Introduction
Conjoint analysis is a statistical analysis in marketing research to know
how respondents value different attributes (characteristics or features) of
products or services.
The goal is to determine what combination of attributes is the most
preferred to respondent choice decisions.
Introduction
Design conjoint
survey
Distribute
survey
Analyze
conjoint data
Number Analytics generates a web link to the survey webpage so that
respondents are asked to choose one of several options available. Once
collected, you can run the analysis and get the individual partial worth at
individual level.
This tutorial shows how to design, distribute and analyze the choice
based conjoint analysis
Conjoint Survey Outline
Starting description > Demographics > Conjoint Questions > Analysis
Analysis
Start a Choice Based Conjoint Survey
Choice Based Conjoint generates online survey line and collect respondent choice data
First you need to set attributes and levels for the product
Set conjoint constraints where unrealistic or infeasibleattribute combinations should not be shown
For example, Starbucks does not sell a coffee below $2,then we need to eliminate the option for $1.5
Limiting condition *Available only for advanced and premium plans
The number of CBC q
uestions set how man
y times the choice tabl
es are shown per resp
ondent
The questionnaire display options
• Standard table without none option
• Promotion type display (without attribute n
ames)
• Standard table with none option
The max number of respondents
Set the number of co
njoint choice alternati
ves in the choice tabl
e
*Available only for advanced and premium plans
Design setting
Post the survey link to social media, website, facebook,or send email to respondents.
Publish your conjoint survey
Data analysis
Collect data with at least 3 respondents (“Run” button turns to be active)
Click “Run” button to see the analysis results
Test the survey in a test mode
Publish the surveyBefore you post or send the link to the respondents, please publish the survey. Once you publish the survey, you can't change the survey. You can only change the survey until the survey is published. So test enough before you publish and post the survey to public. If you changed the survey, the data will be initialized and will be reset.
Once you finalized the survey design, you need to “publish” button then you can’t change it.
“Publish with responses” if you want to keep the respondent data you already collected
Publish without previous responses if you want to collect the respondent data all over
Conjoint results
Attribute importance metrics
as a pie chart
If there is a big variation across attribute levels given the attribute, the importance measure is higher
Attribute mean utilities as column charts at aggregate level
Interpret demographic effects on conjoint
25-34 years old customers like Mc Cafe more than Starbucks
Male customers are more price sensitive
Male customers likes Cappuccino than female customers
Market simulatorDefine competitive market conditions
Set market size and profitability
Predict market share, revenue and profits
Choose the price variable