The Engagement-Prolonging Designs Teens Encounter on Very Large Online Platforms

Paper Title

The Engagement-Prolonging Designs Teens Encounter on Very Large Online Platforms

Publication Info

• Topic area: Analysis of manipulative design strategies in digital platforms targeting teen users.
• Keywords: engagement-prolonging designs, manipulative design, attention economy, teens, social media, eCommerce, cognitive load, emotional manipulation, Very Large Online Platforms (VLOPs), digital regulation.

Background and Problem

• Problem / challenge: Very Large Online Platforms (VLOPs) employ engagement-prolonging designs (EPDs) to maximize user attention, often at the expense of user autonomy and wellbeing. Teens, due to their developmental stage, are particularly vulnerable to these manipulative strategies.
• Significance: Understanding and regulating EPDs is crucial to protect teens from excessive and compulsive platform use, which can disrupt sleep, education, and social interactions, and contribute to mental health challenges.
• Motivation and related work: Previous studies have documented manipulative designs in eCommerce and video games but have not systematically analyzed EPDs targeting teens on VLOPs. This paper builds on existing taxonomies and regulatory discussions to fill this gap by identifying and categorizing EPDs used by VLOPs.

Solution

• Proposed approach: A structured content analysis of 17 VLOPs (websites and apps) to identify and categorize EPDs targeting teen users, resulting in a hierarchical taxonomy of 63 patterns grouped into four strategies.
• Novelty:
  1. Development of a taxonomy that organizes EPDs by underlying manipulation strategies (pressuring, enticing, trapping, lulling).
  2. Quantification of EPD prevalence across 583 instances in 17 VLOPs, with social media platforms employing twice as many EPDs as other types.
  3. Contextual illustrations (vignettes) showing how EPDs combine to prolong engagement.
  4. Contribution of a video dataset documenting 583 EPD instances for further research and regulatory use.
• Procedure and key techniques:
  • Created teen personas and accounts to simulate teen user experiences on VLOPs.
  • Conducted inductive-deductive coding to identify EPD patterns.
  • Developed a hierarchical taxonomy of EPDs based on psychological mechanisms.
  • Conducted inter-rater reliability testing to refine the taxonomy.
  • Analyzed the prevalence and distribution of EPDs across VLOPs, categorizing them by platform type (social media, eCommerce, information).

Results

• Concrete findings:
  • A total of 583 EPD features were identified across 17 VLOPs (334 in apps, 249 in websites).
  • Social media platforms averaged 27.5 EPD features per app, compared to 14.8 for eCommerce and 13 for information platforms.
  • The most common strategies were pressuring (34.4% of app EPDs) and lulling (27.5%), while trapping (17.9%) was least common.
  • Platform-manufactured tasks and autopopulated feeds were the most prevalent patterns, while reinforcement schedules were the least used.
• Advantage over baselines:
  • Comprehensive taxonomy that integrates theoretical and empirical insights.
  • Quantitative data showing the extent and distribution of EPDs across platforms and strategies.
  • Contextual vignettes demonstrating how EPDs interact to amplify engagement.
• Experiments / evaluation:
  • Systematic coding of 17 VLOPs (web and app versions) using a refined taxonomy.
  • Statistical analysis of EPD prevalence by platform type and strategy.
  • Visualizations (e.g., tree diagrams, bar charts) illustrating EPD distribution and interaction.
• Limitations and future work:
  • Did not analyze paid membership features or non-European versions of VLOPs.
  • Short-term observations may miss long-term EPD effects.
  • Future work could involve longitudinal studies with teen participants and collaboration with industry stakeholders for deeper insights.

Summary

This study systematically analyzed engagement-prolonging designs (EPDs) across 17 Very Large Online Platforms (VLOPs) to identify manipulative strategies targeting teen users. A hierarchical taxonomy was developed, categorizing 63 EPD patterns into four strategies: pressuring, enticing, trapping, and lulling. The analysis revealed that social media platforms employed twice as many EPDs as other platform types, with 583 instances documented across apps and websites. Contextual vignettes illustrated how EPDs combine to amplify engagement. The findings highlight the need for regulatory interventions to protect teens from manipulative designs and promote autonomy-supporting digital experiences.