Mapping knowledge recombination for innovation: cognitive proximity and business method patents in the digital era

Purpose This study aims to investigate how knowledge recombination contributes to innovation during digital transformation, focusing on business method (BM) patents. It introduces the concepts of dominant and spark technologies to capture different types of cognitive proximity in the innovation process. The research aims to reveal how these distinct knowledge types influence the creation of BM innovations across regions. By mapping knowledge interactions in patent networks, the study provides new insights into the mechanisms of technological emergence and highlights their strategic relevance for knowledge-intensive organizations and regions navigating digital change. Design/methodology/approach Using patent data from the European Patent Office (1981–2015), the study constructs co-occurrence-based knowledge spaces to map technological proximity and recombination. Two types of technologies are defined: dominant (widely combinable) and spark (highly BM-specific), measured via weighted centrality in network analysis. Regional-level indicators are developed, and fixed-effects negative binomial regression models are employed to evaluate their impact on BM patent output. This approach integrates cognitive proximity theory and evolutionary economic geography into empirical knowledge mapping, enabling analysis of regional innovation through a knowledge management lens. Findings The analysis shows that both dominant and spark technologies significantly contribute to the creation of BM patents, though their effects vary by regional innovation capacity. Dominant technologies, broadly used across patenting activity, support general recombination, while spark technologies are tightly linked to BM-specific innovations. Notably, spark technologies exert greater influence in metropolitan regions with stronger innovation ecosystems. These findings highlight the differentiated roles of knowledge types in innovation and demonstrate that knowledge recombination dynamics are context-dependent. The study also confirms the value of cognitive proximity metrics in predicting innovation outputs. Originality/value This study extends knowledge management theory by operationalizing and empirically testing cognitive proximity through patent-based knowledge spaces. It introduces actionable metrics – dominant and spark technologies – that offer new ways to evaluate knowledge recombination potential. Unlike traditional studies focused solely on knowledge stock or diffusion, this work emphasizes structural relationships between knowledge domains and their implications for business innovation. The findings offer valuable guidance for firms, regions, and policymakers seeking to foster innovation in the digital economy by aligning knowledge strategies with technological convergence patterns and local knowledge structures.