https://iaeme.com/Home/journal/JARET 103 editor@iaeme.com
Journal of Advanced Research Engineering and Technology (JARET)
Volume 3, Issue 1, January-June 2024, pp. 103-116, Article ID: JARET_03_01_010
Available online at https://iaeme.com/Home/issue/JARET?Volume=3&Issue=1
Journal ID: 2120-0202, ISSN Online: 2295-5152
Impact Factor (2024): 16.12 (Based on Google Scholar Citation)
© IAEME Publication
OPTIMIZING CONTEXTUAL ADVERTISING
WITH LARGE LANGUAGE MODELS: A
UNIFIED APPROACH TO AD CONTENT
GENERATION AND TARGETING
Lav Kumar
Salesforce Inc., USA
Karan Khanna
Netflix, USA
ABSTRACT
For businesses to reach their target group, contextual advertising is becoming more
and more important. But traditional methods often use separate steps for creating ad
content and keyword targeting, which wastes time and makes ads work less well than
they could. This article talks about a new method called Contextual Ad Generation and
Targeting (CAGT). It uses Large Language Models (LLMs) to create ad content and
improve keyword targeting all at the same time, within a single structure. To make ads
more relevant and useful, CAGT uses adaptable prompt engineering and reinforcement
learning-based feedback loops. CAGT increases the click-through rate (CTR) by 23%
and the conversion rate (CR) by 18% compared to traditional methods. These results
were found using a dataset of 10 million ad impressions from a major e-commerce site.
When CAGT was used in a commercial ad network, it showed even better results, with
an average increase in click-through rate (CTR) of 27% and an increase in click-
through rate (CR) of 21% across 50 advertising programs over 3 months. The results
show that LLMs could change contextual advertising by making ads more interesting
and improving targeting all at the same time.
Keywords: Contextual Advertising, Large Language Models (LLMs), Ad Content
Generation, Keyword Targeting, Reinforcement Learning
Cite this Article: Gaurava Srivastava and Abhi Ram Reddy Salammagari, Designing
Conversational Bots for Real-World Challenges: A Comprehensive Approach, Journal
of Advanced Research Engineering and Technology (JARET), 3(1), 2024, pp. 103-116.
https://iaeme.com/Home/issue/JARET?Volume=3&Issue=1