Beyond Keywords: A Novel Approach to Operationalize Strategic Attention Through Large Language Model

The Attention-Based View (ABV) of the firm has emerged as a critical theoretical framework for understanding organizational decision-making and adaptation in complex environments. Despite its importance, operationalizing and measuring attention constructs such as intensity and breadth remain significant challenges due to the limitations of traditional methods, including dictionary-based and manual coding approaches. These methods often fail to capture the contextual and dynamic nature of organizational attention, hindering theoretical and empirical progress. This paper introduces a novel framework leveraging large language models to overcome these limitations. By processing text data such as earnings call transcripts and annual reports, the proposed method generates context-sensitive embeddings that enable precise measurement of attention intensity and breadth. Unlike static dictionary-based approaches, our method adapts to linguistic nuances, evolving jargon, and diverse contexts, providing a scalable and flexible solution for operationalizing attention constructs. We validate the proposed method using real-world datasets from U.S. and Chinese firms, demonstrating its robustness and transferability across languages and organizational settings. Empirical results highlight the method’s superior predictive power for organizational outcomes compared to traditional approaches. By advancing the measurement of attention, this study opens new avenues for understanding how attention dynamics influence organizational behavior and performance.