Influencer tests tank-top effect; shows how Instagram, X and YouTube algorithms work
Influencer Zara Dar tested the "tank-top effect" by posting two identical videos in different outfits. The tank-top version gained higher views on Instagram and X, but performed poorly on YouTube.
Zara Dar, a social media influencer, has gone viral for an experiment called the "tank-top effect.” Dar, who left her PhD programme to pursue content creation and OnlyFans, designed a test to study how different social media platforms respond to subtle visual changes in videos.
Her method was simple: she made two identical short clips. The only difference in the videos was her outfit. In one she wore a regular top, and in the other, a tank top. The results revealed surprising differences in how algorithms across platforms reacted.
On Instagram, the tank-top version drew 28 per cent more views. On X (formerly Twitter), it received more than double the engagement. But the same video performed worse on YouTube, which delivered fewer views than the normal top version.
Dar, who describes herself as an independent researcher and content creator, said the findings were unexpected and highlighted the opaque nature of platform algorithms.
Her "tank-top effect” experiment adds to the ongoing conversation about how algorithms shape visibility and engagement online. It suggests that there is no universal strategy for creators. While some platforms reward specific visuals, others appear to penalise them.
The test has also sparked debate among her followers about whether social media platforms consciously promote certain aesthetics, or whether it is simply a reflection of audience behaviour.
One user said, "Which version was uploaded first? As this is almost a duplicate content, the who was uploaded first is important too." Another wrote, "Thank you, youtube numbers say to keep my clothes on in my videos."
A third user added, "Coming from the marketing world, showing skin will get more views but the downside will attract the wrong followers." Yet another asked, "Would like to know the time difference between both the posts. And which one were posted first. Is it really a "tank-top effect" or viewer retention thing where people are spending more time on the tank top ones?"