A competitive intelligence acquisition framework for mining user perception from user generated content

Abstract

Competitive intelligence is an important basis for companies in the process of developing competitive strategies. However, competitive intelligence acquisition methods such as surveys and expert ratings are not able to achieve rapid responses to user perceptions. In addition, the huge amount of unstructured user-generated content makes it more difficult to analyze user perceptions. In this paper, we propose a competitive intelligence mining framework for acquiring user perceptions from user-generated content. The framework covers multiple aspects of competitive intelligence acquisition, mining, analysis, and decision-making. A unified text processing model (WS-TCM) can automatically filter the content irrelevant to competitive intelligence and quickly extract fine-grained user perceptions of competitive attributes from the huge amount of user-generated content. In addition, the quantile-based intelligence mapping method (QB-IM) determines the competitive landscape of an enterprise based on fine-grained user perceptions and provides help for managers’ strategic decisions. In the case study, the method proposed in this paper has significantly improved compared with the baseline model in the stages of competitive intelligence identification, competitive attribute identification, and user perception identification. Especially, the accuracy improvement in identifying user dissatisfaction perception is more obvious. Meanwhile, our model shows strong robustness. In the competitive analysis stage, the results of the framework are consistent with those obtained by market analysts through surveys. This study provides a new approach to competitive intelligence research and fine-grained user perception mining, which improves the efficiency of competitive intelligence acquisition for companies while also improving the reliability, accuracy, and usefulness of competitive intelligence.