Why brands fail in China is the question top management asks once the money is gone. During the whole process, nobody sees the failure coming. It is only after the company has given up that the board asks what went wrong. I have sat in those meetings. The answer is almost never the one people expect.
Ten years ago the diagnosis was easy: they did not understand WeChat, they used a .com instead of a .cn, they translated their slogan with Google. That era is over. The brands failing in 2026 know what Douyin is. They fail for other reasons, and those reasons are more expensive to fix.
Written by Olivier Verot, founder and CEO of GMA. I have been based in Shanghai since 2012 and I have run the China entry of more than 500 foreign brands, including the ones that pulled out. Most of this article comes from post-mortems I did with clients after the budget was spent.
What Changed in 2026
Three things moved at the same time, and together they killed the old playbook.
Domestic brands caught up on perceived quality. This is the big one. Chinese consumers no longer assume that a foreign label means a better product. In fast-moving consumer goods, domestic brands now hold roughly 76% of the market. Homegrown carmakers took 58% of passenger vehicle sales in 2024, close to double their share five years earlier. The guochao movement started as a fashion statement around 2018. It is now a purchase reflex across skincare, electronics, sportswear and packaged food. A Western brand that used to win by default now has to win on argument.
Consumers turned cautious. They still spend, but they justify every purchase. Price transparency is total: a shopper compares your Tmall price, your Douyin price, your daigou price and your duty-free price in about forty seconds. A premium position that is not backed by a visible reason (an ingredient, a certification, a factory, a story that survives fact-checking) does not sell any more. It just sits in the warehouse. Look at what happened in television sets: the average retail price of a 100-inch screen fell from 19,713 RMB in 2022 to 12,941 RMB in the first half of 2025. Foreign brands who tried to hold their margin simply lost the shelf.
Discovery moved to AI and to social search. China had 1.125 billion internet users at the end of 2025, an 80.1% penetration rate, according to the 57th CNNIC report published in February 2026. That number stopped being interesting a while ago. What matters is where those people ask their questions. By March 2026, AI native apps in China reached 440 million monthly active users. Doubao alone had 345 million, Qwen 166 million, DeepSeek 127 million, according to QuestMobile’s Q1 2026 report. Add Xiaohongshu, at 242 million domestic monthly users and growing 13.2% year on year, where a large share of sessions start with a search box rather than a feed. If your brand is invisible in those answers, you are invisible at the exact moment the decision is made.
The exits of 2025 and 2026 tell the story
These are documented departures, not rumours.
- Skoda announced on 26 March 2026 that it will leave the Chinese market by mid-2026. It delivered 341,000 cars in China in 2018, its best year, and 15,000 in 2025. That is a fall of more than 95%, down to under 0.1% market share. Source.
- Samsung announced on 6 May 2026 that it stops selling all home appliances in mainland China: televisions, monitors, air conditioners, fridges, washing machines. Only the phone business stays. It had been selling appliances there since 1992. By 2026 its offline share was down to 3.62% in televisions, 0.41% in fridges and 0.38% in washing machines, and the division posted its first loss in fiscal 2025. Source.
- Mannings closed all its mainland stores and its online business at the end of 2025, after 21 years.
- Galeries Lafayette shut its Beijing flagship on 27 May, after more than a decade.
- Triumph, the German lingerie brand, is closing all its China stores after more than thirty years. Etam pulled out. Wacoal cut close to 10% of its stores.
- Before them, Panasonic, Sony, Toshiba, Sharp, Siemens and Whirlpool had already sold off or shrunk their Chinese appliance business.
Notice the pattern. None of these companies failed because they did not know how to open a WeChat account. They failed on product fit, on pricing power, and on the speed of local competition.
The Five Mistakes That Kill Foreign Brands in China Today
I see the same five, in the same order, almost every month.
1. Chronic underinvestment: a test budget on a market that needs 18 months
This is the number one cause, by a wide margin. A European mid-size company decides to “test China” with 40,000 or 60,000 euros over six months. That budget buys you a Tmall flagship you cannot feed, three KOL posts nobody remembers, and a Baidu account with no content behind it. At month six the numbers are bad, so the board cuts the line. The brand concludes that China does not work.
