Designing adaptive feedback mechanisms with text mining capabilities: An illustration on eBay

Abstract

This research looks at current feedback mechanisms design at an electronic marketplace, notices the shortcomings of underutilized feedback comments, and proposes an alternative design that uses text mining to reveal latent service quality/customer satisfaction dimensions, otherwise potentially unnoticed. We observed the rigidity of many feedback mechanisms that confine users to leave feedback on a narrow palette of options, and we used adaptability theory principles to propose the design of a new feedback mechanism. The proposed feedback mechanism design draws on three studies: (1) the first study shows that feedback comments contain unobserved dynamically latent service quality/customer satisfaction dimensions, (2) the second study shows that some of the dynamically latent service quality/customer satisfaction dimensions are more important than the rigid a priori service quality/customer satisfaction dimensions existent at current electronic marketplaces, and (3) the third study shows that, when revealed, extracted service quality/customer satisfaction dimensions have the potential to change behavioral intentions formed on rigid a priori established service quality/customer satisfaction dimensions. We conclude our research by providing steps on how to implement an adaptive feedback mechanism using text mining.