Why Your Posts Reach Nobody: How Xiaohongshu's Algorithm Decides What Gets Seen

Why Your Posts Reach Nobody: How Xiaohongshu's Algorithm Decides What Gets Seen

"I posted, but nobody's seeing it."

This is the wall most people hit first on Xiaohongshu. The question — "The content isn't bad, so why?" — starts to make sense once you understand how Xiaohongshu's algorithm works.

This article explains how Xiaohongshu evaluates posts and decides who sees them.

First: Xiaohongshu Doesn't Run on Follower Count

On most social platforms, more followers means more reach. That was true of Instagram too, once upon a time.

Xiaohongshu works differently.

Followers matter to some degree — but Xiaohongshu's algorithm fundamentally distributes traffic based on content quality, not audience size. A brand-new account with zero followers can reach tens of thousands of people if the content is strong. Conversely, an account with 10,000 followers will barely reach anyone if its latest post fails to generate a response.

This sounds fair — but since the algorithm decides what "quality" means, not knowing its criteria means you can spin your wheels indefinitely.

Every Post Enters a "Small Room" First

When you publish a new post on Xiaohongshu (called a "笔记" or "note"), it doesn't go out to all users at once.

What happens first is a test release into an initial traffic pool.

The platform shows your post to a very small group — typically in the range of 200 to 500 people. Think of it as a limited preview screening. Based on how those few people respond, the platform decides whether to distribute it further.

If the response is good, the post moves into a larger traffic pool. If the response there is also good, it moves to an even larger one. This cycle repeats, and certain posts eventually reach audiences of tens or hundreds of thousands. That's what it means to "go viral" on Xiaohongshu.

If the initial response is poor, distribution stops there. The post sits dormant, seen by almost no one.

The CES Score: Your Post's Report Card

So what does the platform measure to determine whether a response is "good"?

This is where the CES (Community Engagement Score) comes in — an internal metric that functions as a report card for your content.

The formula is publicly known and works roughly like this:

CES = Likes × 1 + Saves × 1 + Comments × 4 + Shares × 4 + Follows × 8

Looking at this formula, it becomes clear what Xiaohongshu prioritizes.

Follows earn the highest score at 8 points. Comments and shares come next at 4 points each.

A "like" is worth only 1 point.

In other words, Xiaohongshu values deep engagement far more than passive approval. "I want to keep seeing this person's content" (follow), "I want to share this with a friend" (share), "I have something to say about this" (comment) — these signals of genuine involvement are what the algorithm rewards.

Saves: A Special Signal

In the formula above, "saves" appear to be worth only 1 point — but in practice they carry special significance.

On Xiaohongshu, saving a post to refer back to later is extremely common behavior. Recipes, travel plans, skincare routines — users save content that has lasting practical value.

From the algorithm's perspective, a high save count signals: "this post has long-term value." Posts with many saves tend to experience a "long-tail effect" — being recommended again days or weeks after publication, generating stable, sustained traffic over time.

The First Two Hours Determine Everything After

After you post, the most critical window is the first two hours.

The CES data generated in this window largely determines whether the post advances to larger traffic pools. Some data suggests that the performance in the first 30 minutes determines approximately 60% of the post's eventual reach.

This means getting early engagement as quickly as possible after posting is important. Accounts with existing followers benefit from their audience providing this initial push — but for new accounts, this is the hardest part. This is precisely why the timing of posts and the strategy for generating early momentum is where experienced operators earn their keep.

Two Entrances to Traffic

There are two main routes through which a post can be seen on Xiaohongshu.

① Recommended traffic from the Discovery Feed (发现页)

This is the main timeline users see when they open the app — posts selected by the algorithm based on each user's interests. This feed accounts for over 60% of total traffic on the platform. The traffic pool promotion system described above feeds directly into this.

② Search traffic

Posts that appear when users actively search for keywords. As noted in our first article, Xiaohongshu users frequently use the platform to look things up — meaning search traffic comes from users with specific intent, often with high purchase motivation. This makes it particularly valuable for businesses.

By embedding keywords that your target audience is likely to search for into your title and body text, you can capture this search traffic on an ongoing basis. A well-designed post can continue pulling in new visitors month after month — even six months after it was published.

What the Algorithm Dislikes

Understanding CES optimization might tempt some to ask: "Can't we just manipulate the score?" Since 2025, Xiaohongshu's fraud detection has become significantly more sophisticated.

Specifically:

Saves far outnumber comments while comments are zero → flagged as unnatural

Likes arrive in sudden high volumes (like rate exceeding 50% within seconds) → flagged as purchased likes

AI-generated text used verbatim without at least 30% human editing → flagged and restricted

These behaviors result in "降权" — a direct deduction to the account's CES score. Once an account's standing is reduced, every subsequent post starts at a disadvantage.

The Philosophy of "Decentralization"

At the heart of Xiaohongshu's algorithm is a design philosophy called "去中心化" — decentralization.

This means the platform is deliberately designed to prevent traffic from concentrating in the hands of a few large accounts or top influencers. A brand-new account with no followers can receive massive traffic if its content is excellent. But a once-popular account that starts producing mediocre content won't be protected by its past success.

This philosophy makes content competition on Xiaohongshu unusually flat — and unusually demanding. Every post is a test. History offers little protection.

Understanding the System and Operating Within It Are Two Different Things

Having read this explanation, you may feel: "Got it. Now I understand how it works."

Directionally, yes. But effective operation requires building on this understanding with countless additional decisions: Which keywords to choose. How to design the thumbnail. How to generate early engagement after posting. What frequency works for your category.

The algorithm itself is also adjusted frequently — what worked six months ago may not work today. Between "knowing the rules" and "consistently producing results" lies months of observation and iteration.

In the next article, we'll look at the actual content that gets responses on Xiaohongshu — breaking down the anatomy of posts that go viral.