Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

OKG: On-the-Fly Keyword Generation in Sponsored Search Advertising

Zhao Wang, Briti Gangopadhyay, Mengjie Zhao, Shingo Takamatsu

TL;DR

This paper tackles the problem of real-time keyword optimization in sponsored search advertising, where traditional methods rely on static datasets. It introduces OKG, an LLM agent that monitors live KPI signals and real-time data to adapt keyword generation along two directions: wider exploration and deeper exploitation. A publicly accessible dataset with real ad keywords and KPIs is presented as a resource for future research, and the authors demonstrate that OKG improves adaptability and responsiveness over baselines. The work provides code and data to facilitate reproducibility and practical deployment in dynamic SSA environments.

Abstract

Current keyword decision-making in sponsored search advertising relies on large, static datasets, limiting the ability to automatically set up keywords and adapt to real-time KPI metrics and product updates that are essential for effective advertising. In this paper, we propose On-the-fly Keyword Generation (OKG), an LLM agent-based method that dynamically monitors KPI changes and adapts keyword generation in real time, aligning with strategies recommended by advertising platforms. Additionally, we introduce the first publicly accessible dataset containing real keyword data along with its KPIs across diverse domains, providing a valuable resource for future research. Experimental results show that OKG significantly improves keyword adaptability and responsiveness compared to traditional methods. The code for OKG and the dataset are available at https://github.com/sony/okg.

OKG: On-the-Fly Keyword Generation in Sponsored Search Advertising

TL;DR

This paper tackles the problem of real-time keyword optimization in sponsored search advertising, where traditional methods rely on static datasets. It introduces OKG, an LLM agent that monitors live KPI signals and real-time data to adapt keyword generation along two directions: wider exploration and deeper exploitation. A publicly accessible dataset with real ad keywords and KPIs is presented as a resource for future research, and the authors demonstrate that OKG improves adaptability and responsiveness over baselines. The work provides code and data to facilitate reproducibility and practical deployment in dynamic SSA environments.

Abstract

Current keyword decision-making in sponsored search advertising relies on large, static datasets, limiting the ability to automatically set up keywords and adapt to real-time KPI metrics and product updates that are essential for effective advertising. In this paper, we propose On-the-fly Keyword Generation (OKG), an LLM agent-based method that dynamically monitors KPI changes and adapts keyword generation in real time, aligning with strategies recommended by advertising platforms. Additionally, we introduce the first publicly accessible dataset containing real keyword data along with its KPIs across diverse domains, providing a valuable resource for future research. Experimental results show that OKG significantly improves keyword adaptability and responsiveness compared to traditional methods. The code for OKG and the dataset are available at https://github.com/sony/okg.

Paper Structure

This paper contains 32 sections, 6 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Direct Keyword Generation Using Generative Methods
  4. Using GANs for Keyword Generation
  5. NMT for Constrained Keyword Retrieval
  6. Advancements in Keyword Generation Using Large Language Models
  7. LLM for Document Retrieval
  8. LLM for Keyword Generation in Sponsored Search
  9. Limitations of Current Generative and LLM-Based Approaches
  10. Dependence on Large Datasets
  11. Limited Real-Time Adaptability
  12. Lack of Continuous Monitoring
  13. Problem Setting
  14. Methodology
  15. Key Components of OKG
  16. ...and 17 more sections

Figures (4)

  • Figure 1: This visual contrasts the traditional keyword generation strategy with our OKG Agent, demonstrating the motivation behind our work.
  • Figure 2: The architecture of OKG, which fulfills the functionality of online search, real-time keyword and KPI retrieval, adaptive keyword generation, calculation and etc.
  • Figure 3: Comparison Results of Component Ablation.
  • Figure 4: An intuitive example of OKG generation prompt for Sony Bank's Mortgage Service