The worst is not to fail, but to fail to learn from failure: A multi-method empirical validation of learning from innovation failure

An inconvenient truth about innovation projects is that they frequently fail. Innovation and failure are so entwined that the probability of failure increases with the intensity of innovation. However, while some innovation projects fail, some organisations fail to learn from their failures despite the importance of such failures in avoiding failure in the future. Drawing on a scarce stream of work, this study contributes to the literature on Learning From Innovation Failure (LFIF) by drawing on two complementary theories: Social Learning Theory (SLT) and Social Exchange Theory (SET). Moreover, it contributes to the empirical validation of the relationship between LFIF and its determinants. Based on a sample of 436 manufacturing SMEs in Canada and a triangulation of analysis methods (structural equation modelling, SEM, and configurational: fuzzy set qualitative comparative analysis, fsQCA), we show that LFIF is explained by organisational (problem-solving and blaming approaches and psychological safety), interactional (trust among employees), and individual factors (personal mastery). Moreover, while the SEM results confirm synergies between LFIF drivers, the fsQCA results shed light on three pathways of conditions for LFIF. We derive some implications for managers and suggest directions for future research on the links between psychological safety, trust, and problem-solving approaches.