在綫廣告中用戶標籤的價值:標籤數量與結構規律﹑認知情景差異與廣告優化

Translated title of the thesis: Value of User Tags in Online Advertising: Tag Quantity and Structural Patterns, Cognitive Context Differences, and Ad Optimizations

Student thesis: Doctoral Thesis

Abstract

As a core data element in online advertising recommendation systems, user tags accurately depict users’ interests, behavior patterns, and consumption preferences, thereby providing robust support for advertisers to develop personalized marketing strategies. Playing a pivotal role in various aspects such as precision marketing, user profiling, and advertising effectiveness evaluation, user tags have become a vital link connecting advertisers with target audiences. Traditional marketing theories generally posit that a greater number of user tags enables more precise market segmentation and user need identification, thereby enhancing advertising conversion efficiency and dissemination effectiveness to some extent. However, constrained by the limitations in data collection and processing capabilities, existing studies often examine tag value and advertising strategies independently, overlooking the deeper influence of users’ cognitive context and advertising optimization processes on the endogenous transformation of tag value.

This study centers on the transformation of tag value, aiming to systematically explore the intrinsic logic and transformation mechanisms of user tags in online advertising, and to construct an interdisciplinary theoretical framework. Leveraging massive datasets from large-scale advertising platforms (Ad Exchanges), this research conducts empirical analyses: First, from the perspective of tag structure, it investigates the balance between the number and quality of tags. The findings reveal that more tags do not necessarily equate to better performance; instead, a reasonable combination of breadth and depth is essential for precise user targeting. Second, by examining the cognitive context shaped by different advertising channels, the study uncovers how various channels influence users’ cognitive states—positioning some users at the center and others at the periphery—thus creating diversified value transformation paths. Third, focusing on the features of ad content, it identifies which textual and visual features are more effective in driving user clicks, and how the effectiveness of these features depends on user tag characteristics, thereby informing the personalized optimization of ad content and further improving advertising performance.

This research breaks through the linear cognitive framework of “more tags mean higher precision” commonly found in traditional marketing theory. It reveals the moderating role of cognitive context in the transformation of tag value and proposes a mechanism through which channel characteristics shape users’ cognitive states. Additionally, it introduces an approach for optimizing ad content based on user tags. These findings advance the evolution of precision marketing from static analysis toward dynamic mechanisms.

On a practical level, the study offers actionable solutions, including: (1) Tag combination strategies: emphasizing the coordinated configuration of core and contextual tags to avoid ineffective tag redundancy; (2) Cognitive-channel adaptation: developing differentiated tag application rules based on the cognitive characteristics of various advertising channels; (3) Intelligent content generation: dynamically matching ad content with user tags to transition creative optimization from manual experience to a data-driven paradigm. Together, these contribute to a full-chain solution from strategy formulation to effectiveness evaluation for the advertising industry.
Date of Award8 Jul 2025
Original languageChinese (Traditional)
Awarding Institution
  • City University of Hong Kong
SupervisorChoon Ling SIA (Supervisor) & Cheng ZHANG (External Supervisor)

Cite this

'