Data Science specialists structure and analyze large amounts of data, apply machine learning to predict events and detect non-obvious patterns. Based on signs of customer behavior, it can estimate the likelihood that a customer is inclined to choose a different supplier. Using a number of parameters that we identified through a study of customer activity, our team calculated the likelihood of outflow within 1-3 months. We identified groups that are "at-risk zone" and helped the company find out the most relevant text messages, which will motivate customers to buy goods.
What Was Done
During the preparation stage, we analyze the following parameters: frequency and volume of purchases, number of site visits, date of the last purchase, viewed and selected products, positive or negative user experience. As a result, the solution enabled the business to increase the efficiency of communications with users, conversion and repeat purchase rates, and the customer lifetime value (LTV).
Technologies and tools: Python, Pandas, NumPy, Scikit-learn.