Why They Link: An Intent Taxonomy for Including Hyperlinks in Social Posts

TL;DR

The paper investigates reader-perceived intentions behind including URLs in social posts, addressing the lack of URL-specific intent taxonomies for information retrieval. It introduces a hybrid three-phase approach—crowdsourced bottom-up annotation refined by large-language models—to build a taxonomy with $6$ top-level categories, $28$ intention groups, and $442$ fine-grained classes. Through two crowd-labeling studies on $1000$ URL-bearing tweets, it characterizes prevalent intents (e.g., promotion, argument/conversation, sharing) and demonstrates reliability improvements when providing contextual information. It also shows that incorporating hyperlink-intent signals into retrieval improves ranking metrics such as $nDCG@10$ and $MAP$, supporting downstream NLP tasks, crisis informatics, and misinformation mitigation. Overall, the work offers a reproducible, interpretable framework that enhances intent-aware information retrieval and content understanding in social media analyses.

Abstract

URLs serve as bridges between social media platforms and the broader web, linking user-generated content to external information resources. On Twitter (X), approximately one in five tweets contains at least one URL, underscoring their central role in information dissemination. While prior studies have examined the motivations of authors who share URLs, such author-centered intentions are difficult to observe in practice. To enable broader downstream use, this work investigates reader-centered interpretations, i.e., how users perceive the intentions behind hyperlinks included in posts. We develop an intent taxonomy for including hyperlinks in social posts through a hybrid approach that begins with a bottom-up, data-driven process using large-scale crowdsourced annotations, and is then refined using large language model assistance to generate descriptive category names and precise definitions. The final taxonomy comprises 6 top-level categories and 26 fine-grained intention classes, capturing diverse communicative purposes. Applying this taxonomy, we annotate and analyze 1000 user posts, revealing that advertising, arguing, and sharing are the most prevalent intentions. This resulting taxonomy provides a foundation for intent-aware information retrieval and NLP applications, enabling more accurate retrieval, recommendation, and understanding of social media content.

Figures (5)

• Figure 1: An example of a tweet with the URL of a product web page, with the intent of inquiring about it.
• Figure 2: An example of a tweet that promises a free streaming for a sport match throught the URL.
• Figure 3: Our 3-phase framework of developing our intent taxonomy for sharing URLs in Twitter
• Figure 4: The distribution of overall intentions in Study 1.
• Figure 5: Example of two tweets where providing context to the workers (conversation) improved their agreement.