Анализ поведения авторов онлайн-отзывов с позиции клиентской аналитики

In modern society companies are striving to get the attention of the customers and it is more vital for those of them, which users are content creators and the main asset of the business. Customer analytics has become a widely used tool for a successful customer engagement and retention in almost all industries. This research investigates to what extent customer analytics models are applicable for analysis of the authors at the user generated for authors of the reviews for the user generated content (UGC) platforms. This study evolves the results of last year's paper (Kolesnik, Milovidova & Molostova, 2020) which confirmed the applicability of customer analytics models for general UGC platforms such as TripAdvisor.com. Buy Till You Die (BTYD) and Recency - Frequency - Monetary approach (RFM) with an implication of Random Forest and Pareto/NBD modelling become a basis for the research on the data of 45986 reviews and 20 characteristics from 1939 registered users of the TopHotels.ru platform. Both models proved to be precise in the customers behaviour patterns identification and prediction, in particular churn, with regards to their performance, yet the lowest RMSE metric contributes to the Pareto/NBD being the most promising algorithm. Obtained results are useful for TopHotels and TopHotels-like UGC platforms allowing them to address different customer segments based on their previous activities, eliminating the problem of churn.