,allowExpansion)
We are redefining churn processes by utilising advanced technologies to detect churn early and maintain customer relationships.
Long-term customer retention at banks is becoming an increasing challenge. Today, customers have access to a wide variety of financial services and products that they can obtain with just a few clicks. Additionally, comparison portals simplify the process of comparing account or credit terms. In this highly competitive environment, it is crucial to identify customers who are likely to terminate their customer relationships. This is referred to as the "churn" process. Specifically, the goal is to identify customers who can be persuaded by targeted offers or outreach not to end their relationship. After all, it is generally far more profitable to sell additional services or products to existing customers than to acquire new ones.
)
The main problem in churn detection lies in identifying reliable indicators of an impending dissolution of the customer relationship. Traditional methods often rely on simple heuristics and statistics, as well as manual data analyses, which frequently come too late or are too imprecise. Moreover, important predictive signals in the multitude of information about purchasing behaviour, transactions, or customer demographics can be lost with these methods.
Another problem lies in data availability: to reliably detect churn, complete historical data sets from customers who have churned are required. However, storing and using customers' data for business purposes after they have terminated their banking relationships is legally and ethically problematic.
At Fincon Reply, we circumvent the issue of data availability by using appropriate churn definitions. We combine our technical and professional expertise. For example, it is possible to consider existing customers who have significantly reduced their transaction volume or withdrawn parts of their investment assets as "churned" — this is referred to as "soft churn”. Once the customer data has been enriched with the appropriate churn designation, we deploy advanced machine learning methods (ML) such as eXtreme Gradient Boosting (XGB) or artificial neural networks. In this way, we can discern complex data patterns and possible non-linear relationships .
between the predictive signals and the ultimate churn decision In various use cases, we have been able to improve churn detection by more than 10 percent compared to traditional methods in the past.
Simultaneously, the implementation of an effective feedback loop is crucial, where the predictions of the ML model are used to initiate targeted customer retention measures and measure their success. This feedback helps to continuously refine the model and adapt it to changing customer behaviours.
Early detection of customers who are at high risk of ending their customer relationship offers several advantages:
Targeted use of retention strategies
... such as personal counselling sessions or suitable product recommendations, which can increase customer satisfaction and loyalty.
Higher profitability
... as acquiring new customers is often more expensive than maintaining existing customer relationships.
Valuable insights
... into customers' needs and preferences that can be used for product development and improving customer service.
)
Fincon Reply is a business and IT consultancy specializing in the financial services industry. Fincon Reply proactively advises banks, the Sparkassen Finance Group, insurance companies and near-financial companies as well as their suppliers on their digital transformation. The company provides on-site support with specialised teams of consultants and developers and delivers turnkey solutions.