See Beyond a Single View: Multi-Attribution Learning Leads to Better Conversion Rate Prediction

2025-08-21

AI Summary

Traditional CVR prediction relies on a single attribution label (e.g., last-click), ignoring complementary signals from multiple attribution perspectives, thus limiting modeling capacity. To address this, we propose MAL (Multi-Attribution Learning), the first framework to jointly model labels generated by diverse attribution mechanisms (e.g., first-click, last-click). MAL features a dual-module architecture: ATK (Attribution Knowledge Aggregation) for cross-attribution signal fusion and PTP (Probability-Calibrated Prediction) for calibrated output estimation. We further introduce CAT (Cartesian-Augmented Training), a novel training strategy leveraging Cartesian product combinations of multi-attribution labels to enhance knowledge transfer. MAL is deployment-friendly—requiring no modifications to online serving infrastructure. Offline evaluation shows a +0.51% GAUC gain; online A/B testing yields a +2.6% ROI improvement over single-attribution baselines. Our core contributions are: (1) a paradigm for joint modeling of multi-attribution labels; (2) the CAT training mechanism; and (3) a lightweight, efficient dual-module architecture.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Multi-modal Vision
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: User modeling for targeted and personalized online advertising
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Conversion rate (CVR) prediction is a core component of online advertising systems, where the attribution mechanisms-rules for allocating conversion credit across user touchpoints-fundamentally determine label generation and model optimization. While many industrial platforms support diverse attribution mechanisms (e.g., First-Click, Last-Click, Linear, and Data-Driven Multi-Touch Attribution), conventional approaches restrict model training to labels from a single production-critical attribution mechanism, discarding complementary signals in alternative attribution perspectives. To address this limitation, we propose a novel Multi-Attribution Learning (MAL) framework for CVR prediction that integrates signals from multiple attribution perspectives to better capture the underlying patterns driving user conversions. Specifically, MAL is a joint learning framework consisting of two core components: the Attribution Knowledge Aggregator (AKA) and the Primary Target Predictor (PTP). AKA is implemented as a multi-task learner that integrates knowledge extracted from diverse attribution labels. PTP, in contrast, focuses on the task of generating well-calibrated conversion probabilities that align with the system-optimized attribution metric (e.g., CVR under the Last-Click attribution), ensuring direct compatibility with industrial deployment requirements. Additionally, we propose CAT, a novel training strategy that leverages the Cartesian product of all attribution label combinations to generate enriched supervision signals. This design substantially enhances the performance of the attribution knowledge aggregator. Empirical evaluations demonstrate the superiority of MAL over single-attribution learning baselines, achieving +0.51% GAUC improvement on offline metrics. Online experiments demonstrate that MAL achieved a +2.6% increase in ROI (Return on Investment).

Problem

• Integrating multiple attribution perspectives for better conversion prediction
• Overcoming limitations of single-attribution training in advertising systems
• Generating calibrated conversion probabilities aligned with system metrics

Innovation

• Multi-Attribution Learning framework for CVR prediction
• Attribution Knowledge Aggregator integrates diverse attribution labels
• Cartesian product training strategy enriches supervision signals