The study, presented today at the Conference on Human Factors in Computing Systems in Yokohama, Japan, suggests that the key to a having a positive experience online is finding the right level of personal investment – neither too much nor too little.
Conducted by researchers examining digital self-regulation, they reveal distinct user types and propose that social media platforms could be remodelled to support more intentional use.
Lead author Dan Bennett from Bristol’s Faculty of Science and Engineering said explained: “Many people feel the need to better control their time on social media. While social media offers entertainment, social connection and opportunities for personal growth, people feel the need to better manage their engagement, to avoid wasting time and engaging in a way which damages their mood and well-being.
“We know that one size does not fit all for digital self-control. People are affected differently by their social media use and have different needs for managing their time online. However, we have lacked data on what drives different experiences and needs, and how to adapt social media designs to suit these needs.”
This study introduced a person-centred machine learning approach to categorise social media users into groups based on their motivations and behaviours:
- Socially Steered Users – feel strongly constrained by peer expectations and pressures.
- Automatic Browsers – commonly find themselves engaging without thought or purpose, find their social media use “meaningless”, and struggle with overuse and regret.
- Deeply Invested Users – do connect social media to personal meaning, identity and goals but again struggle with overuse and regret.
- Goldilocks Users –see personal value in their social media use, but do not investing much of themselves in it. These users experienced the lowest levels of regret.
This research highlights the potential for personalized digital tools to help users self-regulate their social media habits. Instead of a one-size-fits-all approach, platforms could introduce customized features that support different user needs, such as helping compulsive users regain intentional control, or helping socially constrained users balance the pressures and benefits of social connection.
The findings were based on a survey of 500 participants, using psychological assessments and person-centred machine learning to identify distinct engagement styles.
Dan added: “We identify different types of users on social media — including those who browse without strong intentionality, those deeply invested in their online lives, and those who see value in using social media, but retain personal distance. While the latter group arguably has the best outcomes overall, each group presents unique challenges for self-regulation.
“By tailoring social media designs to these different needs, platforms could help users stay in control and make their time online more purposeful and valued.”
The implications of this work reach beyond social media design into technology use more broadly. In another recent paper the authors found similar groupings of users across a range of technologies, including games, and technologies for personal well-being. Together these results point to data-driven approach to design that can help promote sustainable engagement connected to things that matter to the user, rather than just maximising screen time.
The next phase of this work will explore how social media platforms can identify different user groups, and adapt interfaces to help users engage in a way that aligns with their personal well-being.
Paper:
‘Autonomous Regulation of Social Media Use: Implications for Self-control, Well-Being, and Ux’ by Dan Bennett, Feng Feng and Elisa D Meckler presented at CHI '25: Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems.