Targeted and Personalized Online Advertising in the Age of Artificial Intelligence (AI): A Literature Review and Research Agenda

Abstract

This study aims to provide a comprehensive evaluation of current machine learning (ML) algorithms employed in targeted and personalized advertising. It reveals key findings and conclusions from a wide range of sources, offering readers a concise summary. The study addresses the gap by identifying and analyzing the most significant machine learning-based targeting methods utilized in the recent studies. This helps readers understand the strengths and weaknesses of different approaches and keeps them up-to-date with the most recent advancements and best practices. Employing the PRISMA methodology, the review systematically examines existing literature on ML-driven targeted advertising. It identifies effective ML methods and strategies, presenting real-world examples to illustrate their practical implementation. Reviewing key findings from existing literature, the analysis identifies the most effective ML methods for targeted advertising. It also examines three research questions across three key dimensions: targeting, personalizing, and predicting customer preferences. This study proposes a novel theoretical framework that elucidates the application of ML in targeted advertising. Specifically, the study explores ML algorithms that enhance precision in each dimension. Key models include Long Short-Term Memory (LSTM) networks for analyzing historical customer data, Convolutional Neural Networks (CNN) for image recognition tasks, and Factorization Machines for capturing feature interactions in click-through rate (CTR) predictions. Additionally, traditional models such as logistic regression, decision trees, random forests, and support vector machines (SVM) are utilized for classification tasks, while unsupervised learning techniques like k-means clustering and hierarchical clustering facilitate user segmentation based on behavioral and demographic similarities. These models collectively enable marketers to derive actionable insights, optimize advertising content, and improve overall campaign performance. By consolidating key findings from existing literature on ML-driven targeted advertising, this study offers a valuable resource for understanding current trends and gaps. It also proposes future research directions, highlighting potential areas for further exploration, which can inspire new studies and innovations in the field.