Feature drift-based framework for novel idea recommendation in new product development

Abstract

In these competitive times, designers need to constantly update their tangible products with new features and functionalities to retain market share. This is especially important for designers of emerging products, as customer expectations of them change more frequently. New Product Development (NPD) is an established process which assists product designers in achieving this aim. In the first step of this process, product designers find novel features for upgrading their products (idea generation) and select some ideas (idea selection) for developing a new product concept in the next steps of NPD process. This step plays a vital role in introducing newness into a product and consequently its success in the market. A recent trend in existing literature for generating novel ideas is transfer learning where ideas for improvements are inspired from products in other domains. However, they are heavy/process-centric and time-intensive for the designers of emerging products in finding novel ideas from products of other domains. Moreover, while enormous research has been done to accomplish the idea selection task, there are still no agreed evaluation metrics for the product designers to prioritize their ideas. To address this problem in this thesis a framework, namely FEATURE, is developed which assists product designers to accomplish idea generation and idea selection to enhance emerging products. FEATURE adopts the science and engineering research approach, and it consists of three modules, namely Novel Feature Finder (NFF), Feature Sentiment Analyser (FSA) and Targeted Feature Recommender (TFR). NFF generates ideas by searching the products of cross-domains for their novel features that can be integrated into the product to be improved. To diminish the risk of product failure, FSA then ascertains the popularity of the generated ideas by using the customers’ reviews in social media. Finally, TFR prioritizes ideas based on different decision-criteria. From an industry-applicability perspective, FEATURE provides the designers of emerging products with a systematic light-weight approach to identify and assess realistic ideas in a short time. The functionality and viability of the different tasks of FEATURE are validated by experiments and systematized by a prototype to highlight the effectiveness of the overall proposed solution.