A framework for measuring and predicting customer lifetime value: a case study for online game industry

Customer lifetime value (CLV), which is a measure of customer profitability, is an important research problem of customer relationship management (CRM). In literature, there are numerous models that have been used to measure CLV in different industry domain problems. Particularly, measuring CLV and constructing a CLV prediction model are usually considered for CRM related decision making purposes. However, defining the level of customer value depends on the industries and different CLV models are suitable for specific industrial applications. In practice, when a company attempts to measure their customer profitability, it is difficult to determine the right CLV model for the company. Therefore, in this paper, we propose a framework for measuring and predicting CLV as the guideline for different industries to follow. More specifically, it is composed of five steps including data collection, data pre-processing, CLV calculations, and CLV prediction model construction and evaluation. To demonstrate the effectiveness and usability of the proposed framework, a case company of the online game industry is chosen as an example. In particular, five popular CLV models are compared in terms of CLV prediction performance. As a result, the best CLV model is identified based on the proposed framework.