With the rise of the competition in the movie production market, because of new players such as Netflix, Hulu, HBO Max, and Amazon Prime, whose primary goal is producing a large amount of exclusive content in order to gain a competitive advantage, it is extremely important to minimize the number of unsuccessful titles. This paper focuses on new approaches to predict film success, based on the movie industry community structure, and highlights the role of the casting director in movie success. Based on publicly available data we create an "actor"-"casting director"-"talent agent"-"director" communication graph and show that usage of additional knowledge leads to better movie rating prediction.