US Live Streaming is Hot, But Don't Copy Chinese Livestream Tactics

I saw a set of data: In 2025, short videos accounted for about 50% of TikTok Shop US GMV, mall and search rose to 36%, while livestreaming made up only 14%. On the other hand, Jude, who has been doing US livestreams for five years, says his main account maintains stable online viewership above 5k, averaging 20k orders per day, with livestream contributions exceeding short videos.

Putting these two sets of numbers together creates some conflict. Platform data and top-tier intuition look at different layers; overall market structure and single-account performance often don't align. My conclusion is that US livestream e-commerce is indeed heating up, but don't interpret it as a copy-paste of Douyin (TikTok China) livestream rooms. It's more like the content entry points are split: short videos seed interest, search and malls capture traffic, and livestreams provide the final push.

Last week when I analyzed the "no steering wheel" news, I wrote: break down facts first, then think about product portfolios. The key to US livestreaming is that the in-app transaction path has become complete. Previously, if a video went viral, users still had to jump to independent sites or Amazon; long jumps led to high drop-off rates. Later, TikTok Shop kept transactions in-app, allowing short videos, livestreams, malls, and search to feed data to each other. Jude benefits because he has an account matrix, supply chain, and host team. Ordinary sellers staring at "5k+ online" and "20k daily orders" and rushing into livestreams easily miscalculate costs.

Why is the share of US livestreaming still low? Chinese users have been trained by livestream rooms: cheap prices, limited-time offers, sales pitches, atmosphere—it works well. US users can watch too, but their shopping paths are more scattered. Many are used to searching, browsing malls, adding to cart, and comparing prices. Jewelry, dolls, and cosmetics suit short-video displays; they don't necessarily need hosts shouting constantly. My testing shows the difficulty lies in integrating English scripts, local hosts, inventory rhythm, and return expectations.

This makes me pay more attention to the role of AI tools. Over the past two months, I've fed US comments, livestream bullet chats, and return reasons into Doubao. Last month, I used Feishu tables for classification; this tool saved me the effort of filtering comments. Previously, for jewelry, I had to read many comments myself to judge color, size, logistics, and check if images were mistaken for AI-generated. Now I let the model cluster them first, then I pick high-risk issues. Actual conversion rate improvements aren't as mythical as claimed, but product selection judgment is steadier. Especially for cosmetics: skin tone, ingredients, allergy terms, compliance descriptions—if translation misses a word, the ad image looks good, but so does the return slip.

However, batch operations carry risks. I previously used a robotic arm to grab small items; if speed increased and buffering wasn't sufficient, the gripper would misalign the goods. AI copywriting is similar: generating dozens of titles and scripts at once looks efficient, but the trouble comes later—homogenization, exaggerated promises, infringement terms, and platform reviews all require human oversight. The entry point sounds lively, but execution-layer risk control is the cost center. Whether a livestream room retains viewers depends on whether the host can pitch, but also on whether inventory keeps up and logistics can cover it.

So I lean towards viewing US livestreaming as an enhancer for shelf-based commerce, not a standalone growth myth. Use short videos to test products, search terms to capture traffic, malls for repeat purchases, and livestreams to explain materials, clear stock, and build trust. Small and medium sellers shouldn't put all their budget into livestream rooms. First, get a product running with real clicks, favorites, and add-to-carts via short videos, then consider scaling up.

What's worth watching is whether, as the in-app transaction path becomes increasingly complete, the platform pushes more traffic from "watching for fun" to "deterministic transactions." This change will determine the weighting of the livestream entry point.