to rank different bundles of
attributes
1.
C
2.
E
C
3.
A
D
4.
F
5.
B
6.
D
A
B
E
F
15. Conjoint Analysis - Intuition
• Using the information contained in consumers’
rankings (or ratings), we can back out what their
valuations of different attributes must be:
1.
C
2.
E
3.
A
4.
F
5.
B
6.
D
Suppose both products C and E contains a certain
attribute that is not found in A,F,B,D. Then the
consumer probably has a high valuation for that
attribute
Suppose both products B and D contains a certain
attribute that is not found in C,E,A,F. Then the
consumer probably has a low valuation for that
attribute
16. Conjoint Analysis: Procedure
1. “Design” a set of product concepts (profiles) - this
will usually come in with collaboration with the
engineering team
2. Ask respondents to rate/rank those product
concepts
3. Based on the respondents’ evaluations of the
product concepts, figure out how much utility each
attribute provides
32. Group Exercise: WTP
• $50 = ( 4 ) units
• Apple > Samsung
( 3 ) units = $( 37.5 )
• 20GB > 10GB
( 2 ) units = $( 25 )
• Most important attribute: ( Price )
• Least important attribute: ( Memory )
33. Group Exercise: Attribute Importance Weights
• $50 = (
4 ) units
• Apple > Samsung
( 3 ) units
• 20GB > 10GB
( 2 ) units
• Q1) Partworth range: the difference between
largest and smallest partworths for each attribute
• Q2) Importance weights:
= partworth range / (sum of partworth range)
34. Group Exercise: Attribute Importance Weights
Q1) Partworth range
• $50 difference: 4 units
• Apple > Samsung : 3 units
• 20GB > 10GB: 2 units
Sum of partworth range = 4 + 3 + 2 = 9
Q2) Importance weights:
• Price = 4/9 = 44.4%
• Brand = 3/9 = 33.3%
• Memory = 2/9 = 22.2%