Customer success: marketing preditivo como estratégia de sustentabilidade para empresa de receita recorrente

Abstract

The monographic study presents fundamentals of data science that enable predictive marketing strategies, considering as fundamentals the areas of consumer experience, management of relationship and customer success as a sustainability strategy for recurring companies. Based on indicators that seek to guarantee the sustainability of the business and the relationship, such as the forecast of purchases, the probability of evasion and the value of the customer's useful life, the study develops a methodology of mathematical and statistical models called “Buy 'til you die models” that model a database of historical transactions to predict the future of the company-customer relationship. The overall objective of this monograph is to develop predictive analytics for B2B businesses with technologies related to data science. Additionally, its specific objectives are, through the Python language, to enable estimates of purchases, probability of evasion (churn) and customer value (CLV) for the entire database extension (big data operation). For this, two methodologies were applied: RFM and BG/NBD linked to the Gamma-Gamma model. The methodologies made it possible to model statistical indicators of success for the entire customer base for the next four years. Thus, a strategic and relationship action plan was developed based on the forecasts obtained.