Modeling and segmentation of the customer lifetime value in the banking industry

银行业客户终身价值的建模和细分

• 批准号: 492137-2015
• 负责人: Augustyniak, Maciej
• 金额: $ 1.79万
• 依托单位: Université de Montréal
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=587192
• 关键词: Modeling; segmentation; customer; lifetime; value

Due to the strong competition in the Canadian banking market, acquiring and retaining profitable customers is more challenging than ever. Because customer demographics, buying behavior and needs are evolving rapidly, banks must now develop dynamic statistical models that will help them identify the right products to offer to their clients. These challenges are pushing the National Bank of Canada (NBC) to understand its Customer Lifetime Value (CLV) at an individual customer level, across all their financial products (deposits, loans, investments, credit card, etc...). The CLV can be defined as the revenue that a client is expected to generate for the bank throughout his or her relationship with the bank. The broad problem is therefore to develop advanced statistical methods to precisely evaluate and project the CLV. The CLV will then help the NBC to define, understand, and predict in a novel way the relationship between a given client and the NBC. The model is therefore intended to drive marketing decisions and customer relationship management. This research project has two objectives. The first objective is the identification and segmentation of the variables that are relevant for measuring the CLV. Within this objective, we will investigate modern segmentation techniques that have yet to be considered in the CLV context, such as self-organizing maps and fuzzy clustering. The second objective is the development of an original predictive model that measures CLV by way of semi-Markov chains. This research project is intended to benefit the Canadian banking industry in the customer analytics field by improving banking services to better fit users.

由于加拿大银行市场竞争激烈，获取和留住有利可图的客户非常重要 比以往任何时候都更具挑战性。由于客户人口统计、购买行为和需求正在迅速变化， 银行现在必须开发动态统计模型，帮助他们确定向客户提供的正确产品。 他们的客户。这些挑战促使加拿大国家银行 (NBC) 了解其客户 个人客户级别的所有金融产品（存款、贷款、 投资、信用卡等）。 CLV 可以定义为客户预期产生的收入 银行在他或她与银行的关系中。因此，最广泛的问题是开发先进的 精确评估和预测 CLV 的统计方法。然后 CLV 将帮助 NBC 定义， 以新颖的方式理解并预测特定客户与 NBC 之间的关系。模型是 因此旨在推动营销决策和客户关系管理。 该研究项目有两个目标。第一个目标是识别和细分 与测量 CLV 相关的变量。在此目标范围内，我们将研究现代 尚未在 CLV 环境中考虑的分割技术，例如自组织映射和 模糊聚类。第二个目标是开发测量 CLV 的原始预测模型 通过半马尔可夫链。 该研究项目旨在通过以下方式使加拿大银行业在客户分析领域受益： 改善银行服务，更好地贴合用户。