For identifying and selecting the most profitable customers in terms of the shareholder value, the Customer Lifetime Val-ue (CLV) gained broad attention in marketing literature. However, in this paper, the authors argue that the CLV does not take into account the risk associated with customer relationships and consequently does not conform to the principle of shareholder value. Therefore, a quantitative model based on financial portfolio selection theory is presented that considers the expected CLV of customer segments as well as their risk. The latter includes the correlation among the segments. It is shown how imperfect correlation among segments may be employed to maximize the value of the customer portfolio. Since portfolio selection theory does not allow for the consideration of fixed costs, it is extended by a heuristic method consisting of two algorithms, referred to as “subtract”- and “add”-approaches.