Stuck on you: How to make social media good again
This IPPR paper examines how social media platforms have shifted from user-led spaces to algorithm-driven, influencer-dominated environments through ‘sticky gatekeeping’. A UK survey found only 18 per cent of feed posts were personal content. The report recommends regulatory reform, algorithmic transparency, and development of a public social media platform.
Please login or join for free to read more.
OVERVIEW
Introduction
This paper examines how social media platforms have changed over the last decade and how these changes have shaped culture and politics in the UK. More than 70 per cent (5.6 billion) of people across the world now have social media accounts, and people in the UK spend around an hour and a half on social media every day (p.7). While harms such as disinformation and polarisation receive significant media attention, for most people social media plays a more modest role in daily life. The paper focuses on structural and collective implications, rather than individual behaviour.
Mapping the landscape
Platforms now divide into three types: closed messenger platforms, creator-led broadcast platforms, and specialised spaces such as Reddit, LinkedIn and Discord. According to Ofcom’s 2024/25 Online Nation report, audiences spend more time on broadcast platforms than other kinds (p.10). This paper examines broadcast platforms through the lens of gatekeeping.
How gatekeeping has changed
Strong gatekeeping historically saw single editors choose content disseminated to thousands or millions. Torquay’s Herald Express sold 21,000 copies per day in 2010 but only around 2,500 copies per week by 2025 (p.11–12). While strong gatekeeping enforced shared public life, it also marginalised political groups, immigrants and LGBTQ+ communities.
Weaker gatekeeping emerged between 2005 and 2015 as social media grew in popularity. Movements such as #BlackLivesMatter and #MeToo demonstrated how more people could contribute to public discourse. However, algorithmic recommendation systems introduced in 2009 reshaped who and what we see (p.12–13).
Sticky gatekeeping is the result. Stickiness — originally a 1990s marketing term — now dictates the content we see. Platforms engineer a sense of intimacy and belonging through personalisation and influencer relationships, prioritising time-on-platform over genuine social connection (p.13–14).
Personalisation
Platforms have shifted from the ‘social graph’ (who you know) to the ‘interest graph’ (what you can be sold). A survey of 1,000 UK participants found 35 per cent of feed posts were from influencers, public figures and recommended content; 29 per cent were ads and brands; and only 18 per cent were personal content from someone the user knew (p.15–16). Friend content is more likely to appear in the first post (24 per cent) versus the fourth post (14 per cent) (p.16). People aged 18–34 spend 1,470 minutes per month on TikTok versus 510 minutes per month on WhatsApp (p.15).
The UK social commerce industry was valued at £9 billion in 2025, with 200,000 UK businesses selling via TikTok (p.17). Focus group participants expressed a clear desire for non-commercial, non-algorithmic platforms oriented around genuine social interaction. Personalisation deprioritises groups and content that are not sticky, making community building and connective politics harder.
Parasocial relationships
Platforms prioritise vertical influencer-follower relationships over horizontal peer-to-peer interaction, making ordinary users less visible to each other. Ofcom found that fewer adults ‘feel freer to be themselves’ online than offline in 2025 — 25 per cent compared with 30 per cent the previous year (p.19). Nearly a third of all social media users post less than a year ago, especially Gen Z adults (p.19). A February 2026 paper found that X’s algorithms promote conservative influencers and deprioritise traditional media, with persistent effects on users’ political attitudes (p.19–20). These dynamics distort the perception of what the public thinks is important, shaping what politicians and decision-makers consider salient.
Recommendations
The paper makes five recommendations: extend prominence requirements to social media platforms beyond news to include a broader range of public benefit content; amend the Online Safety Act to ban manipulative algorithmic ‘dark patterns’, aligned with the EU Digital Services Act; revive the BBC’s Open Door programme to platform underrepresented groups without editorial input; encourage UK and European public service broadcasters to explore developing a public social media platform with a public benefit objective; and simplify the Public Interest Test and introduce a public digital infrastructure exemption in competition assessments to allow public broadcasters to collaborate and compete with large technology companies.