Managing Information Overload: The Case of Online Product Review Categorization

Abstract

Online product review sites contain an impressive amount of positive and negative consumer feedback. However, many marketing or community managers have difficulties handling these huge amounts of information, and thus these qualitative data often are wasted nowadays. The authors seek a practical solution to automatically determine consumers’ satisfaction levels according to the linguistic dimensions that appear in these product reviews. Based on a sample of 1, 014 real-life online consumer reviews, the authors propose an automated classification model that achieves an unprecedented time/performance ratio for consumer feedback classification. The authors test its performance against the categorization performance of manual encoding of the consumer narratives by 507 experienced marketing managers. Ultimately, this research contributes to marketing science by proposing an alternative way of measuring consumer satisfaction.