China does not fail slowly, it just takes time to start. Realistic timeline: three to six months to build content mass and get indexed, six to twelve months before organic search and Xiaohongshu content produce steady inbound, twelve to eighteen months before repeat purchase makes the unit economics work. Any plan shorter than that is a plan to lose money with certainty. If you cannot fund eighteen months, do not start. Spend the money on a market where six months is enough.
2. The trademark that belongs to somebody else
China is first-to-file. Not first-to-use. Whoever files at CNIPA first owns the mark, no matter how famous you are at home, no matter how long you have been selling. I have watched brands discover on the eve of their launch that a squatter, often a former distributor or an agent they met at a trade show, registered their name three years earlier. At that point you have three options: buy it back at a price the squatter sets, fight for two to four years, or rebrand for China. All three are worse than filing early.
It has also become harder to file. CNIPA tightened examination through 2025, and overall rejection rates climbed to around 40%, close to double the level of three years before, as the EU IP Helpdesk documented in January 2026. Examiners now read foreign-language marks properly and refuse descriptive ones. Which means: file in the right classes, file the Chinese-character version too, and file before your first trade show, not after. Our guide on how to protect your brand in China covers the classes that actually matter.
3. Shipping the same product you sell at home
Format, texture, sweetness, size, packaging, claim. Chinese consumers are not a subset of your existing customers with different eyes. A few concrete cases from our client base: a food brand whose 500g family pack did not sell because urban households buy small and often; a skincare brand whose rich cream was rejected in summer because the finish felt heavy in humid coastal cities; a wine importer whose label carried no Chinese story and no serving occasion, so nobody knew when to drink it.
Adapting is not betraying the brand. It is answering a question the Chinese buyer is actually asking. The brands who win test three formats in one city before rolling out nationally.
4. Refusing to hand control to a local team
Headquarters wants to validate every post. Legal in Paris or Munich reviews Xiaohongshu copy. Approval takes eleven days. In China, a trending topic lasts four. By the time the post is approved it is dead content, and the algorithm treats it as such.
The brands that grow here give their China team a real budget, a pre-approved tone of voice, and the right to publish without asking. Headquarters keeps veto on brand codes and reviews monthly, not daily. If you cannot accept that, you will always be one cycle behind the domestic competitor who decides in the morning and publishes at noon.
5. Believing cross-border e-commerce builds a brand
Cross-border is a good way to test demand. Import and export volume in cross-border e-commerce reached 2.75 trillion RMB in 2025, so the pipe works. But cross-border is a sales channel, not a brand strategy. On a cross-border store you get a transaction, and often you do not even get the customer data. You have no offline presence, no local WeChat relationship, limited advertising options, longer delivery, and a permanent price ceiling because everyone knows a cross-border price is comparable with the overseas price.
Brands who stay on cross-border for three years plateau at the same revenue and cannot explain why. The reason is simple: nobody in China knows who they are. Cross-border validates the product. Building the brand still requires content, search visibility, social proof and a local entity when the volume justifies it.
What Western Brands Are Doing Wrong: The Classic Mistakes That Never Went Away
From luxury fashion labels and fast-food chains to tech giants and carmakers, plenty of Western brands entered China with high hopes. Succeeding here is not simple. The market is varied and it moves faster than most head offices. You cannot reuse the strategy that works elsewhere. You adapt to what Chinese consumers want and how they think, or you leave.
As you can see above, many American giants eventually gave up, often after years of struggling. Here are the mistakes that keep repeating.
A bad marketing or branding campaign turns into a public mess
There were times when products ruled and manufacturing was king. Then the market moved, and branding started selling the product. Good branding can carry an average product a long way. Bad branding sinks a good one in 48 hours.
In China, poor understanding of consumers and of how they shop has ended a lot of international ambitions. The country still offers real expansion for retailers. Yet the list of Western brands that fell short of the Chinese consumer keeps getting longer.
Being inauthentic
A basic reason brands fail is a lack of authenticity and transparency. Without real insight into what customers need, and with attention on the wrong things, a business loses trust fast. Chinese consumers check. They read the reviews under your product page, they search your brand name on Xiaohongshu before buying, and they compare what you claim with what other buyers say.
