Time-Decay Attribution Model Using Half-Life in Exponential Decay for E-Commerce Conversion Analysis

Abstract

With the rapid growth of e-commerce platform users, it has become essential for companies to analyze the customer journey leading to a purchase. Users may interact with or be exposed to multiple devices and touchpoints/channels throughout this process. So now, to identify which of those touchpoints resulted in a positive conversion, the companies use various types of attribution models. By identifying this, the company then allocates its marketing spend accordingly in the future on the chosen channels to further optimize its sales and customized user experience. Over time, various types of attribution models have been developed, each model having its pros and cons. This study proposes a novel Time-Decay Attribution Model using Half-life in Exponential Decay (TD-HLED). We also provide a comparative analysis with six different Models: First-touch, Markov Chain Model, Linear-Touch, Last-touch, and our proposed model, TDHLED. Furthermore, extensive experiments have proved that the model attributions in TD-HLED have more accurately reflected how the users have interacted with the marketing than the existing simplistic rule-based architectures.