The goal of this step is to identify useful and distinct attributes and levels that can be combined into profiles for testing consumer responses and preferences. Conjoint analysis study respondents are presented with combinations of attributes and levels that they rank or rate. This data becomes the basis of the conjoint analysis.
When the number of attributes and levels is small, it is possible to create profiles that present the entire set of combinations. If all combinations are used, consumers will experience some repetition as they review the profiles. The inevitable repetition should be balanced against the utility of making such fine distinctions, and the likelihood that all of the profiles generated have the potential to actually be used in manufacturing or service provision. If the number of relevant attributes and levels is large, a technique called fractional factorial design will need to be employed. Fractional factorial design is used in situations where not all the combinations of attributes and levels can functionally be presented to consumers and when the objective is to provide a reasonable menu of profiles that will permit optimal use of the information collected from respondents.