E152: Peeking Behind TikTok’s Curtain - What a Data Reporter Learned from 15M Videos

In this episode, Kevin talks with Jeremy Merrill of The Washington Post about an ambitious investigation into how TikTok’s recommendation algorithm really works. Jeremy explains how he and his colleagues analyzed data from more than 1,000 volunteers and 15 million videos to better understand personalization, stickiness, and scale on the platform. They discuss what the data reveals about how feeds are shaped, why some content is harder to escape than others, and what this kind of reporting can—and can’t—tell us about influence, attention, and transparency in algorithm-driven media.

Read, "We teamed up with our audience to measure TikTok’s unmistakable draw..."
View this story's interactive map, "Are you in TikTok’s cat niche? What 121,000 videos reveal."
Subscribe to The Washington Post