Xiaohongshu (RED): how a French brand really launches there

Launching on Xiaohongshu (RED, RedNote) in 2026 means choosing between three scenarios: reach the Chinese already present in France, sell cross-border into China, or give up. The entry fee is 600 RMB a year for business certification, non-refundable if refused; the real spend is producing Chinese-language notes designed for search, not media. Allow 3 to 6 months, 60 to 120 notes, a Chinese domestic guarantor if you sell, and a trademark registered in China before any account is opened. The rest of this article details the three paths, line by line. According to Rest of World in January 2025, the app was number one on Apple's App Store in the United States from 13 to 23 January 2025, and it had stayed top of Google Play since 14 January.

Xiaohongshu, RED, RedNote: what exactly are we talking about in 2026?

Xiaohongshu (小红书, “little red book”) is a mass-market search engine disguised as a social network, whose dominant format remains the image-and-text note with a vertical cover. RED is the international name, RedNote the one popularised in January 2025 by the arrival of American “TikTok refugees”. Same app, three names, and a misunderstanding that costs French brands dearly: treating RED as “the Chinese TikTok”.

The public figures are dizzying and contradict each other, which is information in itself. At the end of May 2026, via Huxiu and picked up by Daxue Consulting, the platform announced more than 400 million monthly active users and around 800 million searches a day. Yucheng IP Law, on 4 August 2026, cited 350 million MAU at the end of 2025 with more than 70% aged 18-34. The Qian Gua 2025 report, relayed by Fashion China, stayed at 300 million MAU, a male/female ratio of 3:7, and 35% Gen Z. And pages that still rank well on Google talk about 190 million. If your supplier comes out with 190 million, their file is out of date.

The figure that changes strategy is none of those. It is the search penetration rate: around 70% of monthly users open the app to search for something, according to Marketing Chine in July 2026. People do not scroll RED like a feed; they type “巴黎 买什么” (Paris, what to buy) before a trip. The direct consequence for production: a note that is badly titled and badly tagged does not exist, whatever the quality of the image.

On the business side, Marketing to China puts 2026 annual revenue at around 42 billion RMB, up 40% year on year, nearly 76% of it from advertising, and a valuation above $50 billion on track for a Hong Kong IPO. The platform is also pushing towards a male audience: the broadcast of the 2026 World Cup in China was announced on 28 May 2026. The image of “RED = beauty and urban women” is still true, but is starting to shift.

Which scenario does your brand fall under: tourists in France, cross-border, or neither?

There is not one RED launch but three, and confusing the first two is the fastest way to burn €40,000 without selling a single unit. Four questions are enough to decide. Where is your customer physically located: in Paris, Nice and Lyon, or in Shanghai? Is your brand registered in China? What is your average basket and your gross margin after bonded logistics? Do you have, in-house or with a partner, someone who writes RED's Chinese, not translated Chinese, platform Chinese?

For the vast majority of French brands we see, the honest answer is scenario A: reach the Chinese already in France. Tourists, students, the diaspora. No Chinese entity required, a logic of place notes and local search, a first usable signal within a few weeks. A restaurant in the 8th arrondissement, a leather goods shop on rue Saint-Honoré, an aesthetics clinic, a concept store in Cannes: their 90-day ROI is there, not in Hangzhou. On the student side, according to Campus France's figures for 2023-2024, China is the third country of origin of international students in France, and the number of Chinese students there rose by 6%.

Scenario C deserves bluntness: it does not exist commercially. Redshop, the platform's cross-border unit, started in June 2026 with 50 pilot merchants covering Hong Kong, Macao, the United States, the United Kingdom, Australia, Canada, Singapore and Malaysia, according to Jing Daily in April 2026. France is not among them. If your customer buys in euros and pays in France, the real choice lies elsewhere, on the side of French social commerce on TikTok Shop, not a Chinese app with no local payment funnel.

What does it take to open a Xiaohongshu business account without a Chinese company?

A French company can apply for a certified business account (the famous “Blue V”) without a Chinese entity: Kbis extract, the legal representative's passport, an authorisation letter if an agency files for you, and 600 RMB of certification fees valid for one year. Marketing Chine was still confirming this route in 2026 for French, Hong Kong and Taiwanese companies.

