Behavioural insights to empower social media users

Published: 21 May 2024

Enabling users to control what appears on their feed could contribute to fostering media literacy and keeping users safe online. This discussion paper covers two online experiments run by Ofcom’s Behavioural Insight Hub that tested how platform choice architecture affects use of content controls among adult users. Our aim was not to steer users towards a particular choice, for example, reducing the amount of sensitive content they receive, but to empower them to make a choice that is right for them.

We used a mock-up of a typical social media platform. Users could choose the amount of sensitive content on their feed (“All content types” or “Reduced sensitive content”). The trials focused on two stages of the user journey.

Sign-up trial

The Sign-up trial varied how information and choices about content controls are presented to users when they set up a new social media account.

  • When “All content types” was pre-selected at sign-up, only 15% opted for “Reduced sensitive content”. The figure was 24% with no pre-selected option.
  • Presenting information with examples of sensitive content on the decision page increased the selection of “Reduced sensitive content” at sign-up by 5 percentage points, to 29%.
  • Users had a strong tendency to continue with their initial choice – 88% of participants did not change setting after seeing the feed, regardless of what their initial choice was.

Check and Update trial

The Check and Update trial tested the effect of prompting users who have already set up an account to check and update their content controls while browsing.

  • Prompts were effective in encouraging users to check settings: without a prompt, only 4% of participants checked their content settings. 17%-23% checked when prompted.
  • Prompts emphasising the ease of changing settings proved more effective than prompts aiming to provide a sense of control (23% vs 17% checked).
  • Moreover, prompts after engagement with sensitive content encouraged more participants to check their settings than prompts before engagement (21% vs 18%).

These experiments provide new insight into the ways that platform design choices shape use of content controls. Online experiments of this kind provide strong evidence but cannot fully replicate ‘real world’ behaviour. We therefore consider the relative impact of the interventions to be more robust evidence than the absolute size of impact we found.

Read the report

Behavioural insights to empower social media users – discussion paper (PDF, 1.4 MB)

Sign-up trial – technical report (PDF, 5.4 MB)

Check and update trial – technical report (PDF, 3.1 MB)

Back to top