81% of the Money Comes From Customers Who Already Bought

Every brief I receive about Xiaohongshu is an acquisition brief. Reach these people. Grow this account. Find new customers.

In April the platform published its own numbers on where merchants actually make money, and they point somewhere else entirely.

The news: what the merchant conference actually said

On 8 April 2026, Xiaohongshu held its GROW merchant conference in Shanghai. 雷峰网 covered it, and the figures behind the slogan are more useful than the slogan.

Eighty-one per cent. On a platform everybody buys as a discovery channel, four fifths of the money comes from people who had already bought once.

That single number reframes most of the Xiaohongshu budgets I see. If 81% of revenue comes from returning customers and your entire plan is built around reaching new ones, you are funding the smaller half of your own business.

NPL, and the letter everyone skips

The platform packaged its advice as NPL. Notes, people, livestream.

N, 笔记. Notes carry the product’s value and filter for the right audience. This is the part European brands already do, often well.

P, 人群. The platform’s own phrasing is that group chats turn followers into users you can reach and convert repeatedly. This is the part almost nobody outside China does, and it is the part that produces the 81%.

L, 直播. Livestream locks in high-intent buyers and closes the sale.

Most brands I work with run N and L and skip P entirely. They publish, they occasionally livestream, and between the two there is nothing. No group, no reason for a buyer to still be in contact three months later, no way to tell her the restock landed.

P is unglamorous work. It is a person answering messages in Chinese, every day, in a group of four hundred women who bought a coat. It does not photograph well in a board deck. It is where the repeat purchase comes from.

The platform has been building toward this for a while, including with the group buying feature launched this year, which only makes sense if you have a group in the first place.

Two product cases the platform put on stage

Xiaohongshu cited ECODAY, whose woven bags sold close to 80,000 units for more than 25 million RMB across the year, and Unicare, whose jacket recognised close to 6 million RMB in 2025.

Neither is a luxury house. Both are single products, sold repeatedly, to an audience the merchant talks to directly. A woven bag at roughly 300 RMB, sold eighty thousand times, is a better business on this platform than a beautiful thing sold two hundred times. That is uncomfortable for premium brands and it is worth sitting with rather than dismissing.

What I would change first

Measure the thing. Most brands genuinely cannot tell me their Chinese repeat rate. Before any new budget, find out. If it is under 15%, no amount of reach will fix the economics.

Own the customer or accept the ceiling. If a distributor holds the relationship, you are buying attention and renting the value. That is a legitimate choice. It is not a legitimate surprise.

Staff the middle. One Chinese-speaking person answering messages properly will do more for repeat purchase than a second KOL campaign. It is the least fashionable line in a budget and the highest return.

Give people a reason to come back on a calendar. Restocks, colour drops, small capsules. Repeat purchase needs an occasion, not just goodwill.

Why this only works now

None of this was possible when discovery and purchase lived in different apps. You cannot build a repeat-purchase business on a platform where the sale happens somewhere else. As Ecommerce China Agency put it in their piece on the closed loop: “Content and commerce finally sit together.” That is the precondition for everything above.

If you are building the content side of this, we set out the cadence in how to build a Xiaohongshu strategy for a fashion brand.

My honest reservation

These figures come from the platform, presented at the platform’s own conference, using merchants the platform selected. The 81% is almost certainly true of the merchants who succeed there. It tells us nothing about the ones who quietly closed.

And there is a real question about whether a 32% repeat rate is achievable for a house selling a 12,000 RMB coat, where two purchases a decade is normal. Possibly not. For accessories, knitwear, jewellery and beauty, I think it is, and those are the categories growing fastest on the platform anyway.

What I am confident about is the direction of the mistake. Nobody I work with is over-investing in their existing Chinese customers.

Do you know your repeat purchase rate in China? If you do, I would like to hear it. If you do not, that is the more interesting answer.