The paper describes the main approaches used in the development of predictive analytics algorithms, both of a general nature and in the field of well drilling. When describing the algorithms used, the basic assumptions of each and the ensuing limitations of the developed model are explained. The necessity of building hybrid models, involving the full amount of information collected, including calculations carried out during the construction of the well, both from the engineering and geological sides, is shown. The paper discusses the culture of working with data in the oil and gas field in general and in particular well drilling. The current level does not allow companies to create scalable solutions of predictive analysis. As a result of the work done, the basic rules were proposed, the first step was taken towards standardizing the process of collecting and analyzing information flow arising during the construction of wells.