An Approach to Discovering Product/Service Improvement Priorities: Using Dynamic Importance-Performance Analysis

Abstract

The cost budget and resources of a business are limited. In order to be competitive sustainably in the market, it is necessary for a businesses to discover the improvement priorities of their product/service features effectively and allocate their resources appropriately for higher customer satisfaction. Online customer review mining has been attracting increasing attention for businesses to discover priorities of product/service improvement from online customer reviews. Despite some prior related studies, their methods have several limitations, such as simply using the frequencies of mentioned product features in reviews as an indicator of importance; neglecting the market competition; and focusing only on the static importance and performance of the target product/service features. To address those limitations, this study proposes a novel approach to discovering a product/service’s improvement priorities through dynamic importance-performance analysis of online customer reviews. It first clusters similar features into a feature group and calculate the relative performance of the feature groups using sentiment analysis. Next, the importance of each feature group’s performance to overall customer satisfaction is measured by the factor categories based on the Kano’s model. The factor categories are determined by the significance values of each feature group in both positive and negative sentiment polarities derived from the constructed decision tree. Finally, feature improvement priorities of a target product/service will be discovered based on the dynamic performance trend and predicted importance using a dynamic importance-performance analysis. The evaluation results show that the dynamic importance-performance analysis approach proposed in this study is a much better approach for product/service improvement priorities discovering than the product opportunity mining approach proposed in the prior studies. This study makes new research contributions to automatic discovery of product/service improvement priorities from large-scale online customer reviews. The proposed approach can also be used for product/service performance monitoring and customer needs analysis to improve product/service design and marketing campaigns.