← All articles

By David Shang5 min read

What China's live commerce taught me about selling on TikTok Shop in the US

Four lessons from projects I ran or advised on Douyin, China's TikTok, and how I would apply each one to TikTok Shop in the United States.

  • TikTok Shop
  • Live commerce
  • Social commerce

Short answer: Selling on TikTok Shop in the US uses the same building blocks I worked with on Douyin, China's TikTok: short video, livestreams, creators and paid traffic. The lessons I took from my own projects are to test creative fast and let the data choose, to treat a livestream as a system you can tune, to make paid and organic traffic work together, and to get product and supply right before you scale. The method carries over. The specific tactics have to be tested again with American shoppers.

I'm David Shang, founder of Shang & Partners. From 2020 to 2023 I worked in live commerce on Douyin, running my own stores and client accounts. In that time I built a top-5 live-commerce education channel on Douyin in three months, taught courses on platforms such as Juliangxue and Douchacha, and trained 500+ sellers. In 2025 I spoke at the TikTok Overseas Expansion Conference in Shenzhen.

Why look at China at all?

Douyin is one of the most competitive short-video commerce markets in the world. Sellers there have had to work out, under pressure, what makes a person buy from a video or a livestream.

And the US channel is growing. EMARKETER reported on December 9, 2025 that TikTok Shop US sales grew 108.0% in 2025 to $15.82 billion, 18.2% of US social commerce.

What results are these lessons based on?

These are projects I ran or advised on Douyin between 2020 and 2023. Amounts are in RMB; US dollar figures are approximate conversions.

  • Qian Songni (千颂妮), an apparel creator on Douyin: monthly sales grew from ¥20M to about ¥100M (roughly $2.8M to $14M).
  • Xinyu Outfit Diary (芯宇穿搭日记), a womenswear brand's store livestream: daily sales grew from ¥300K to ¥1M (roughly $42K to $140K).
  • BGI Nutrition (华大营养), a probiotics brand: daily sales grew from ¥20K to ¥200K (roughly $2.8K to $28K) at a 10× return on ad spend.
  • Victor Solution, my own loungewear brand, built from zero: ¥3M in sales in the launch month (roughly $420K).
  • DDCAT (叮当猫) pet supplies, built from zero: ¥2M in monthly sales (roughly $280K), from product selection to creator distribution.

I list them so you can see where the lessons come from. They are results from another market and another time, not a forecast for your business.

Lesson one: test creative fast, and let the data choose

The playbook I built on Douyin has two parts: fast creative testing and data-led scaling. In practice that means making many versions of a video or an opening line, putting a small budget behind each, and moving the budget to the ones that sell. Opinions about which version is best, mine included, lose to the numbers.

The same applies on TikTok Shop. Do not spend weeks perfecting one video. Make several rough versions that differ in one clear way, such as the opening seconds, the offer or the person on camera, and let real viewers decide. This is the step my team in Hangzhou now runs with AI, producing variants overnight US time so a test can start the next morning.

Lesson two: a livestream is a system you can tune

One case shows what I mean. The livestream room of Xinyu Outfit Diary grew from ¥200K to ¥800K a day (roughly $28K to $113K) through three changes:

  1. Reworking which items come back for an encore during the session.
  2. Adding product showcase videos for the livestream room.
  3. Adding real-time clips cut from the live broadcast.

Each of these is a part of the session that can be planned, measured and adjusted. In my experience that matters more than a charismatic host.

If you sell live in the US, write a run of show before you go on air. Decide which products open, which return and when. Prepare short videos that show the product in use. Cut clips while the stream is still running.

Lesson three: paid and organic traffic work together

One of the live classes I taught on Douchacha was titled "Using paid traffic to unlock organic reach". I still work that way: ads and content are not separate departments with separate budgets. Ads show quickly which products and videos people respond to. The content plan follows what the ads reveal. Strong content, in turn, makes the ad budget go further.

Return on ad spend, the measure quoted for BGI Nutrition above, is the number I watch while scaling.

Lesson four: product and supply come first

Two of the projects above started from zero, and on DDCAT the work ran from product selection to creator distribution. Later I set up a pet-food factory from zero: site, equipment, product, production and sales.

What I took from that is a plain rule: traffic exposes a product, it does not fix it. If the product is ordinary, the offer is hard to explain or stock runs out when a video takes off, more views only make the problem more expensive. Before you scale on TikTok Shop, check that you can fulfil a sudden jump in orders, that the offer can be explained on camera in one sentence, and that creators who sell for you have samples and clear selling points.

What does not carry over to the US?

I would not copy Douyin tactics line by line. Shoppers, platform rules, logistics and what counts as a fair claim in an ad are all different, and I expect the right session length, prices and selling style to differ too. So I keep the method and retest the tactics: start small, measure, and scale only what the data supports.

This is the work my team does for American businesses: AI creative, AI media buying and managed operations. If you sell a product and want to see how this playbook fits it, tell me what you sell and who buys it. I read every brief myself.