Without focus, a business loses its edge. It is hard to run a strong strategy on a startup budget. What makes companies succeed in China is their ability to pivot quickly, and a lack of focus removes the ability to make the necessary adjustments.
Be honest, be transparent, and find ways to give more rather than take. Rare in business, and necessary if you want to still be here in five years.
Breaking your brand promise
When marketing is built to cover reality, the brand does not last. One rule stands above the others in branding: make a promise to your customers and keep it. Not everything has to be told. But the product has to match the promise, or the fall is brutal.
Deception no longer works. Chinese consumers are well informed, competition is dense, and they switch brands without hesitation. A broken promise costs brand equity and damages the brand image for years. Screenshots live forever on Weibo.
Failure to connect with the target audience
Every business says the customer comes first. Few act that way. Businesses that fail lose touch with their Chinese customers.
Your brand exists in the minds of your Chinese customers. To manage it, you keep your attention on their needs, which change fast. The minute a brand disconnects from those needs, it has failed, even if the sales figures take another year to show it.
Watch the values your Chinese customers are moving towards. Ask for feedback. Do they still love the product? What new features do they ask for? Are you reading the comments? I once spoke to the CEO of a training company who told me they never replied to negative reviews because those were unimportant. That company is no longer in China.
Failure to optimize conversions
Traffic without conversion is a subsidy to the platform. Most companies spend everything on reach and nothing on what happens after the click. Fix conversion early, so there is a positive return on any ad spend. Then you have a business that can survive its own growth.
You cannot rely only on organic traffic. Even niche brands need income-producing, conversion-optimizing activities while they build an audience. Without that, it is only a matter of time before the money runs out and management scrambles to keep the doors open.
Telling weak brand stories and delivering no real value
Your Chinese customers spend their day on their phone under a constant flow of marketing. If you want your story heard, it has to be meaningful and specific. Brands that fail are brands that never told a memorable story.
At the centre of any business is value. The companies that win deliver the most of it. Under-promise, over-deliver. And define your customer in numbers, not in adjectives.
Dolce and Gabbana: the case every marketing team should still study
In November 2018, the luxury label released a campaign full of ethnic stereotypes. The ad showed a Chinese model trying and failing to eat Italian dishes with chopsticks. People were angry at the depiction of Chinese culture as unrefined.
Then it got worse. Diet Prada, a fashion account on Instagram, posted screenshots of a private exchange attributed to founder Stefano Gabbana with insulting comments about China. The post went viral, the backlash was immediate, and the hashtag #BoycottDolce trended on Weibo.
The founders apologised, but the Shanghai runway show was cancelled, at a cost of millions. Eight years later, the brand is still paying for it in Chinese search results. That is the real lesson: in China, a reputation incident does not fade, it gets archived and resurfaced by every AI assistant that reads the web.
Having a chaotic brand architecture
Parent brands and sub-brands are not a problem in themselves. The problem starts when the hierarchy stops making sense to anyone outside the company. Brand architecture should make cross-promotion between your sub-brands easy, not confusing.
Most companies also lack experienced leadership on the ground. Without real China experience, newcomers drown in the volume of decisions. When problems arrive, and they do, there is nobody in the room who has seen the same problem before. Build a small board of advisors who have operated here.
Unable to compete against market leaders
Staying afloat is much harder when the competition is dense and local players have you in their sights. If a foreign company cannot compete head-on with domestic brands, it has to find a defensible niche. That takes judgement and nerve.
Consistency is a quality of strong brands. But consistency often turns into complacency, then into stagnation. The Chinese market waits for nobody. Stay quick and stay current. Avoiding failure means avoiding irrelevance.
Most companies also build products without involving their customers. Chinese brands do the opposite: they launch, read the comments, and ship a revised version in eight weeks. If you are serious about China, build with Chinese customers in the room.
Client Case: Ingrid, a Norwegian Brand That Almost Left
Ingrid runs a Norwegian marine collagen and omega-3 supplement brand, family-owned, around 18 million euros of turnover in Europe. She came to us in 2024 after two years in China.
The situation. She had a cross-border Tmall Global store and had spent about 220,000 euros over 24 months. Result: 340,000 euros of cumulative sales, a customer acquisition cost of 61 euros for an average basket of 74 euros, and a repeat purchase rate of 9%. The board had voted to close the store at the end of the year.
