The Stopper Effect: Visual Stopping Power in the Attention Economy
Research · 2026-09-12
Key findings from our latest study on how visual content captures attention on social media.
What the study asks
Every scroll is a sequence of small decisions. Most visuals pass by; a few make the thumb stop. The Stopper Effect study looks at that moment: which visual properties interrupt the flow of attention inside a social feed, and how consistently they do so across different people.
How the study is set up
Participants scroll a simulated feed on their own phones and computers. Test visuals are mixed between realistic filler posts, so nobody knows which posts are being tested. There are no instructions and no highlighted targets; the feed behaves like a real one. Behaviour is recorded as it happens: where the scroll slows, where it stops, how long each visual stays in view and what is tapped. In sessions with webcam calibration, gaze samples are added on top of the scroll data.
What is measured
Stopping rate: how often a visual produced a stop rather than a pass
Time to first stop: how quickly the stop happened after the visual entered the screen
Dwell time: how long the visual held attention once the scroll stopped
Content transition speed: how fast the scroll resumed afterwards
Relative attention performance: all of the above compared with the other visuals in the same feed
How to read the findings
Stopping power is not a single property. In the sessions analysed so far, visuals that stopped the scroll tended to combine a clear figure against a quiet ground, one dominant element rather than several competing ones, and contrast that still reads at feed size. Visuals that were passed were rarely poorly made; they were busy, evenly weighted or indistinguishable from the posts around them. The study treats these as patterns to keep testing, not as rules.
Why it matters before launch
Performance is usually measured after publishing. The Stopper Effect framework moves part of that question earlier: a visual can be scrolled past in a controlled feed before it has to compete in a real one. The full report describes the feed design, the measures and how they feed into the Stopper Score.