How financial services companies can use existing customer data to identify cross-selling opportunities
Financial services companies wishing to increase their sales may look to their existing customer base for cross-selling opportunities. This study demonstrates a method for using historical information to both identify potential customers for cross-selling and assess their ‘risk profile’.
Insurance policies are financial products that involve a long-term relationship between customer and company. An insurance company can use this relationship to sell more products to preferred customers in its portfolio. Data on the customers' past behaviour is stored in the company's database and this can be used to assess whether or not more products should be offered to a specific customer. In particular, data on past claiming history, for insurance products, or past information on defaulting, for banking products, can be useful for determining how the client is expected to behave with other financial products.
The study investigates identification of customers to whom additional products should be offered, by estimating a customer's specific risk profile with the use of behavioural data from other products. A standard multivariate credibility model was applied to a portfolio of customers owning several financial products from one company. The model allows us to take into consideration the possible (positive) correlation in customer behaviour between different financial products and estimate the customer specific risk profiles, for a specific product not owned by the customer, without having observed any customer specific information with respect to that particular product. Instead, data on customer behaviour, with respect to the other (owned) products, is the only necessity for estimating the risk profile.
In this empirical study we analyse our methodology on real data from a large Swedish insurance company.
This research can help to improve marketing to existing customers and to earn higher profits for the company. It can be downloaded via the link below.