What she had tried, and why it failed. Three rounds of KOL campaigns with mid-tier beauty influencers on Xiaohongshu, roughly 90,000 euros. Good views, almost no conversion, because the posts said “Norwegian marine collagen” and nothing else. Chinese buyers of collagen already knew Japanese and domestic brands with published clinical data and clear dosage. Ingrid’s product had all of that in Norwegian regulatory files, and none of it in Chinese. She had also refused to change the packaging, a 60-sachet box, when the local habit is a 30-day trial format.
What worked, and why. We did three things over eleven months.
- We rebuilt the proof, in Chinese. Translated and published her Norwegian purity certifications, the fishing zone, the heavy-metal test results and the absorption study. Not as a PDF nobody opens, but as 40 short Xiaohongshu notes and 12 long articles answering the questions people actually type: is marine collagen better than bovine, how many milligrams per day, what does the smell mean. This gave the KOC posts something to say beyond “it is Norwegian”.
- We switched from big KOLs to 180 KOCs on affiliate terms. Small accounts, 2,000 to 30,000 followers, paid mostly on commission with a free product. The mechanism matters: Xiaohongshu’s search surfaces posts that already have saves and comments, so 180 modest posts that get real engagement outrank three sponsored posts that get views and nothing else. Cost per acquisition fell because we stopped paying for reach we could not convert.
- We built a private domain on WeChat. Every buyer got a QR code in the parcel leading to a WeCom advisor and a small group. The advisor sends the reorder reminder at day 25, when the box is running low. That single message is why repeat purchase moved.
She also finally made the 30-sachet trial box.
Result after eleven months. Acquisition cost down from 61 to 27 euros. Repeat purchase rate from 9% to 34%. Monthly revenue from about 14,000 euros to 71,000 euros. Not a miracle, and not fast. It cost her another 130,000 euros before it turned. The point is that the money she had already spent was not wasted on the wrong market, it was spent on the wrong sequence.
How to Avoid the Same Fate: What Works in 2026
Online and offline still work together. Trade shows, distributor networks and in-person meetings still open doors, especially in B2B. But the research now happens online before anyone shakes your hand. Here is where to put the effort.
Search on Baidu, and now inside AI assistants
SEO on Baidu remains the cheapest long-term source of qualified leads. It is not a one-time job. You publish useful Chinese content on a Chinese-hosted site, consistently, and it compounds.
What is new is generative engine optimisation. When a buyer asks Doubao or DeepSeek “which foreign collagen brands are reliable”, the model builds its answer from what it has read: Baidu pages, Xiaohongshu notes, Zhihu threads, industry media, Baike entries. It does not read your English website. The mechanism is straightforward and the work is concrete: write pages that answer full questions in Chinese, get your brand named in third-party content, keep your facts consistent everywhere so the model does not find contradictions, and make sure a Baidu Baike or Zhihu presence exists. WeChat’s integration of DeepSeek made this urgent for anyone selling in China, and we explained the consequences in our piece on WeChat and DeepSeek for brand marketing.
Xiaohongshu as a search engine
Stop treating Xiaohongshu as a social feed where you post pretty pictures. A large share of its 242 million monthly users open it to search before buying. Your content strategy should therefore be built on queries, not on campaigns: one note per real question, with the answer in the first two lines and the keyword in the title. Volume of honest notes beats budget on a single sponsored post. And read the comments, they are the cheapest market study you will ever get.
Douyin and interest-based commerce
Douyin does not wait for demand, it creates it. The algorithm pushes a product video to people who never searched for the category, and the purchase happens inside the app in seconds. That changes what you produce: not brand films, but short demonstrations that show the problem in the first two seconds and the product solving it before the fifth. Live selling then handles the objections in real time. A growing number of brands are shifting budget from Tmall to Douyin for exactly this reason, as we detailed in our analysis of brands leaving Tmall for Douyin.
KOC and affiliation instead of one big KOL
One celebrity post costs a lot and disappears in a day. Two hundred small accounts posting genuine reviews create a body of evidence that stays searchable for years and feeds the AI answers mentioned above. Pay on commission where you can. The brands with the best China economics I have seen run 70% of their influence budget on performance terms.
