# TWICE: Two-Clock, Two-Window Learning for Long-Horizon Conversion Prediction in Online Advertising

arXiv: https://arxiv.org/abs/2607.25404 · HTML: https://mlsanity.com/paper/2607.25404/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 13 of its 28 references that are in the map.
- First posted on arXiv 28 Jul 2026.

## Builds on (most cited first, 13 of 13)

- [Optimized Cost per Click in Taobao Display Advertising](https://mlsanity.com/paper/1703.02091/) (2017)
- [ESCM2: Entire Space Counterfactual Multi-Task Model for Post-Click Conversion Rate Estimation](https://mlsanity.com/paper/2204.05125/) (2022)
- [Addressing Delayed Feedback for Continuous Training with Neural Networks in CTR prediction](https://mlsanity.com/paper/1907.06558/) (2019)
- [Real Negatives Matter: Continuous Training with Real Negatives for Delayed Feedback Modeling](https://mlsanity.com/paper/2104.14121/) (2021)
- [A Nonparametric Delayed Feedback Model for Conversion Rate Prediction](https://mlsanity.com/paper/1802.00255/) (2018)
- [A Feedback Shift Correction in Predicting Conversion Rates under Delayed Feedback](https://mlsanity.com/paper/2002.02068/) (2020)
- [Asymptotically Unbiased Estimation for Delayed Feedback Modeling via Label Correction](https://mlsanity.com/paper/2202.06472/) (2022)
- [Unbiased Delayed Feedback Label Correction for Conversion Rate Prediction](https://mlsanity.com/paper/2307.12756/) (2023)
- [Generalized Delayed Feedback Model with Post-Click Information in Recommender Systems](https://mlsanity.com/paper/2206.00407/) (2022)
- [Learning Classifiers under Delayed Feedback with a Time Window Assumption](https://mlsanity.com/paper/2009.13092/) (2020)
- [Handling many conversions per click in modeling delayed feedback](https://mlsanity.com/paper/2101.02284/) (2021)
- [Personalized Interpolation: Achieving Efficient Conversion Estimation with Flexible Optimization Windows](https://mlsanity.com/paper/2501.14103/) (2025)
- [Delayed Feedback Modeling with Influence Functions](https://mlsanity.com/paper/2502.01669/) (2025)

## Abstract (arXiv)

Long-horizon conversion prediction under delayed feedback creates a two-clock, two-window learning problem in online advertising. A short base observation window releases recent clicks on the click clock before their outcomes mature, whereas conversions continue to arrive on the conversion clock throughout a longer target conversion window. The click clock provides timely but partially observed status supervision. The conversion clock reveals long-tail delays, but the delay composition within an arrival-time slice is weighted by historical click cohorts with different traffic volumes and target-window conversion rates. We present TWICE, a framework that factorizes long-horizon post-click conversion rate (CVR) into a target-window conversion probability and a grouped elapsed-delay cumulative distribution function (CDF). The two clocks provide complementary supervision. Click-clock records train the target-window CVR head through a current-status likelihood over the base observation window. Newly arrived conversions train the delay model on the conversion clock. To account for the cohort mixture, TWICE uses fixed click-time predicted CVR (pCVR) mass as cohort exposure in an arrival-conditioned likelihood. This accounts for differences in cohort traffic and conversion propensity. The resulting aggregate records are self-contained. A single learned CDF produces monotone predictions for all requested horizons up to the target conversion window. Serving requires neither historical lookup nor convolution. Experiments on a public benchmark and an industrial advertising dataset demonstrate the effectiveness of TWICE. In an online A/B test in Kwai's advertising system, TWICE increased expected revenue, revenue, and conversions by 2.486%, 1.858%, and 2.061%, respectively. It was subsequently deployed to full traffic.