Delayed Feedback Modeling for Post-Click Gross Merchandise Volume Prediction: Benchmark, Insights and Approaches

TL;DR

This work tackles delayed feedback in post-click GMV prediction by introducing TRACE, a benchmark that captures complete transaction sequences from each ad click to support online streaming modeling. It then proposes READER, a RepurchasE-Aware Dual-branch prEdictoR with a dedicated sample router and debiasing mechanisms, designed for online learning where labels evolve over the attribution window. TRACE experiments show online streaming training yields clear accuracy gains and that single-purchase versus repurchase samples benefit from separate modeling. READER achieves superior performance over strong baselines (e.g., ~0.86% AUC gain and ~2.19% ACC gain) and ablations confirm the effectiveness of its dual-branch architecture, routing strategy, and debiasing modules, providing a practical path for improved GMV prediction in dynamic online advertising.

Abstract

The prediction objectives of online advertisement ranking models are evolving from probabilistic metrics like conversion rate (CVR) to numerical business metrics like post-click gross merchandise volume (GMV). Unlike the well-studied delayed feedback problem in CVR prediction, delayed feedback modeling for GMV prediction remains unexplored and poses greater challenges, as GMV is a continuous target, and a single click can lead to multiple purchases that cumulatively form the label. To bridge the research gap, we establish TRACE, a GMV prediction benchmark containing complete transaction sequences rising from each user click, which supports delayed feedback modeling in an online streaming manner. Our analysis and exploratory experiments on TRACE reveal two key insights: (1) the rapid evolution of the GMV label distribution necessitates modeling delayed feedback under online streaming training; (2) the label distribution of repurchase samples substantially differs from that of single-purchase samples, highlighting the need for separate modeling. Motivated by these findings, we propose RepurchasE-Aware Dual-branch prEdictoR (READER), a novel GMV modeling paradigm that selectively activates expert parameters according to repurchase predictions produced by a router. Moreover, READER dynamically calibrates the regression target to mitigate under-estimation caused by incomplete labels. Experimental results show that READER yields superior performance on TRACE over baselines, achieving a 2.19% improvement in terms of accuracy. We believe that our study will open up a new avenue for studying online delayed feedback modeling for GMV prediction, and our TRACE benchmark with the gathered insights will facilitate future research and application in this promising direction. Our code and dataset are available at https://github.com/alimama-tech/OnlineGMV .

Figures (5)

• Figure 1: Label accumulation processes in GMV prediction, where the final label equals to the sum of the transaction prices of all purchases within the attribution window $w_a$.
• Figure 2: The hourly evolution of the average GMV label.
• Figure 3: The cumulative proportion of GMV over time.
• Figure 4: The GMV distributions of single-purchase and repurchase samples differ significantly.
• Figure 5: The workflow of our READER model.