Using Time Features for Increasing Email Conversion

The development of personalized CRM e-mails is currently one of the fastest growing areas for improving the efficiency of omnichannel communications. This project is aimed at creating a communication tool for e-mail newsletters using data provided by the Instamart company, the purpose of which is to identify the optimal sending time for clients and calculate the probability that the client will open an e-mail newsletter. As methods for obtaining data on customer interaction with e-mail newsletters, the method of exploratory data analysis was chosen, as well as the construction of a classification model based on the data. The result of the research and project is to study the effectiveness of classification models (logistic regression algorithm), including using the gradient boosting method and the XGBoost and CatBoost frameworks. According to the results, a better option is to use the CatBoost algorithm to predict the opening of e-mails on certain days and times. In conclusion, the results of the work performed are summed up, as well as recommendations for improving the performance of the model and further implementation of this approach.