Abstract: Social network users connect and interact with one another to fulfil different kinds of social and information needs. When interaction ceases between two users, we say that their tie becomes dormant. While there are different underlying reasons of dormant ties, it is important to find means to revive such ties so as to maintain vibrancy in the relationships. In this work, we focus on designing an online experiment to evaluate the effectiveness of personalized social messages to revive dormant ties. The experiment carefully selects users with dormant ties so that no user gets mixed treatments and be affected by the responses of other users undergoing treatment. Our results show that personalized message content plays an important part in reviving dormant ties. Specifically, we find the message containing friend's recent activity information is more effective than that containing inter-friend activity information. We observe that the quality of engagement of at least 50% of the revived ties can effectively be restored to the level before the ties become dormant. We also observe that it is easier to revive dormant ties that involve users from the same country than those involving users from different countries.
Introduction
In the era of "social overloading," digital relationships often fade into silence. In social network theory, these are known as dormant ties—connections where interaction has ceased. While humans have a limited cognitive capacity for active relationships (the "Dunbar number"), social platforms strive to keep these ties alive to prevent user churn and maintain a vibrant ecosystem. Once a tie goes silent, it rarely restarts spontaneously. This study investigates whether a platform can act as a "matchmaker" to jumpstart these stalled relationships using personalized social messages, specifically testing which types of information motivate a user to reconnect.
Methodology
The researchers collaborated with a large-scale mobile social network (the "MG network") to conduct a randomized controlled trial. The study utilized a dataset of several million users, initially identifying 1,973 active users with 5,128 dormant ties. After filtering for independent ties to avoid network effects—ensuring no user in the experiment shared a common contact with another user in a different treatment group—the final experimental subset consisted of 666 dormant ties.
As shown in the workflow above, the experiment involved three distinct groups:
- M1 (Friend Activity): Users received a message stating: "Your friend [Name] has sent X messages. Check it out [Link]."
- M2 (Inter-friend Activity): Users received a message stating: "[Name] and Y other friends have exchanged X messages with your friend [Name]. Check it out [Link]."
- Control Group: Received no intervention.
The team monitored both "passive engagement" (clicks via a custom URL shortener) and "active engagement" (actual messages sent between the dormant pair) from October through December 2012.
Key Results and Insights
The experiment revealed a clear hierarchy in message effectiveness. The "Direct Interest" advantage was prominent: users receiving M1 (direct friend activity) were significantly more likely to re-engage than those in other groups.
The M2 Backfire
Surprisingly, the M2 message (mutual friend activity) performed no better than the control group. This suggests that seeing a friend interacting with others might actually discourage a user, potentially making them feel excluded or signaling that the friend is already socially occupied.
Engagement Restoration and Geography
The intervention was not merely a temporary "ping." For at least 50% of the ties revived by the M1 message, interaction intensity returned to pre-dormancy levels. Long-term tracking (2.5 months post-treatment) showed that the M1 group maintained a median engagement restoration value of 1.02, while the control group’s naturally revived ties decayed to a median of 0.63.
Furthermore, geography played a role: ties between users in the same country were significantly easier to revive than international ties. This suggests that shared cultural context or the potential for offline interaction enhances the effectiveness of a digital "nudge."
Conclusion
This study demonstrates that social networks can successfully play an active role in maintaining relationship vibrancy. By providing simple, direct information about a friend's activity, platforms can effectively revive dormant ties and restore them to their original quality. Future research is needed to explore the psychological reasons behind the failure of mutual-friend notifications and to determine if these tactics remain effective for "weak ties" or relationships that have been dormant for several years.
Summary updated