Overview of Choice-Based Conjoint

Choice-based conjoint is a survey technique for quantifying the relationship between a product or service’s design – expressed in terms of attributes – drives preferences (e.g., market share, sales).

CBC is shorthand choice-based conjoint (see also Variants of Conjoint Analysis).

Agree on objectives

The start-point for a choice-based conjoint study is agreeing on the objectives for the study. At a high level, this involves working out the overarching goal of the study (e.g., is it about understanding consumer needs, forecasting sales, or something else?).

The main applications of CBC are:

• Testing the appeal of a new product
• Understanding product deletions
• Portfolio optimization
• Assessing the impact of changes in product design
• Pricing optimization
• Understanding psychology
• Purchase hierarchies
• Preferences for attribute levels
• Product optimization
• Computing brand equity
• Calculating Willingness-To-Pay (WTP)
• Market segmentation

This is discussed in more detail in Applications of Choice-Based Conjoint.

For example, the study of the chocolate market sought to understand:

• The proportion of people that prefer sugar-free chocolate.
• Preference for sugar-free versus 50% reduced sugar.
• The extent to which people would pay a price premium for sugar-free and reduced-sugar chocolate.
• Other features to include in a bar of sugar-free chocolate. That is:
  • Branding
  • Price for a two-ounce (57g) bar of chocolate 
  • Cocoa
  • Origin
  • Nuts
  • Ethical chocolate.

Create the attributes and levels

A set of attributes and levels created for the chocolate study are shown in the table below.

For more detail, see:

• Examples of Attributes and Levels
• Identifying Attributes and Attribute Levels
• How to Check Attributes and Attribute Levels

Create choice questions using an experimental design

Choice-based conjoint studies present people with one or more choice questions, also known as choice tasks. For example:

Choice-based conjoint studies can be conducted with as few as one such question, but it is more common to ask each person 8 to 20 such questions. The descriptions of the products vary from question to question, as shown below. Typically, they also vary from respondent to respondent.

The questions are created using an experimental design. See How to Create Experimental Designs for Choice-Based Conjoint.

Conclusions about the appeal of attribute levels can be derived from the choices

Information about preferences can be derived from the answers to such choice questions. Consider Question 3 above. If a person chooses the Godiva option on the left, we can deduce that they prefer Single-origin Venezeuealen Criollo Beans over Sugar-free, when purchasing Godiva dark chocolate for $1.49. 

By making additional assumptions it is possible to make more powerful conclusions. For example, we may assume that preference for Venezeuealen Criollo Beans over Sugar-free is true regardless of the brand, price, and type of chocolate.

In practice, the analysis of CBC data is performed using advanced statistical methods, such as hierarchical Bayes, which have been created with widely accepted assumptions built in.

See Statistical Analysis for Choice-Based Conjoint for more information.

Reporting

A key way of presenting the results of CBC data is via a simulator, as shown in the animation below, where users can simulate the effect of changing product descriptions.

Such simulators can be used to create demand curves for products or product features. 

For more information, see Reporting for Choice-Based Conjoint.