Data-driven innovation development: an empirical analysis of the antecedents using PLS-SEM and fsQCA
研究基于TOE框架，通过264家澳大利亚企业的调查数据，用PLS-SEM和fsQCA两种方法分析了影响数据驱动创新发展的技术、组织和环境因素，发现数据质量、技术导向领导力等是关键条件，并针对不同规模企业给出了组合路径。
Abstract Data-driven innovation (DDI) is a primary source of competitive advantage for firms and is a contemporary research priority. However, what facilitates the development of DDI has largely been understudied in literature. Through a systematic literature review, this study finds technological, organizational, and environmental variables under the TOE framework, which would drive effective DDI development. We thus develop a research model, which is tested using survey data from 264 Australian firms engaged in DDI development. The data have been analysed using both symmetric (partial least squares based structural equation modelling (PLS-SEM)) and asymmetric (fuzzy-set qualitative comparative analysis (fsQCA)) methods. The mixed method enhances the confidence in our empirical analyses of the antecedent variables of DDI development. PLS-SEM has revealed that technological readiness (i.e., data quality and metadata quality), and organizational absorptive capacity and readiness (i.e., technology-oriented leadership and availability of IT skilled professionals) affect DDI development. Our fsQCA results complement and extend the findings of PSL-SEM analysis. It reveals that quality of data and metadata, technology-oriented leadership, and exploitation capacity individually are necessary—but are not sufficient—conditions for high DDI development. Further, it identifies three different solutions each for small, medium, and large firms by combining the TOE factors. Additionally, this study suggests that the TOE framework is more applicable to small firms, on DDI context. Findings of our study have been related with theoretical and practical implications.