Digitalization, on the one hand, significantly increases the volume of opportunities and alternative resources available for learning. However, on the other hand, it leads to a complication of the environment where learning takes place. This, in turn, has a negative impact on the structure of the information, the effectiveness of the learning and the return on human capital investment. The purpose of this article is to test hypotheses about the impact of quantitative diversity in social, spatial and digital contexts, network activity and the ability to self-organize on the effectiveness of employees' self-learning in the digital environment. The methodology of contextualism, methods of factorial and regression analysis are used. The authors use data from a survey of 354 young professionals from the Sverdlovsk region under the age of 35. The novelty of the study lies in the development of an approach to the measurement of explicit variables for the evaluation of theoretical constructs such as the effectiveness of self-learning, learning context, self-organization and network activity. To measure contextual diversity, a method for determining the number and types of contexts (spatial, digital and social) in which employees self-learn is proposed. The results of the study showed that the number of spatial and numerical contexts does not affect the effectiveness of learning, whereas the social context, basic digital competences and the ability to self-organize are crucial for learning in a digital environment. The results of the study can be used by the creators and managers of educational programs to study the influence of a set of contextual variables on the effectiveness of training and the formation of human capital.