The Impact of Conversational Survey Interfaces on Consumers’ Written Self-Expression
(A2021-94741)

AUTHORS

Meike Zehnle, Doctoral Candidate/University of St. Gallen/Institute of Marketing; Christian Hildebrand, University of St. Gallen

ABSTRACT

With advancing methods of automated text analysis, text data is becoming increasingly valuable for understanding consumers and predicting their behavior. To collect such data, companies increasingly replace traditional survey interfaces with conversational survey interfaces. The present work provides evidence from a series of three tightly controlled experiments showing that conversational survey interfaces systematically decrease the breadth of consumers´ writing while its emotionality remains unchanged. Moreover, we show that text features provide equivalent validity for predicting company-relevant outcomes when collected with a conversational survey interface as when collected with a traditional survey interface. Overall, the results of this work suggest conversational survey interfaces as promising market research tool for reducing customer effort and thus company costs without sacrificing customer insight.