Three things China-based teams get wrong on TikTok

We have worked with a number of China-based teams starting out on TikTok. Creative capability is usually fine and budget is usually adequate, yet the first two months look bad. Reviewing them, the failures cluster in three places — and none of them are execution details. They are structural.

One: treating localization as translation

The common move is to take domestic creative that performs, swap the subtitles, redub the voiceover, and run it. Completion rate falls off a cliff.

Short-video pacing is cultural. Feed advertising in the Chinese market conditions you to open with strong conflict or a strong offer in the first three seconds. TikTok audiences have less tolerance for that rhythm and swipe away from anything that reads as an ad much faster.

Localization is about narrative, not language. What works is briefing local creators to reshoot around the same proposition, rather than translating the original. It costs more upfront and usually produces a lower cost per conversion.

Two: importing a domestic account structure

Many teams concentrate budget into a handful of ad groups and manage bids manually and frequently. On TikTok, that behaviour keeps knocking campaigns back into the learning phase.

The auction needs a sufficient volume of conversion signal before it stabilizes. Every budget or bid change resets that. What has worked for us:

1. Fund each ad group to support at least 20 expected conversions per day
2. Leave bids alone during learning unless something is clearly broken
3. Cap any single adjustment at 20%, and wait at least 24 hours between changes
4. Test with new ad groups rather than editing established ones

Three: ignoring per-market ad policy

The same creative that clears review in Southeast Asia can be rejected outright in North America. Beauty, health, finance, and weight management are especially sensitive, and EU markets layer data compliance on top.

Worse, rejection history affects account health. An account that repeatedly triggers policy flags will see delivery suppressed across all its campaigns — and the platform will not spell that out for you in the dashboard.

Separate accounts per market, and clear creative per market. Far cheaper than appealing after the fact.

One more fundamental point

For the first month abroad, do not set ROI as the objective. Set accumulating valid conversion signal as the objective.

While the model is unstable, any ROI figure is noise. Spend at a level you can absorb until the system has enough conversion data to bid consistently, then start optimizing. Most teams lose money in month one because an ugly ROI number makes them pull back early — which writes off every bit of learning they had already paid for.