The China plan you wrote three years ago is already out of date

None of the three exactly, which is why the category confuses people.

The model YASO operates is closest to an extended China team: e-commerce operations, store setup and merchandising, pricing and promotions, supply chain and payments, content and creator partnerships, customer service, offline retail relationships, and reporting.

The important structural distinction is that we are not a distributor. We do not buy your stock. We work on a consignment model, taking a monthly or quarterly order and paying the brand monthly. That matters more than it sounds, because stock ownership determines who controls your pricing, and pricing determines whether your brand positioning survives.

Commercial models vary widely across the market, so the useful thing is to know which questions to ask rather than which number to expect.

Ours has three components:

• A revenue share on China net sales, calculated after tax, returns and cancellations, which covers the operational work rather than charging fees for it.
• A small monthly software fee for dashboard and AI discoverability tool access.
• A growth budget committed by the brand, which funds marketing activity directly, with no management fee or retainer taken out of it.

What to probe with any partner: is the operational work charged as fees or aligned to sales, is the revenue share calculated on gross or net, and does your marketing budget get skimmed before it reaches the market.

There are two routes into China, and the difference between them shapes everything else.

Doing either alone from the UK is harder than it looks, largely because of language and the need for someone local to file. Cross-border is manageable. General trade is where most brands find they need help.

The sequencing point from part one applies here too: launch cross-border, learn what sells, then spend general trade registration money only on the products that have already proved themselves.

For cross-border, compliance comes down mainly to ingredients rather than paperwork. You are not changing labels or formulations unless you contain something highly restricted, CBD being the clearest example.

One practical addition that brands rarely think about: packaging durability. Bottles that shatter easily will cost you stock in transit within China. It is worth reviewing your packaging format for the logistics, not just for the shelf.

This one is recent and catches people out. Following a scandal involving Chinese-owned brands presenting themselves as foreign, platform regulation on health supplements tightened considerably.

Supplement brands now need a certificate of origin for each SKU they want to launch, including on cross-border channels. If you are bringing a supplement range into China, budget time for that documentation rather than assuming the standard eight-week cross-border timeline applies.

Less than you would expect. The meaningful split is not by category but by function.

Discovery platforms, RedNote and Douyin, are stronger for awareness, credibility and peer review. They are visual, they scale, and they are where new brands get found. The margin is thinner, because creators and livestreamers take commission.

Traditional platforms, Tmall and JD.com, have lost share but remain where repeat purchase, healthier margin and long-term business live.

The answer for almost every brand is both, in sequence: build demand where discovery happens, harvest it where the margin is.

It depends on positioning, and the tiers are fairly distinct.

• Premium: Sephora, SKP, and high-end department stores such as Kerry Centre in Shanghai.
• Mass and personal care: Watsons and Mannings. Faith in Nature, one of our brands, went into Mannings on that logic.

Trademark strategy needs a trademark lawyer, and we would rather say so than guess.

What we can say from experience is that timing is the part brands get wrong. We have had to walk away from a solid UK brand because someone had already squatted their name and logo in China. Squatters target brands that are visibly growing, and then charge for the return of their own identity.

Register as early as you can afford to, well before you launch.

Expect a period of not breaking even. For most brands the first six to twelve months is a growth phase, and the length depends on what you are optimising for.

Two reference points are useful here. Perfect Diary, during its rise, ran a marketing spend ratio above 50 per cent of revenue, effectively buying market share as fast as it could. L'Oreal, decades into the market, runs closer to 25 to 30 per cent across its portfolio. That range tells you roughly what competing at different ambitions costs.

If profitability matters more to you than share, there are levers: diversify platforms, diversify your SKU mix, measure each campaign against return rather than reach, and keep early activity sales-focused before spending heavily on brand building.

What matters most is agreeing with your partner, in advance, which of those two games you are playing.

Different in shape, not in difficulty.

In Europe, rising acquisition costs are pushing brands back towards omni-channel retail because pure e-commerce is hard to sustain profitably. That escape route is narrower in China, because digital penetration is so high. More than 80 per cent of Chinese consumers buy beauty products online, so online is not a channel choice, it is the market.

China's equivalent of rising CAC is creator and livestreamer economics. The large livestreamers take a significant commission and thin out your take-home.

The way through is to convert rented attention into owned assets:

• When a creator posts about your product, remake it in your brand's format and publish it on your own channels.
• When a livestreamer features you, capture high-quality footage and reuse it.
• Over time, move towards running your own livestream studio, where the margin is meaningfully better.

It is a two-to-three year build rather than a first-year fix, but the brands that skip it never stop paying rent on their own audience.