Predicting customer value per product: From RFM to RFM/P

Recency, frequency, and monetary (RFM) models are widely used to estimate customer value. However, they are based on the customer perspective and do not take the product perspective into account. Furthermore, predictability decreases when recency, frequency, and monetary values vary among product categories. A RFM per product (RFM/P) model is proposed to first estimate customer values per product and then aggregate them to obtain the overall customer value. Empirical applications for a financial services company and a supermarket demonstrate that RFM/P opens up the possibility to combine customer and product perspectives. Additionally, when there are changes in customer purchase behavior regarding recency per product and frequency per product, which is usual, RFM/P prediction accuracy was found to be better than traditional RFM.

Automated summary

Traditionally, RFM models estimate customer value based solely on customer perspective, ignoring the product perspective, which can lead to decreased prediction accuracy when RFM values vary across product categories. RFM/P model first estimates customer value per product, allowing for better prediction accuracy when customer purchase behavior changes with regards to recency and frequency per product.