More than 99% of the companies in the world are micro, small, or medium size enterprises and account for ~70% of the jobs, on average, in OECD countries. However, due to a lack of productivity, among other factors, only a third of them survive beyond 42 months. This paper explores the potential associations between behavioral management patterns and business growth and productivity in micro and small enterprises in Latin America. We analyze survey data collected from Mexico and Colombia and observations from company visits and workshops conducted in Mexico with managers of micro and small firms. We observe that risk, delegation, and goal setting are all influential predictive features for business growth and productivity. We also find that the associations between business performance and these behavioral patterns are better captured through non-linear models when compared to linear models. For example, when evaluating the out-of-sample accuracy for revenue growth, the non-linear model performs ~27.29% better than the linear model. This suggests that behavioral patterns are not independent from each other, but rather interact and combine in ways that can create different formulas for successful behavioral management. In addition, our results suggest that behavioral patterns should not always be viewed in terms of extreme terms such as “high” or “low”, as suggested by linear models; but rather that there are optimal, potentially moderate, bounds for the levels of each behavior. For example, the non- linear model for employee growth shows that managers with moderate levels of risk-tolerance have an increased probability of high growth compared to those who exhibit risk levels outside of the optimal bounds. Last, we also observe evidence that suggests that the willingness to adopt new technologies and processes as a behavioral management pattern has little predictive contribution to business growth and productivity, and may instead, be an indicator of the manager’s inability to perform a particular task or job well.