Transactions from thousands of customers and hundreds of stores were registered in databases without being used for marketing purposes up to this moment. The work consisted of:
The predictive model calculated the probability of a customer returning to the store in the upcoming months, and the segmentation model profiled the customers considering the basket of purchased products. The models guided how to approach each consumer. Pilot tests with customers from specific stores were carried out in order to validate the concept, followed by direct marketing actions.
The stores were directly involved and offered inventory items to customers with higher probability of returning to the store. Customers were selected for reopening and collection launch events, and for visits during special hours, acknowledging them as special clients.
The databases, implemented for recurrent use, were enhanced by the identified attributes: the likelihood of each customer to continue buying, the segment of interest, and the relationship strategies recommended for each customer. This produced a new way of thinking about and dealing with information inside the company, from which it was possible to taking input for planning and execution. The project brought not only positive financial results, but also created a new culture in the company, which became proficient in dealing with data, information and mathematical models in order to optimize the communication budget and customer satisfaction.