In this work, the authors present a robust framework to enrich new product design process by dynamically capturing customer preference trends. The framework autonomously captures customer preference trends from publicly available product review data which is abundantly available but grossly underutilized. The method overcomes a major challenge that has plagued the product design community -the lack of large scale, realistic customer data and its meaningful interpretation to guide new product design process. The challenge is from conventional, prevalent use of customer surveys or focus group interviews that are usually costly and time consuming while the size of available data is usually small scale. The framework is composed of three steps-retrieval of customer review texts, mining product feature texts, and predicting future trend of product preference.