Case Study: Leading Household Electronic Appliance Company achieves accurate prediction of future customer purchases with Cloudera

Intel and Cloudera Help a Company Use Predictive Analytics to Foresee Future Purchasing Behaviors

A French retailer of household electronics and white goods wanted to predict customers’ future buying behavior—specifically which households were most likely to purchase within the next 12 months and from which product lines—using large, disparate customer datasets (purchase history, demographics, service requests, etc.). The company needed a reliable, scalable analytics approach to move from guesswork to data-driven planning.

Intel and Cloudera implemented a Cloudera Hadoop (CDH) cluster with R to ingest nine data sources, create 106 deterministic variables, and apply random-forest variable selection with a Bayesian/logistic regression scoring process to rank households and predict product-line purchases. The solution accurately hindcasted historical purchases (68% correct), identifies top high-probability households and likely product groups, and gives the retailer actionable insights for targeted marketing and better purchasing strategy.