9:20 AM - 9:40 AM
[2Q1-J-2-02] LTV Prediction Based on RFM and Customer Characteristics
Keywords:lifetime value, probabilistic predictive models
Lifetime Value (LTV) as we know it, is an important indicator of customer evaluation. To build long-term
relationships with the right customer, it is important to predict LTV with increasingly higher accuracy levels.
Once we attain that, we would be able to communicate with them through appropriate marketing actions. While
predicting LTV in a non-contractual setting, three indicators, namely; Recency, Frequency and Monetary Value
(RFM) are widely used. RFM is used as an indicator of customers’ buying behaviour on the whole, however
normally dimensions like demographics are not considered. In this paper, we propose a model for predicting LTV
based not just on RFM, but also other customer characteristics. To support our proposal and its effectiveness we
have also provided the details of the experiments, their outputs and our inference using a real dataset.
relationships with the right customer, it is important to predict LTV with increasingly higher accuracy levels.
Once we attain that, we would be able to communicate with them through appropriate marketing actions. While
predicting LTV in a non-contractual setting, three indicators, namely; Recency, Frequency and Monetary Value
(RFM) are widely used. RFM is used as an indicator of customers’ buying behaviour on the whole, however
normally dimensions like demographics are not considered. In this paper, we propose a model for predicting LTV
based not just on RFM, but also other customer characteristics. To support our proposal and its effectiveness we
have also provided the details of the experiments, their outputs and our inference using a real dataset.