Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

HOB: A Holistically Optimized Bidding Strategy under Heterogeneous Auction Mechanisms with Organic Traffic

Qi Li, Wendong Huang, Qichen Ye, Wutong Xu, Cheems Wang, Rongquan Bai, Wei Yuan, Guan Wang, Chuan Yu, Jian Xu

TL;DR

This paper tackles optimal bidding in a modern advertising ecosystem with heterogeneous auctions and organic traffic. It develops a holistically optimized bidding (HOB) framework that (i) derives an efficient optimal bid in first-price channels accounting for zero-cost organic traffic using a ZIE-winning-price model and (ii) introduces Marginal Cost Alignment (MCA) to achieve global efficiency across SPA and FPA channels. The approach is formalized as a surplus-maximization problem, transformed into tractable subproblems (MCKP) and solved with duality, while MCA equalizes marginal costs across channels. Extensive offline tests on synthetic and real-world data, robustness checks, and large-scale online A/B experiments (including deployment in Alibaba) demonstrate consistent GMV and ROI gains, with practical latency suitable for production. Overall, HOB provides a scalable, deployable solution for multi-channel bidding that robustly leverages organic traffic and heterogeneous auction rules to maximize advertiser value.

Abstract

The E-commerce advertising platforms typically sell commercial traffic through either second-price auction (SPA) or first-price auction (FPA). SPA was historically prevalent due to its dominant strategy incentive-compatible (DSIC) for bidders with quasi-linear utilities, especially when budgets are not a binding constraint, while FPA has gained more prominence for offering higher revenue potential to publishers and avoiding the possibility for discriminatory treatment in personalized reserve prices. Meanwhile, on the demand side, advertisers are increasingly adopting platform-wide marketing solutions akin to QuanZhanTui, shifting from spending budgets solely on commercial traffic to bidding on the entire traffic for the purpose of maximizing overall sales. For automated bidding systems, such a trend poses a critical challenge: determining optimal strategies across heterogeneous auction channels to fulfill diverse advertiser objectives, such as maximizing return (MaxReturn) or meeting target return on ad spend (TargetROAS). To overcome this challenge, this work makes two key contributions. First, we derive an efficient solution for optimal bidding under FPA channels, which takes into account the presence of organic traffic - traffic can be won for free. Second, we introduce a marginal cost alignment (MCA) strategy that provably secures bidding efficiency across heterogeneous auction mechanisms. To validate performance of our developed framework, we conduct comprehensive offline experiments on public datasets and large-scale online A/B testing, which demonstrate consistent improvements over existing methods.

HOB: A Holistically Optimized Bidding Strategy under Heterogeneous Auction Mechanisms with Organic Traffic

TL;DR

This paper tackles optimal bidding in a modern advertising ecosystem with heterogeneous auctions and organic traffic. It develops a holistically optimized bidding (HOB) framework that (i) derives an efficient optimal bid in first-price channels accounting for zero-cost organic traffic using a ZIE-winning-price model and (ii) introduces Marginal Cost Alignment (MCA) to achieve global efficiency across SPA and FPA channels. The approach is formalized as a surplus-maximization problem, transformed into tractable subproblems (MCKP) and solved with duality, while MCA equalizes marginal costs across channels. Extensive offline tests on synthetic and real-world data, robustness checks, and large-scale online A/B experiments (including deployment in Alibaba) demonstrate consistent GMV and ROI gains, with practical latency suitable for production. Overall, HOB provides a scalable, deployable solution for multi-channel bidding that robustly leverages organic traffic and heterogeneous auction rules to maximize advertiser value.

Abstract

The E-commerce advertising platforms typically sell commercial traffic through either second-price auction (SPA) or first-price auction (FPA). SPA was historically prevalent due to its dominant strategy incentive-compatible (DSIC) for bidders with quasi-linear utilities, especially when budgets are not a binding constraint, while FPA has gained more prominence for offering higher revenue potential to publishers and avoiding the possibility for discriminatory treatment in personalized reserve prices. Meanwhile, on the demand side, advertisers are increasingly adopting platform-wide marketing solutions akin to QuanZhanTui, shifting from spending budgets solely on commercial traffic to bidding on the entire traffic for the purpose of maximizing overall sales. For automated bidding systems, such a trend poses a critical challenge: determining optimal strategies across heterogeneous auction channels to fulfill diverse advertiser objectives, such as maximizing return (MaxReturn) or meeting target return on ad spend (TargetROAS). To overcome this challenge, this work makes two key contributions. First, we derive an efficient solution for optimal bidding under FPA channels, which takes into account the presence of organic traffic - traffic can be won for free. Second, we introduce a marginal cost alignment (MCA) strategy that provably secures bidding efficiency across heterogeneous auction mechanisms. To validate performance of our developed framework, we conduct comprehensive offline experiments on public datasets and large-scale online A/B testing, which demonstrate consistent improvements over existing methods.
Paper Structure (28 sections, 3 theorems, 13 equations, 5 figures, 5 tables, 3 algorithms)

This paper contains 28 sections, 3 theorems, 13 equations, 5 figures, 5 tables, 3 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Auto Bidding
  4. Bid Shading
  5. Bidding Across Multiple Channels
  6. Algorithms
  7. Problem Formulation
  8. Local optimal in individual channels.
  9. Global optimal across different channels.
  10. Optimal Bid in FPA Channel with Organic Traffic
  11. Winning Price Distribution Estimation
  12. Surplus Maximization
  13. Global Optimal with Marginal Cost Adjustment
  14. Marginal Cost under SPA
  15. Marginal Cost under FPA
  16. ...and 13 more sections

Key Result

theorem 1

Let the surplus of placing a bid $x \in [0, V]$ be $g(x) = (V - x) F(x)$, where $V = \eta v$. If the winning price follows a ZIE distribution with parameters $\pi \in [0, 1)$ and $\lambda > 0$, such that $F(x) = \pi + (1 - \pi)(1 - e^{-\lambda x})$, then the surplus function $g(x)$ is strictly unimo

Figures (5)

  • Figure 1: Illustration of the procedure. Our work focuses on the design of holistically optimized bidding strategy.
  • Figure 2: Comparison between the actual CDF (black) and CDFs fitted using nonlinear least squares with different distribution models.
  • Figure 3: The architecture of our proposed HOB. While the Optimal Bid Calculation module (left) computes a locally optimal bid for a primary channel, achieving the global optimum across heterogeneous channels (right) is orchestrated by the Marginal Cost Adjustment module, which dynamically coordinates strategies using cross-channel feedback.
  • Figure 4: HOB's improvements on different budget levels.
  • Figure 5: Analysis on different channel proportions.

Theorems & Definitions (3)

  • theorem 1: Unimodality of the Bidding Surplus Function
  • Lemma 1: Marginal Cost Equalization
  • theorem 2: Marginal Cost under FPA with Shading