In an era of rapidly changing and evolving markets, firms’ marketing agility has become a key driver of competitive advantage. Marketing agility, however, requires very timely and ongoing insights into markets. Over the past decade, marketers have thus turned to a rich, abundant and very timely data source, namely user-generated-content (UGC). Extracting insights from UGC, however, is challenging since such data are usually very big and unstructured. To overcome this challenge, researchers recently started to employ methods from the field of natural language processing (NLP) such as the popular word2vec model. However, it suffers from two major shortcomings: 1) it is unable to distinguish between multiple meanings of a word, and 2) it does not capture changes over time. We overcome the shortcomings of word2vec with a new model named Dory that is inspired by human memory-systems. As we show by simulation and in an empirical application, Dory’s additional qualities of human memory-systems enable it to detect patterns in consumers’ brand perspectives that remain undetected by extant models.
Professor Ringel is a managerial data scientist interested in creating insights into today’s large markets. His research interests include competitive analysis, data visualization, neural learning, unsupervised learning and the analysis of unstructured data. He currently studies the evolution of markets, the attention economy, and temporal information in natural language processing. His goal is to provide firms with tools and models that inform their strategic decision making. Professor Ringel worked for more than 10 years in management consulting.