Optimal customer selection for cross-selling of financial services products

• Published in: Expert systems with applications Vol. 40; no. 5; pp. 1748 - 1757
• Main Authors: Kaishev, Vladimir K., Nielsen, Jens Perch, Thuring, Fredrik
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier Ltd 01-04-2013; Elsevier
• Subjects: Applied sciences; Asymptotic properties; Call center; Contracts; Cross-sales; Exact sciences and technology; Expert systems; Finance; Financial services; Firm modelling; Insurance; Insurance industry; Marketing; Maximization; Mean–variance; Methodology; Multivariate Bühlmann–Straub credibility; Operational research and scientific management; Operational research. Management science; Optimization; Portfolio theory; Profit optimization; Queuing theory. Traffic theory; Call center; Mean–variance; Multivariate Bühlmann–Straub credibility; Cross-sales; Insurance industry; Profit optimization; Financial services; Marketing; Stochastic model; Information credibility; Very large databases; Probability distribution; Optimization; Variance; Swedish; Mortgage; Profit; Mean-variance; Derived product; Probabilistic approach; Cross-selling; Multivariate Bühlmann-Straub credibility; Loan; Asymptotic normality; Experimental design; Insurance; Insurance company; Contract; Portfolio management; Financial service; Service economy
• ISSN: 0957-4174; 1873-6793
• DOI: 10.1016/j.eswa.2012.09.026

► Cross-sale prospects are usually selected based on their propensity to buy. ► This neglects any stochasticity in profit, given sale, among the prospects. ► A model is presented combining sales probability and customer specific profit. ► Customer specific expected profits are then available prior to the sales approach. ► The method doubles the profit in comparison to a standard model for propensity to buy.
A new methodology, for optimal customer selection in cross-selling of financial services products, such as mortgage loans and non life insurance contracts, is presented. The optimal cross-sales selection of prospects is such that the expected profit is maximized, while at the same time the risk of suffering future losses is minimized. Expected profit maximization and mean–variance optimization are considered as alternative optimality criteria. In order to solve these optimality problems a stochastic model of the profit, expected to emerge from a single cross-sales prospect and from a selection of prospects, is developed. The related probability distributions of the profit are derived, both for small and large portfolio sizes and in the latter case, asymptotic normality is established. The proposed, profit optimization methodology is thoroughly tested, based on a real data set from a large Swedish insurance company and is shown to achieve considerable profit gains, compared to traditional cross-selling methods, which use only the estimated sales probabilities.