Innovation Learning-Outcome Variability: A convergence-divergence model for population-level learning

Abstract

This study challenges the intuitive assumption in much of the organizational learning literature that greater implementation experience inherently promotes convergence in how an innovation is implemented by a population of organizations. We follow up on early conceptual arguments in the population-learning literature, speculating that such convergence may not always occur. To substantiate these claims, we engaged in an abductive investigation to assess the plausibility of this outcome using empirical data from the United States professional baseball industry (1923–1940) and applying visual descriptive, anecdotal qualitative, and formal quantitative analyses. Results revealed increased variation in a key innovation feature across organizations. However, the innovation’s performance impact across organizations became more similar over time, creating an unexpected pattern of innovation-feature divergence and performance convergence. Additional phenomenological mapping generated a novel conceptual framework that explicates four alternative scenarios of how industry-level experience can affect variability in innovation features and performance impact over time. This framework and the reported empirical support for one of the four divergence-convergence patterns advance population-learning theory with important implications for industry-level adaptation.