| Summary: | Customer Lifetime Value (CLV), the present value of the future net cash flows associated with a particular customer, is considered as one of the ways to increase competitive advantage. This thesis proposes a new method to utilize CLV models in the marketing management of retail e-commerce in Central Europe. The method focuses on 3 stakeholder goals: firstly a self-assessment of the needs for Customer Value Modeling, secondly deciding on the appropriate approach of CLV modeling, and finally mapping specific applications of CLV models for marketing management and experiments.
Literature research demonstrates these companies differ from other industries, e.g. telecommunications and finance, mainly due to non-contractual business settings. Preliminary qualitative research with 4 marketing managers to identify and refine problem context and barriers to adoption of CLV models discovered 6 important factors, although there is no single common barrier.
Out of many researched CLV models, 12 different types suitable for e-commerce are compared and their predictive performance is empirically evaluated on 10 e-commerce datasets. The datasets are very recent and robust – from the years between 2008 and 2017, ranging from 104 to 381 weeks of data and covering from 15,000 to almost 800,000 customers. No single model has outperformed the rest in all selected evaluation criteria. It was demonstrated that BG/NBD and EP/NBD models had consistently performed well, and could thus be considered good and stable for e-commerce businesses. Probabilistic models outperformed both naïve and econometric models. Alternative and more complex models that incorporated non-transactional data, haven’t beaten the performance and shall not be considered as good models for e-commerce in the default settings.
The main output of this thesis is a stand-alone method to utilize CLV. It also includes two interactive workbooks to help business users navigate through business settings self-assessment and an analysis workbook for introductory Customer Base Analysis. Market validation of the proposed artifact via in-depth interviews with 4 marketing managers and by 6 expert opinions proved to satisfy all 3 defined stakeholder goals. The method is thus properly designed and ready to be utilized by business practitioners. |