Two details kill half of all applications. The first: those 600 RMB are not refunded if the application is refused, and the certification must be renewed every year with a review of the file; the platform's official document, confirmed by TheRedScroll in September 2026, is unambiguous on this. The second: the account name must match a brand over which you hold enforceable rights in China. China applies first-to-file. If a third party registered your name in class 3 or 25 before you, your application fails, you lose the 600 RMB, and you discover that your brand belongs to someone else in the territory.

Recommended operational order: trademark filing in China, then certified translation of company documents, then account certification, then, only if you sell, shop opening. Not the other way round. “We'll open the account and see later” is the most common and most expensive mistake. The European Commission's China IP SME Helpdesk reminds us that China applies the first-to-file system and recommends putting your intellectual property protection strategy in place at least a year before doing business with the country.

How much does it cost to launch on Xiaohongshu, line by line?

Platform fees are low and well documented; the line item that explodes is editorial production in Chinese, note by note. Let's take them in order.

Fixed costs

• Business certification: 600 RMB per application per year (platform source, Sept. 2026).
• Overseas merchant shop: 600 RMB audit fee and a deposit from 20,000 RMB frozen, refundable on closure if there is no infringement, according to Aspire Hong Kong (2025-2026). Other sources, such as Marketing to China in 2026, give a deposit range of 1,000 to 50,000 RMB depending on category. Use the highest in your forecast.
• Merchant commissions: around 5 to 8% depending on category plus 0.5% in technical fees (Marketing to China, 2026); AppInChina mentions 3 to 5% for business shops, textiles at 3%, electronics at 5%. New merchants are given three months' exemption from commission.
• Trademark filing, certified translation, legal advice: the real budget of a full Chinese file, to be priced with a law firm, not with a media agency.

Media: Juguang, and the end of the 0.3 RMB myth

Juguang is the self-serve ad platform. Entry ticket announced in September 2025 by an agency relayed via Eastmoney: 10,000 RMB prepaid, plus the 600 RMB annual certification. CPCs sit between 0.8 and 3 RMB, higher in beauty and mother-and-baby, with CPMs of 10 to 30 RMB. For 2026, Marketing to China gives CPMs of around 20 RMB per 1,000 impressions on a basic banner, up to more than 100 RMB on themed pages and live promotion. Shenzhen Hometop, via Marketing Chine in August 2026, cites 1 to 5 RMB per click.

The “0.3 RMB per click” still found in a dozen French articles is the floor price of a search auction, not the real cost of a campaign. If someone presents that figure to you as a budget, change contact.

Production, the real line item

A credible brand on RED publishes. Not three times, sixty times. On our quotes, a complete RED note, visuals, cover, Chinese copy, keywords, compliance review, comes to between €60 and €350 depending on the scenario and the volume committed. Multiply by 80 notes over a quarter and you have your real bill. The tier logic is exactly the one described in our cost-per-creative grid by volume: below 20 units a month, the unit cost stays high because the editorial setup does not amortise.

Why your Reels are not RED notes, and what to produce instead?

Your TikTok and Reels library cannot be reused as is on RED, because the unit of purchase changes: you do not publish a video there, you publish an indexed note. A carousel of vertical images, a cover with Chinese text overlaid, a title of 18 to 20 characters containing the target query, a structured body text, hashtags chosen for search and not for trend. Video exists, it works, but it remains a minority of what users consume when they search.

The equivalent of the three-second hook, on RED, is the cover + title pair. Nothing else. If the cover does not say at a glance what the note promises, and if the title does not contain the words the person typed, the note does not come up in results. The ratio also follows a different logic: you work in 3:4 for the cover, not systematically 9:16, the same discipline as described in our article on choosing the ratio by placement, applied to a platform with its own rules.

The last three lines do not exist on a Reel. That is where the extra cost lies, and it is why a “we'll recycle what we have” ends in dead notes. What really gets recycled: product rushes, detail shots, usage sequences. What does not: the edit, the title, the voice, the promise.

On cadence, the order that works according to feedback from Hangzhou-based agencies relayed by Marketing Chine in 2026: RED notes and Douyin short videos first, live afterwards. Never live first. A live with no stock of indexed notes behind it is an audience that looks for your brand, finds nothing, and leaves.