Private domain on WeChat
Platform traffic is rented. WeChat contacts are owned. Move buyers into WeCom groups and one-to-one advisor chats, then use AI-assisted service to answer the repetitive questions at any hour and hand over to a human for the rest. Cost per repeat sale in a private domain is a fraction of the cost of acquiring a new customer on Tmall. This is where margin comes back.
Rich content and brand value
Published content protects your expertise and keeps brand awareness alive between campaigns. Chinese B2B buyers read before they contact anyone. Baidu rewards fresh Chinese-language content on your own site, and so do the AI models. Good content on your own domain is still the best marketing asset you can own in China.
Social media, carefully
Chinese social media is part of daily life. People open WeChat, Weibo, Douyin and Xiaohongshu dozens of times a day. Brands know it and are all there.
Used well, social media builds awareness and brings customers. Used carelessly, it produces the next screenshot on Weibo. Before publishing anything with cultural content, have it reviewed by Chinese colleagues who are allowed to say no. If a campaign makes you hesitate, ask two more people. That reflex costs nothing and has saved several of our clients from a very expensive week.
FAQ
What is the single most common reason Western brands fail in China in 2026?
Underinvestment. Not ignorance of the platforms, not cultural mistakes. Companies allocate a six-month test budget to a market that needs eighteen months to produce compounding results, then read the early numbers as proof that China does not work. The brands that succeed decide upfront how much they can lose for a year and a half, and stop revisiting the decision every quarter.
How much budget do I need to enter China realistically?
For a consumer brand testing one category seriously, plan 150,000 to 300,000 euros over eighteen months, covering trademark, a Chinese site, content, KOC seeding and one commercial channel. Below 100,000 euros you can validate demand with a small cross-border listing and content, nothing more. In B2B the entry cost is lower, often 60,000 to 120,000 euros a year, because the volume of leads needed is smaller.
Should I register my trademark before or after testing the market?
Before. Always before. China is first-to-file, so the first party to register at CNIPA owns the mark regardless of your history elsewhere. Registering costs a few hundred euros per class. Recovering a squatted mark costs tens of thousands and two to four years, when it is possible at all. File the Latin name, the Chinese-character name, and the core classes, before your first trade show or distributor meeting.
Is cross-border e-commerce enough to build a brand in China?
No. It is enough to test whether people want your product, and that is genuinely useful. But cross-border limits you: little customer data, no offline presence, restricted advertising, longer shipping, and a price that everyone can benchmark against your home market. Brands that stay purely cross-border usually plateau. Once volume is proven, a local entity or a general trade route opens up the rest.
How do Chinese domestic brands beat foreign ones now?
Speed and price. They launch a product, read consumer comments for two weeks, ship a revised version, and repeat. Their marketing decisions happen in hours. On top of that, quality perception has closed: domestic brands hold about 76% of the FMCG market and 58% of passenger car sales. Foreign brands who still price on the old prestige premium, without a visible reason for it, lose the comparison instantly.
What is GEO and do I really need it?
Generative engine optimisation means making sure AI assistants describe your brand correctly when a consumer asks them for a recommendation. In China this matters because AI native apps reached 440 million monthly users by March 2026. The work is practical: Chinese pages that answer complete questions, mentions in third-party content, consistent facts across sources, and a presence on Zhihu and Baidu Baike. If the model has never read about you, it will recommend a competitor.
How long before I see results?
First qualified inbound leads usually arrive between month four and month six if content production is steady. Search visibility on Baidu builds from month six. Repeat purchase, which is where profitability lives, needs a full year of buyers going through their second and third order. Anyone promising a strong return in three months is selling you paid traffic, not a brand.
Do Not Repeat Their Mistakes: Let Us Talk
GMA is a China-based agency that has handled the market entry and recovery of foreign brands since 2012, from Shanghai. We are called in for two situations: brands about to enter who want to avoid the mistakes above, and brands already losing money here who need a diagnosis before the board decides to leave. We work on trademark, Chinese search, Xiaohongshu, Douyin, KOC and private domain, with our own team on the ground.
If your China numbers are not working and you want an honest read on why, send us your last twelve months. We will tell you if the problem is the market, the product, the sequence, or the budget. Sometimes the right answer is to stop. More often it is to change the order of operations. Contact us, let us have a chat.
