“Tell Me Why You’re Asking”: Exploring How to Increase Engagement in Preference Feedback for Intelligent Notification Systems

Paper Title

“Tell Me Why You’re Asking”: Exploring How to Increase Engagement in Preference Feedback for Intelligent Notification Systems

Publication Info

• Topic area: User engagement and preference elicitation in intelligent notification systems.
• Keywords: Intelligent notification systems, preference feedback, user engagement, personalization, timing, transparency, adaptive systems, human-AI interaction, notification management, co-learning.

Background and Problem

• Problem / challenge: Existing intelligent notification systems rely heavily on user input for personalization but primarily collect preferences during initial setup. This approach fails to account for dynamic and situational changes in user needs, and sustained engagement in preference feedback remains a challenge.
• Significance: Understanding how users prefer to express notification preferences is critical for designing systems that adapt meaningfully to their needs while respecting their effort and attention. This has practical implications for reducing notification overload and improving user satisfaction.
• Motivation and related work: Prior research has explored implicit and explicit preference elicitation mechanisms, timing of notifications, and user engagement strategies. However, gaps remain in understanding when, how, and under what conditions users are willing to provide feedback, especially in the context of dynamic notification systems.

Solution

• Proposed approach: Semi-structured interviews with 33 participants using design probes to explore user willingness, strategies, and expectations for expressing notification preferences.
• Novelty:
  1. Identifying the importance of justifiability in timing for preference elicitation, particularly embedding requests within notification-handling routines.
  2. Highlighting the need for clarity in preference formation and consequences to sustain engagement.
  3. Revealing user expectations for notification systems to act as evolving partners capable of distinguishing short-term situational needs from long-term preference changes.
• Procedure and key techniques:
  • Speed-dating approach with speculative scenarios to prompt reflection on preference expression.
  • Use of mockups illustrating diverse interaction modalities and notification-management functions.
  • Iterative recruitment and grounded theory analysis to identify themes influencing user engagement.

Results

• Concrete findings:
  • Lightweight mechanisms like binary inputs are preferred for low-effort interactions, but users are willing to invest effort for nuanced control.
  • Timing of feedback requests is critical; users prefer requests aligned with their notification-handling routines and initiated after engaging with a notification.
  • Transparency and immediacy in system response are essential for trust and sustained engagement.
  • Users expect systems to adapt dynamically to situational and long-term changes in preferences.
• Advantage over baselines:
  • Extends prior work by emphasizing the construct of justifiability in timing and identifying routine-specific timing requirements unique to notification systems.
  • Highlights the importance of mutual learning between users and systems for dynamic adaptation.
• Experiments / evaluation:
  • Semi-structured interviews with 33 participants aged 20–64, recruited across three waves to maximize diversity in notification habits and generative-AI familiarity.
  • Analysis using constructivist grounded theory to identify themes influencing preference expression.
• Limitations and future work:
  • Lack of real-world prototype testing; findings based solely on interviews and speculative scenarios.
  • Design probes may have constrained ideation by focusing on specific functions.
  • Results may not generalize across cultural contexts or device ecosystems.
  • Future research should explore alternative device contexts and cross-cultural applicability.

Summary

This study investigates how users prefer to express notification preferences in intelligent systems, revealing that willingness depends on timing, transparency, and clarity in preference formation and consequences. Feedback requests are most acceptable when embedded within users’ notification-handling routines and aligned with their immediate context. Sustained engagement requires systems to provide immediate, verifiable feedback and adapt dynamically to both situational and long-term changes in preferences. These findings suggest that intelligent notification systems should function as adaptive companions, fostering trust and collaborative sensemaking to better support users in managing notifications.