With intense competition and increasing globalization in the retail banking environment, having an effective customer retention strategy is becoming increasingly important. The longer a bank can retain a customer, the greater the revenue and cost savings from that customer
However, this is easier said than done, especially when customers are spoiled for choice and offered nearly identical financial products by banks and financial services providers. How can your bank stand out and compete more effectively?
Data can be your key differentiator. It can help you to understand why customers churn, predict who is about to churn, and proactively offer relevant promos or programs in order to keep valuable customers. How is this done?
Churn Analysis
Churn analysis is highly dependent on how customer churn is defined. For example, in the credit card business customers can easily start using another credit card, so the churn indicator for the previous card company is declining transactions. By taking a look at attributes such as the Recency, Frequency and Monetary value of transactions and a customer’s Length of Association with your bank (or RFML attributes), we can calculate the weighted Lifetime Value (LTV) of a customer. A higher LTV would indicate loyalty, while a lower LTV would indicate a tendency to churn. You can then segment customers into clusters such as “loyal," “likely to churn,” and “churned."
Together with demographic information, these same RFML attributes can then be used to analyze new customers and predict their churn behavior. Want more info? Make sure you to contact us for all these trending topics and more.
Taking Appropriate Action
Using the transaction history of current customers, we can build a model to predict their future activity including whether they will churn or spend more, and on which merchants. Marketing departments can then make decisions on what kinds of campaigns or promotions to run, based on these predicted behaviors. What’s more, they can personalize offers so that promos are directed towards customers most likely to act on them.
The more information you have on a customer, the better the models can fit them to a behavioral segment and the more accurate the predictions will be. This information can help you predict the balance and hence the interest income we can expect to generate from these customers. This can then be used in calculating potential returns on planned marketing campaigns. To see a summarized version of the scenario in discussion along with a shortened version of the working code, please have a look at our notebook on churn prediction for credit card customers. To get a detailed view of the code, you may want to have a look at our GITHUB code repository.