Analyzing Customer Needs of Product Ecosystems Using Online Product Reviews

Abstract

It is necessary to analyze customer needs of a product ecosystem in order to increase customer satisfaction and user experience, which will, in turn, enhance its business strategy and profits. However, it is often time-consuming and challenging to identify and analyze customer needs of product ecosystems using traditional methods due to numerous products and services as well as their interdependence within the product ecosystem. In this paper, we analyzed customer needs of a product ecosystem by capitalizing on online product reviews of multiple products and services of the Amazon product ecosystem with machine learning techniques. First, we filtered the noise involved in the reviews using a fastText method to categorize the reviews into informative and uninformative regarding customer needs. Second, we extracted various customer needs related topics using a latent Dirichlet allocation technique. Third, we conducted sentiment analysis using a valence aware dictionary and sentiment reasoner method, which not only predicted the sentiment of the reviews, but also its intensity. Based on the first three steps, we classified customer needs using an analytical Kano model dynamically. The case study of Amazon product ecosystem showed the potential of the proposed method.