Jointly Optimizing Debiased CTR and Uplift for Coupons Marketing: A Unified Causal Framework
Researchers have developed a new framework called UniMVT to address confounding bias in online advertising, particularly for coupon marketing. This model disentangles user preferences from the effects of interventions like coupons, allowing for more accurate prediction of base Click-Through Rates (CTR) and estimation of uplift. UniMVT handles multi-valued treatments and has shown significant improvements in predictive accuracy, calibration, and business metrics through real-world A/B tests. AI
IMPACT Improves accuracy in ad targeting and coupon distribution by mitigating bias in CTR prediction.