Dynamic prediction of product competitive position: A multisource data-driven competitive analysis framework from a multi-competitor perspective

Abstract

Competitive analysis is critical in enhancing product quality and facilitating strategic adaptation, with user-generated content (UGC) offering new avenues to improve product competitiveness. Existing UGC-based SWOT analysis methods have made progress in minimizing subjective bias. However, they still lack in analysis scope, enhancement of dynamic forecasting, and improvement of factor prioritization. Given the rapid changes in the market environment, companies urgently require innovative approaches to secure a competitive advantage. From a multi-competitor perspective, this study introduces a multisource data-driven dynamic SWOT (D-SWOT) analysis framework, which broadens the analysis to include all potential market competitors, aiming to dynamically predict the focal product's position in a specific competitive environment and evaluate factor priorities. Initially, competitors are identified from multisource data through a competitor identification algorithm. Subsequently, attributes are identified, their importance is assessed using the LDA model, and their performance is evaluated through the bidirectional long-short-term memory model. This study aims to comprehend market dynamics and formulate effective strategies by employing a grey prediction model to forecast future attribute trends. Based on the core principles of SWOT analysis, rules for determining factors have been clarified, and a D-SWOT matrix has been constructed to predict changes in the product's market position. Finally, integrating attribute performance differences with their relative importance, a competitive priority index apt for competitive analysis has been developed to guide product enhancements and strategic realignments. Through a case study involving tablet products, the framework's substantial benefits in terms of competitive analysis have been validated.