2026年,小红书完成了新一轮算法升级,最核心的变化是:CES(Community Engagement Score)评分公式中,评论和转发的权重大幅提升至点赞的4倍,关注行为的权重更是高达8倍。这意味着「深度互动」正式取代「浅层互动」成为推流的核心驱动力。与此同时,平台引入了双阶段流量池考核机制和三层账号权重体系,传统的「刷量」和「水评」策略不仅失效,还会直接导致账号降权。理解这些变化并调整运营策略,是2026年在小红书获得流量的必修课。In 2026, Xiaohongshu completed a major algorithm upgrade. The most critical change: in the CES (Community Engagement Score) formula, comment and share weights surged to 4x that of likes, while follow actions now carry 8x weight. This means "deep engagement" has officially replaced "surface interactions" as the core driver of content distribution. The platform also introduced a dual-stage traffic pool evaluation system and a three-tier account weight framework. Traditional tactics like fake engagement and low-quality comments now actively harm your account. Understanding and adapting to these changes is essential for gaining traffic on Xiaohongshu in 2026.
CES 评分新公式:评论和关注成为核心杠杆New CES Formula: Comments & Follows Become Core Levers
2026年最新的CES评分公式为:CES = 点赞×1 + 收藏×1 + 评论×4 + 转发×4 + 关注×8。从公式可以清楚看到,不同互动行为的权重差异巨大。点赞和收藏仅计1分,是「最便宜」的互动行为;评论和转发权重达到4分,代表用户「想参与」的深度互动;而关注行为权重高达8分,代表用户「想长期跟随」的最高价值行为。这套公式的底层逻辑是:平台更看重用户的真实参与深度,而非表面的数据繁荣。The 2026 CES formula is: CES = Likes×1 + Saves×1 + Comments×4 + Shares×4 + Follows×8. The weight differences are stark: likes and saves score just 1 point each—the "cheapest" interaction; comments and shares earn 4 points, representing genuine participation; follows carry the highest weight at 8 points, signaling long-term user commitment. The underlying logic: the platform now prioritizes authentic engagement depth over surface-level vanity metrics.
水评反噬:低质量互动不再有用,反而有害Low-Quality Comments Backfire: They're Now Harmful
2026年算法的另一个重大变化是对「水评」的严厉打击。无意义的评论(如「哈哈哈」「好看」「不错」「666」等)不仅不会被计入CES评分,还会直接拉低账号权重。平台通过NLP(自然语言处理)技术识别评论的真实度和信息含量,只有包含实质内容的评论才会被算法认可。此外,3小时内的高质量互动权重是24小时后互动的5倍。这意味着笔记发布后的黄金窗口期内,作者需要引导和回复的是有深度、有信息量的真实评论,而非简单的表情和应付式回复。Another major 2026 change is the crackdown on "water comments" (low-quality filler). Meaningless comments like "haha", "nice", "cool", or "666" no longer count toward CES—and they actively lower your account weight. The platform uses NLP to assess comment authenticity and information density; only substantive comments are recognized. Furthermore, high-quality interactions within 3 hours carry 5x the weight of those after 24 hours, making the golden post-publish window even more critical for meaningful engagement.
双阶段流量池:1小时初始池 + 3小时晋级池Dual-Stage Traffic Pools: 1-Hour Initial + 3-Hour Promotion
旧版赛马机制主打「前1小时定生死」,但2026年4月算法完成精细化升级,形成双阶段流量池考核机制。第一阶段为1小时初始池,笔记发布后系统推给200-500个冷启动用户,根据初始互动数据(主要看点击率和互动率)决定是否晋级。第二阶段为3小时晋级池,如果在3小时内高质量互动达标,笔记可以冲刺万级精准曝光。关键策略调整:发布后的前3小时是你必须保持在线、积极回复评论、引导深度讨论的核心窗口期。The old system was "first hour decides everything," but the April 2026 upgrade refined this into a dual-stage traffic pool. Stage 1 is the 1-hour initial pool: your post is shown to 200-500 cold-start users, and initial engagement data (CTR and interaction rate) determines advancement. Stage 2 is the 3-hour promotion pool: if high-quality interactions meet the threshold within 3 hours, the post can reach 10,000+ precision impressions. Key takeaway: the first 3 hours post-publish are your critical window for staying online, actively replying to comments, and driving deep discussions.
三层账号权重体系:基础权重决定流量上限Three-Tier Account Weight: Base Weight Sets Your Traffic Ceiling
平台新增三层基础权重体系,账号基础权重不达标,即便内容质量优秀也会被隐性限流。维护账号权重的核心要素包括:保持每周稳定更新2-3篇(断更超7天动态权重下降30%);坚持单号单赛道(垂直领域账号权重比泛领域账号高50%);杜绝搬运、违规内容;完成实名认证。对于做矩阵账号的品牌和运营者来说,宁做10个垂直小号,也不做1个全能大号。每个账号只聚焦一个细分场景和品类,才能获得最大化的推荐权重。The platform introduced a three-tier base weight system. If your account's base weight is subpar, even excellent content gets throttled. Key weight factors: publish 2-3 posts weekly consistently (going dark for 7+ days drops dynamic weight by 30%); stick to one niche per account (vertical accounts get 50% higher weight than generalist ones); avoid reposts and policy violations; complete real-name verification. For brands running multiple accounts, ten focused niche accounts outperform one catch-all account.
实操策略:如何在新算法下获得更多推荐Practical Strategies: Winning More Recommendations Under the New Algorithm
基于CES新公式和双阶段流量池,这里给出具体的实操建议:第一,在笔记结尾抛出场景化的精准问题(如「你在XXX的时候是怎么处理的?」),引导用户写出有内容的真实评论。第二,制作场景化干货合集,突出实用价值,激发用户收藏和转发需求。第三,在结尾输出长期价值承诺(如「关注我,每周更新XX系列干货」),提升精准关注权重。第四,发布后3小时内保持在线,快速回复每一条评论,对优质评论点赞并展开二次互动。第五,善用「语音问一问」等平台热点功能,贴合场景引导互动,获取平台额外加权。同等互动数量下,高质量真实互动的笔记排名比普通刷量内容高出10倍。Based on the new CES formula and dual-stage pools, here are actionable tactics: First, end posts with specific scenario questions ("How do you handle XXX?") to elicit substantive comments. Second, create practical resource compilations that drive saves and shares. Third, close with long-term value promises ("Follow me for weekly XX series") to boost follow weight. Fourth, stay online for 3 hours after posting—reply to every comment quickly, like quality comments, and spark follow-up discussions. Fifth, leverage platform features like "Voice Q&A" for extra algorithmic boost. Posts with authentic high-quality interactions rank 10x higher than those with inflated metrics.
搜索与推荐打通:搜索意图匹配更加关键Search-Recommendation Integration: Intent Matching Matters More
2026年小红书进一步深化了搜推互通机制,搜索流量与推荐流量不再是两个独立的系统。平台新增了「意图识别」模块,能区分「信息型搜索」(如「怎么去黑头」)和「交易型搜索」(如「黑头仪推荐」),并据此匹配不同类型的内容。长尾词的权重显著提升,如「油敏肌夏季水乳」比泛词「水乳」更容易获得曝光。同时,图文搜索流量占比回升至42%,说明信息密度高的图文笔记在搜索场景下依然有独特优势。建议围绕「产品词 + 场景词 + 痛点词」构建关键词矩阵,精准承接用户的搜索需求。In 2026, Xiaohongshu further unified its search and recommendation systems. The platform added an "intent recognition" module that distinguishes "informational searches" (e.g., "how to remove blackheads") from "transactional searches" (e.g., "blackhead remover recommendations"), matching different content types accordingly. Long-tail keywords now carry significantly more weight—"oily-sensitive skin summer moisturizer" ranks more easily than the generic "moisturizer." Image-text search traffic has recovered to 42%, confirming that information-dense posts retain unique advantages in search. Build your keyword matrix around "product + scenario + pain point" terms to capture search intent precisely.