![[2602.12972] Jointly Optimizing Debiased CTR and Uplift for Coupons Marketing: A Unified Causal Framework](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.12972] Jointly Optimizing Debiased CTR and Uplift for Coupons Marketing: A Unified Causal Framework
Summary
This paper presents a novel framework, UniMVT, for optimizing debiased Click-Through Rate (CTR) and uplift in coupon marketing, addressing confounding biases and improving predictive accuracy.
Why It Matters
With the increasing reliance on coupons in digital marketing, understanding and accurately predicting user engagement is crucial. This framework tackles biases that can distort marketing strategies, leading to more effective coupon distribution and improved business outcomes.
Key Takeaways
- UniMVT framework effectively disentangles confounding factors in CTR prediction.
- The model improves accuracy in estimating both debiased CTR and uplift.
- Real-world A/B tests validate the framework's effectiveness in enhancing business metrics.
Computer Science > Social and Information Networks arXiv:2602.12972 (cs) [Submitted on 13 Feb 2026] Title:Jointly Optimizing Debiased CTR and Uplift for Coupons Marketing: A Unified Causal Framework Authors:Siyun Yang, Shixiao Yang, Jian Wang, Di Fan, Kehe Cai, Haoyan Fu, Jiaming Zhang, Wenjin Wu, Peng Jiang View a PDF of the paper titled Jointly Optimizing Debiased CTR and Uplift for Coupons Marketing: A Unified Causal Framework, by Siyun Yang and 8 other authors View PDF HTML (experimental) Abstract:In online advertising, marketing interventions such as coupons introduce significant confounding bias into Click-Through Rate (CTR) prediction. Observed clicks reflect a mixture of users' intrinsic preferences and the uplift induced by these interventions. This causes conventional models to miscalibrate base CTRs, which distorts downstream ranking and billing decisions. Furthermore, marketing interventions often operate as multi-valued treatments with varying magnitudes, introducing additional complexity to CTR prediction. To address these issues, we propose the \textbf{Uni}fied \textbf{M}ulti-\textbf{V}alued \textbf{T}reatment Network (UniMVT). Specifically, UniMVT disentangles confounding factors from treatment-sensitive representations, enabling a full-space counterfactual inference module to jointly reconstruct the debiased base CTR and intensity-response curves. To handle the complexity of multi-valued treatments, UniMVT employs an auxiliary intensity estimation task to ...
