@inproceedings{1061, abstract = {{It is well-established that both average online ratings and the number of ratings positively impact product sales. Yet, the economic implications of the information contained in the online review texts is not that well understood. In this study, we contribute to the understanding of online review texts and its economic implications by conducting and validating an unsupervised machine learning algorithm, the latent dirichlet allocation, to identify online reviews that mention product failures. Furthermore, we show that the textual information on product failures are associated with lower product sales. Our results help online review system designers, e.g., amazon, to identify these reviews and to make them easily accessible to potential customers to support the customer’s purchasing decision. Academics can build on our results by applying our validated topic identification strategy and by linking reviews mentioning product failure to a range of different outcomes.}}, author = {{Gutt, Dominik}}, booktitle = {{Proceedings of the Multikonferenz Wirtschaftsinformatik 2018 (MKWI), Lüneburg, Germany}}, title = {{{Sorting Out the Lemons - Identifying Product Failures in Online Reviews and their Relationship with Sales}}}, year = {{2018}}, }