Touchy Issues in Adaptive Choice-Based Conjoint Analysis: An Abstract

Abstract

The advancement of smart mobile technology has changed the way market research works. Importantly, an increasing number of consumers use direct touch (e.g., touchscreens) instead of indirect touch (e.g., mouse/keyboard combo) interfaces. However, the touch interface type’s role in market research methods’ performance has been largely neglected.

Choice-based conjoint analysis (CBC) is one of the most widely applied market research methods to elicit consumer preferences (Schlereth and Skiera 2017). Recent research, however, finds deficiencies in CBC. Therefore, its results may have only limited managerial relevance (Schlereth and Skiera 2017). Consequently, research on machine learning has introduced an upgraded version, namely adaptive choice-based conjoint analysis (ACBC) (Johnson and Orme 2007; Wackershauser et al. 2017).

During an ACBC analysis, respondents undergo three mandatory stages: a Build-Your-Own (BYO) stage, a Screening stage, and a Choice Tournament stage. All three stages are interrelated and allow for learning efficiently about consumers’ preferences (Orme 2014, p. 132). The BYO stage is of pivotal importance as its results set the stage for the remainder of the ACBC study. Within the BYO stage, consumers configure their ideal product by selecting specific product features at varying feature prices.

In this research, we argue that the type of touch interface a consumer uses in ACBC influences the BYO stage’s outcomes. This is because research in other contexts, such as car configurators, suggests that consumers accept higher-priced options when configuring products using a direct vs. an indirect touch interface (Hildebrand and Levav 2017). We first meta-analyze 12 ACBC data sets provided by market research agencies and academics to address this concern. In line with our concerns, the majority of ACBC studies indicated that the configured BYO products of the direct (vs. indirect) touch interface users had a higher price.

We subsequently conducted two studies (one online ACBC and one lab experiment) to analyze the interface type’s effects. We highlight its consequences for the ACBC’s results in terms of the estimated utility functions, relative price importance, willingness to pay, and predicted market demand. The lab study additionally provides an explanation of the interface type’s impact. Especially consumers with a high autotelic need for touch (Peck and Childers 2003) experience higher study enjoyment when using a direct (vs. indirect) touch interface during an ACBC. We extract highly relevant managerial implications. Researchers should control for respondents’ interface type in order to adjust the interface type used in ACBC studies with the one future customers will use when purchasing the focal product and services.