How do you work with KOCs without getting fleeced?

The volume of small creators beats the headliner: according to the 21st Century Business Herald, picked up by Daxue Consulting in 2026, the platform sends about half of its traffic to accounts with fewer than 1,000 followers. The idea that you need “a big KOL to launch” is the most expensive in the market.

The orders of magnitude, as they circulate in 2025-2026. Via Pugongying, the official marketplace: around 100 to 800 RMB per image note for a creator with fewer than 5,000 followers, and 80,000 to 300,000 RMB above a million followers (Marketing to China). Long Advisory, cited by TheRedScroll in June 2026, gives $20 to $60 per post for nano-KOCs with 1,000 to 5,000 followers, and $10,000 to $35,000 for a structured KOL campaign. Small creators, according to the 21st Century Business Herald, run at around 500 RMB per image note and 600 RMB per video note.

Going through Pugongying costs a 10% platform commission on declared deals (TheRedScroll, 2026). Many brands try to get around it by paying off-platform. Bad idea: an undeclared commercial collaboration is exactly the reason a brand account gets suspended, and you lose traceability of your spending. The 10% is the price of insurance.

For the brief, the four locks we apply to French creators, frame, proof, constraint, deliverable, remain valid, with a fifth lock specific to RED: language and keywords. You cannot approve what you cannot read. Either you have a Chinese-reading reviewer in the loop, or you pay blind. And comparing €1,500 of Chinese KOC seeding with €1,500 of French micro-influence is a healthy exercise before committing: the second gives you assets reusable in paid, the first does not.

Which RED metrics matter, and which ones lie?

Views are the most misleading metric on the platform; saves, ranking on your target queries and the search volume for your brand name are the only ones that predict a sale. A user who saves a note to a collection is preparing a purchase or a trip. A user who scrolls is preparing nothing.

The minimum dashboard, for scenario A as for B:

• Ranking of your notes on 5 to 10 Chinese queries defined in advance, checked manually every week. This is tier 1.
• Save rate (收藏) relative to views, per note. A note with many views and few saves is a false positive.
• Search volume for your brand name in the app, month after month.
• Comments “哪里买” (where to buy) and requests for an address: this is the most direct intent signal on the platform.
• For scenario A: a question at reception or at the till, “how did you hear about us”. Rustic, and it is all you will get.

Attribution is the weak point, and we may as well say so frankly. From France, you have no usable pixel, no cross-platform tracking, no clean reconciliation between a note viewed in Chengdu and an order on your site. Our four-tier measurement protocol applies, provided you replace tier 1, retention, with search ranking. The other three tiers do not change.

We publish no ROAS range for RED. No properly sourceable data exists in 2026 for a French brand, and inventing a range would do more harm than saying nothing. The only public benchmarks are social commerce conversion rates of 3 to 8% announced by Hashmeta in 2025, not specific to France, and the fact that merchants in the Rise100 ranking multiplied their GMV by 2.6 in a year in 2025, with 81% of sales made to already-acquired customers (GROW 2026 merchant conference, via Yucheng IP Law).

That last figure deserves a pause: 81% of sales come from already-acquired customers. RED is not a cold recruitment machine. It is a channel for confirmation and retention. The brands that succeed there maintain a relationship; they do not buy reach.

What legal and operational risks should you weigh before signing?

Four concrete risks, in order of probability: loss of the trademark filing, logistics downgrade, account suspension, and value leaking to the grey market.

Trademark filing in China

First-to-file. Whoever files first holds. Every year French brands discover, when they come to open their shop, that their name has been registered by a former distributor or a professional squatter. Account certification fails, the 600 RMB is lost, and buying back the rights is negotiated at five figures. This item comes before the creative budget, always.

The bonded warehouse

According to AppInChina in 2026, direct shipping from abroad leads to a downgrade in search. It is not a moral penalty, it is the mechanical consequence of delivery times: the platform favours merchants able to keep a short delivery promise. The bonded warehouse (Ningbo, Zhengzhou, Qianhai) is therefore not a comfort option, it is a condition of existence in the results. Budget for it on a par with the commission.

