Not logged in.
Quick Search - Contribution
Contribution Details
| Type | Conference Presentation |
| Scope | Discipline-based scholarship |
| Title | How to predict customer lifetime value? A comparison of state-of-the-art approaches for non- contractual business settings |
| Organization Unit | |
| Authors |
|
| Presentation Type | keynote |
| Item Subtype | Original Work |
| Refereed | No |
| Status | Published electronically before print/final form (Epub ahead of print) |
| Language |
|
| Event Title | INFORMS Marketing Science Society Conference 2014 |
| Event Type | conference |
| Event Location | Atlanta, Georgia, USA |
| Event Start Date | June 11 - 2014 |
| Event End Date | June 14 - 2014 |
| Abstract Text | With firms’ increasing interest in marketing analytics, customer valuation receives more attention than ever. This is clearly reflected in recent advances in customer lifetime value calculation for non-contractual business settings such as grocery retailing. However, due to the novelty of these developments there is little empirical evidence on when which model performs best. To calculate customer lifetime value in non-contractual settings it is essential to statistically model both, how long a customer will stay with the firm as well as the frequency of a customer’s future transactions. The Pareto/NBD model (Schmittlein, Morrison, and Colombo, 1987) was the first model to address both issues and is currently the de facto standard in marketing practice and research. Recently, alternative models have been proposed, such as the BG/NBD model (Fader, Hardie and Lee, 2005), the GGompertz/NBD (Bemmaor and Glady, 2012), and the Normal/NBD model (Jain and Singh, 2013). Besides, some initial work has been done on the inclusion of covariates into these models. However, no empirical comparison of those approaches across multiple real-world scenarios exists. We contribute to the literature on customer lifetime value by providing a structured review of recent modeling approaches for non-contractual business settings. Additionally, we compare the models’ performance by using four different datasets from the grocery retail, luxury goods, travel, and video-on-demand industry. Our findings have strong implications for both, marketing practice and research. Besides giving detailed recommendations on when to use which modeling approach, we also provide practical advices for estimating these models. |
| Export |
BibTeX
|