A design strategy for improving adaptive conjoint analysis

Adaptive conjoint analysis (ACA) is a market research methodology for measuring utility in business-to-business and customer studies. Based on partial profiles, ACA tailors an experiment’s design to each respondent depending on their previously stated preferences, ordered in a self-assessment questionnaire. The purpose of this paper is to describe advantages and disadvantages of using a partial-profile randomised experiment, the usual system, and to propose a new design strategy for arranging profiles in blocks that improve its performance.

The authors propose a comparison between their design with the commonly used designs, as random designs and the so-called “mirror image”, in their resolution capacity for the estimations of main factors and two-factor interactions with the lowest number of profiles.

Comparing the proposed design over the other two designs highlights certain aspects. The proposed design guarantees more estimation for each experiment than the others and allows the researcher to tailor the design to his or her goals. The authors’ procedure will help researchers to determine an experiment’s resolution capacity before carrying it out, as well as to estimate main factors and two-factor interactions alike.

The authors propose a new design strategy for arranging the profiles in blocks for improving the performance of ACA. This proposal is based on the use of a full-profile approach in which profiles are arranged in two-level factorial designs in blocks of two, and the levels of each factor are codified vectorially.