The 9 June 2026 terms and data

The Paris law firm Hassan Kohen published in 2026 an analysis of RedNote's terms of 9 June 2026: who legally manages the account, what data flows to China, what transparency obligations weigh on creators working with brands, what recourse exists in case of a banned account. The firm notes a discrepancy between the Google Play listing, which states that no data is shared, and the app's privacy policy. On the Chinese side, the PIPL requires local storage of Chinese users' data and regulates cross-border transfers (Yucheng IP Law, August 2026). If you entrust the account to an agency, have it written in black and white who holds it and what happens at the end of the contract. It is the same clause as for recovering your video assets, with an additional stake: a lost certified account cannot be rebuilt. In May 2025, following an investigation into transfers of European users' data to China, the Irish Data Protection Commission fined TikTok a total of €530 million and gave it 6 months to bring its processing into compliance.

Daigou and the grey market

Your notes will generate demand before your shop is ready. That demand will be served by resellers, at the price they decide, with the photos they want. We have seen brands create their own unfair competition by launching content six months before logistics. Release the notes when you can take the money, or accept that you are feeding the grey market.

Where to start if you decide to go?

A serious Xiaohongshu launch starts with a 90-day test on a closed budget, on the one scenario your business can actually serve. For a brand with physical points of sale in France: 40 to 60 notes, 10 to 15 Chinese student KOCs based in Paris or Lyon, zero media, weekly ranking check on 8 queries. For a brand targeting mainland China: legal and logistics first, content afterwards, and not a euro of Juguang before you have 60 indexed notes.

The stop criterion, the one nobody writes: if your average basket is below €60 and your gross margin under 55%, the cross-border scenario does not pay for itself for two years once you add up the frozen deposit, customs, the 5 to 8% commission and the editorial cost. In that case, the right call is often to put that budget back into your monthly volume of creatives on your existing channels. We say it knowingly: it is advice that makes us sell fewer exotic projects and more useful production.

What we do not know yet, and what will change everything: if Redshop opens to continental Europe after its pilot phase of 50 merchants, scenario B will become accessible to much smaller brands. Nobody can date that opening today. In the meantime, the only thing that does not expire is the stock of indexed notes you will have built. They keep coming up in search three years later. It is the only place in social where that remains true.

Frequently asked questions

Xiaohongshu, RED, RedNote: are they the same thing?

Yes. Xiaohongshu (小红书, “little red book”) is the Chinese name, RED the international name, RedNote the one popularised in January 2025 by the mass arrival of American users. One app, a mass-market search engine dominated by image-and-text content, with about 800 million daily searches announced by the platform at the end of May 2026.

Can you open a Xiaohongshu business account without a Chinese company?

Yes for a certified business account. You need a Kbis extract, the legal representative's passport, an authorisation letter if an agency files for you, and 600 RMB of certification fees valid for one year, renewable. Selling cross-border, on the other hand, requires a heavier structure, with a joint-liability Chinese domestic guarantor according to Yucheng IP Law in August 2026.

How much does a RED launch cost in the first year?

Platform fees are low: 600 RMB for certification, plus 600 RMB for the audit and a frozen deposit from 20,000 RMB if you open a shop. The dominant item remains production: 60 to 120 Chinese-language notes over a quarter, between €60 and €350 each on our quotes depending on the scenario and the volume committed.

Do you need a big KOL to get started on Xiaohongshu?

No. According to the 21st Century Business Herald, picked up by Daxue Consulting in 2026, the platform sends about half of its traffic to accounts with fewer than 1,000 followers. Thirty KOCs at 300-800 RMB a note produce more search signal than one influencer at 150,000 RMB. Go through Pugongying and accept the 10% commission.

Does RedNote let you reach a French-speaking audience?

Not in a commercially usable way. Redshop, the cross-border unit, started in June 2026 with 50 pilot merchants covering Hong Kong, Macao, the United States, the United Kingdom, Australia, Canada, Singapore and Malaysia according to Jing Daily. France is not among them. For a customer who pays in euros, the real choice lies on French channels.

My RedNote account has been banned: what recourse do I have?

The 9 June 2026 terms, analysed by the Paris firm Hassan Kohen, set out the recourse, but it remains limited and handled in Chinese. The most frequent reasons on the brand side: undeclared commercial collaboration outside Pugongying, prohibited product claims, an account name not matching a brand registered in China.