Purpose The purpose of this article is to analyze data-driven insights as a technological factor capable of generating knowledge from the failures of the organization's previous innovation projects. Furthermore, this new knowledge plays a critical role in expanding the organization's knowledge resources, enabling it to better access and leverage the knowledge provided by its external partners. Design/methodology/approach The three-way interaction model was tested on a sample of 197 Colombian firms operating in medium- and high-digital-intensity sectors, which are characterized by the development of more sophisticated projects and the use of more advanced digital technologies. For this purpose, partial least squares structural equation modeling (PLS-SEM) was employed due to its explanatory and predictive capabilities. Findings The main findings reveal that innovation failure is a traumatic experience that triggers evasive and cautious behaviors within the organization and among its members. Moreover, this behavior, which is counterproductive to digital innovation, can only be reversed through prescriptive insights. This is because the nature of such insights enables the automation and decentralization of decision-making, thereby fostering greater agility and responsiveness to market needs. Research limitations/implications This article emphasizes the importance of broadening the organization's innovation approach to include external knowledge and technology, aiming to expedite the development of digital solutions and maximize their potential for market success. Moreover, this study challenges the dominant perspective that views technological and market knowledge gained from unsuccessful innovations as sufficient to prevent failures in subsequent innovation projects. Finally, it addresses the recent call in the literature for the design of innovation models integrating tools or mechanisms aimed at transforming the traumatic experiences of innovation failure into new knowledge assets to improve the performance of subsequent projects. Practical implications The main practical contribution of the article lies in recognizing the significant differences between descriptive and predictive insights versus prescriptive insights in decision-making, due to their distinct nature. Accordingly, the primary recommendation is to prioritize prescriptive insights, which requires the organization to ensure the availability of adequate technological infrastructure, data architecture and quality, as well as specialized technical expertise. Finally, the organizational learning strategy should be oriented toward the systematization of innovation failure in order to generate inputs for data analytics. Social implications This study has important social implications, as it encourages a cultural shift toward viewing innovation failure as a valuable source of collective learning. By highlighting the potential of prescriptive insights to transform failure into actionable knowledge, the article promotes more inclusive access to organizational learning, fosters interorganizational collaboration, and emphasizes the need for knowledge management practices. Moreover, the automation and decentralization of decision-making processes imply evolving skill requirements, underscoring the importance of digital literacy and workforce adaptability in an increasingly data-driven society. Originality/value The originality of this article lies in the application of descriptive, predictive and prescriptive insights to generate organizational knowledge from past innovation failures. This responds to the call in the literature to identify technological tools capable of fostering organizational learning, as well as to the practical need for organizations to recover value from resources invested in failed innovation projects. Furthermore, the study contributes to understanding the combined effect of failure-based learning, enabled through insights, on the organization's ability to leverage both internal and external knowledge for digital innovation.
Vélez-Jaramillo et al. (Mon,) studied this question.