Xiaohongshu and Fake Marketing Play a Game of Cat and Mouse

```

Author | Wang Xiaojuan

Editor | Huang Yu

"False marketing not only undermines the authentic community ecosystem of Xiaohongshu, but also disrupts fair commercial order."

At the "Crackdown on False Marketing" open governance day event held on September 17, the captain of Xiaohongshu's "Crackdown on False Marketing Squad," Lego, spoke these words.

At this event, Xiaohongshu's "Crackdown on False Marketing Squad" also made its official debut. At its first appearance, this special squad announced some astonishing figures: since its establishment in March this year, Xiaohongshu has banned over 12 million fake accounts, handled 13.76 million false marketing posts, cleared more than 360 million fake comments, and the exposure rate of false marketing content has dropped by 60% overall.

This battle over the community ecosystem reflects the growing pains Xiaohongshu faces as China's largest "grass-planting" (recommendation) platform.

In its 12 years since inception, Xiaohongshu has always been sought after by users for its community atmosphere of real and equal sharing. Now, it has developed into a comprehensive platform combining social interaction, content creation, and e-commerce, with over 300 million monthly active users. It has become the preferred platform for China's new generation of smartphone users to obtain travel reviews and lifestyle advice.

At the same time, as users have developed habits of searching Xiaohongshu before buying, many merchants have also set their sights on Xiaohongshu and conduct a series of false marketing activities, continuously impacting ordinary users' experience.

So-called false marketing refers to acts that disguise as real in order to achieve marketing goals. Its governance challenge lies in its profit-driven "strong concealment" and "strong confrontation." Lego explained, "False marketing usually shows features of crossing accounts, posts/comments, and multiple links. It’s difficult to eradicate just by letting individual teams manage their own separate areas."

Faced with this challenge, Xiaohongshu's "Crackdown on False Marketing Squad" has built five governance sectors: account, content, industry, brand, and search.

Specifically, account governance attacks false marketing at the source. The team captures "abnormal behaviors" of accounts and combines this with multiple dimensions such as content and associated violating brands for identification.

Bixing, head of account governance, introduced that the team captures "abnormal behaviors" of accounts, combines this with content and violating brands in multiple dimensions for identification, and has banned over 12 million fake accounts in six months, aiming to let no "drama queen" fake account slip by."

Content governance accurately identifies and clears fake posts and comments. Team leader Lijia explained that content governance especially targets posts with embedded ads and fake self Q&A comments, as these are most likely to mislead users into thinking they are real recommendations from ordinary users—this is the main focus of their crackdown.

With account and content governance in place, why does Xiaohongshu still need an industry governance team?

Industry governance team leader Xingzhuo explained, it's because in important consumption decision areas like studying abroad, real estate, marriage and relationships, parenting, and tourism, a single fake recommendation could cost users tens of thousands yuan more, or even affect their "major life events." The industry governance team must dig deep into industry inside practices, sometimes even disguising as clients to inquire about details, ultimately formulating strict entry standards and clear content norms, and has purged a large number of violating accounts.

The squad also clearly recognizes that no matter how fake marketing tactics change, it eventually exposes its brand. Brand governance leader Sanye explained that Xiaohongshu has established a "brand purification system": putting violating brands into a "blacklist," comprehensively scanning for false marketing behaviors, pinpointing the brands behind such operations, and applying graduated penalties to brands—"hitting faking brands as soon as they show up."

When the cost of a single false marketing attempt exceeds the long-term benefit to a brand, the motivation to fake is weakened.

Even after multiple layers of filtering, some brands still distribute mass fake review content, hiding within search result long-tail traffic. The team deeply analyzes fraud processes—from low-quality articles to gang distribution and profit behaviors—for full identification, shuts down exposure in search, and governs "escaped" fake content in search areas.

All these are realized with the help of technology.

In Xiaohongshu's false marketing crackdown governance system, the team has comprehensively upgraded large AI models for the "strong confrontation, complex scenarios" features of false marketing, tailoring for different governance scenarios. This "one scenario, one model" design greatly improves identification accuracy and efficiency, effectively avoiding situations where generic models fail to fit the scenes of false marketing well.

Lingyi, from the Xiaohongshu technical team, explained that false marketing posts often use "template writing." Fraud agencies sum up user-favored expressions and use them in titles, main text, cover images, videos, etc. The team keeps summarizing typical features of false marketing posts, continuously expanding the template library, and has upgraded its "template post identification large model" to batch identify and intercept homogenous content. In just two months, the violation exposure rate of template posts dropped by 82%.

Especially worth mentioning is the construction of the "Eco Sentinel" system and "Black Seed Database." "Eco Sentinel" systematically reviews new posts within a period; when a large amount of homogenous abnormal content appears, it can quickly identify it by algorithms.

The "Black Seed Database" has accumulated 180,000 violation samples; new content is compared against these samples in multiple modes, and if the similarity is high, it is called back for review, intercepting false marketing in real time and efficiently.

"Cracking down on false marketing is a long-term battle. 'Raising the cost of false marketing above the gains' is our long-term goal," said Lego. "We also hope more brands, creators, and users will join us in resisting false marketing and maintaining a true and friendly community atmosphere on Xiaohongshu."

As the governance efforts deepen, subtle changes have occurred on Xiaohongshu: the overall exposure of false marketing content has dropped by 60%, user negative feedback on search results has dropped by 20%, and user satisfaction with community content has increased by 10%.

In the long term, Xiaohongshu’s battle against false marketing is not just a platform governance issue, but also a deciding point for the future direction of content communities.

Risk Warning and DisclaimerThe market is risky and investment must be cautious. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are appropriate to their particular circumstances. Investing accordingly is at your own risk. ```