In this project, the overall goal is to predict the churn of bank customers. From a business perspective, this is very relevant for the effort to retain customers with the ultimate end goal of increasing profitability.
Customer churn is defined as the percentage of customers that stopped using a company’s product or service offering in a defined time frame. One might consider that customer churn is not so important as long as more new customers are acquired than lost to the company. This is fogetting entirely the cost of acquiring new customers. Bringing in new customers is a lot less profitable than retaining customers. In financial services, for example, a 5% increase in customer retention produces more than a 25% increase in profit. The reason for that is because returning customers spend on average more than already existing customers. In online services, a loyal customer spends on average 2/3 more than a new one. At the same time there is a cost associated with acquiring new customers, which decreases when less new customers have to be acquired. Keeping existing customers thus allows for a reallocation of funds away from the need of growing by acquiring new customers.
Customer churn can be reduced by pooling resources into keeping the most profitable customers, instead of focusing on keeping overall customer numbers (even unprofitable ones). Another option would be to find out why and when customers are leaving, thus targeting in a customer lifetime this specific point and put effort into avoiding churn. In either case, the customer churn has to be thoroughly analyzed, which is what this small example project is designed to deliver.
Outline
This churn prediction project follows this outline:
1. Dataset description
2. Descriptive visualizations using Tableau
3. Data extraction, transforming, and loading (ETL)
4. Analysis of the dataset
Stages 1 and 2 are covered in Bank customer churn prediction: Identifying the question.
Stage 3 is presented in Bank Customer Churn Prediction: ETL.
Stage 4 is detailed in Bank Customer Churn Prediction: MVP.