[{"data":1,"prerenderedAt":818},["ShallowReactive",2],{"home-latest-blog":3,"home-latest-news":354},[4,130,238],{"id":5,"title":6,"author":7,"body":8,"canonical":114,"category":115,"cover":114,"date":116,"description":117,"draft":118,"extension":119,"keywords":114,"meta":120,"navigation":121,"noindex":118,"ogImage":114,"path":122,"seo":123,"seoDescription":114,"seoTitle":114,"stem":124,"tags":125,"updated":114,"__hash__":129},"blog\u002Fblog\u002Fcitation-click-landing.md","被 AI 引用之后，用户点进来落在哪一页","元析智能",{"type":9,"value":10,"toc":103},"minimark",[11,15,19,22,25,28,31,34,37,40,44,47,50,53,57,60,63,66,70,73,76,79,82,85,88,91,94,97,100],[12,13,14],"p",{},"2026 年 11 月我们查一个客户的 AI 引用数据，发现有内容的页面被引用最多，但那些页面的留资几乎为零。引用带来了访问，访问走到了死路。",[16,17,18],"h2",{"id":18},"引用点击会落在哪",[12,20,21],{},"大部分情况落在被引用的那个页面，因为用户点的是答案里的来源链接。",[12,23,24],{},"少数情况用户不会点，而是看完回答直接搜品牌名。这部分流量会被记成品牌词搜索或直接访问。",[12,26,27],{},"两种情况都要求被引用页面本身能承接。",[16,29,30],{"id":30},"承接断在哪",[12,32,33],{},"被引用的页面通常是文档、博客、术语解释，这类页面的设计目标是讲清楚一个问题，不是促成转化。",[12,35,36],{},"于是出现一个断点：用户读完，觉得有用，但不知道下一步做什么。",[12,38,39],{},"页面上没有入口、没有相关产品、没有更多资料，用户就走了。",[16,41,43],{"id":42},"第一步找到被引用的页面","第一步：找到被引用的页面",[12,45,46],{},"从引用记录里整理页面清单，按被引用次数排序。",[12,48,49],{},"再和流量数据交叉，看这些页面实际有多少访问。",[12,51,52],{},"两个数据都靠前的页面，优先改。",[16,54,56],{"id":55},"第二步加一个不打扰的下一步","第二步：加一个不打扰的下一步",[12,58,59],{},"在文章末尾加一句，指向更详细的内容或者试用入口，比如\"想看完整的配置步骤，这里有文档\"。",[12,61,62],{},"在正文中间插一个轻量入口，比如\"这一节涉及的具体参数，我们的产品页有完整列表\"。",[12,64,65],{},"不要弹窗，不要遮内容。被引用页面的读者是来找答案的，打扰就流失。",[16,67,69],{"id":68},"第三步别把文档页改成销售页","第三步：别把文档页改成销售页",[12,71,72],{},"这是最容易犯的错。看到文档页流量大，就往上加各种 CTA 和表单。",[12,74,75],{},"结果内容变得难读，引用减少，流量也没了。",[12,77,78],{},"我们的原则是内容主体不动，只在合适的节点加导航，加完之后自己读一遍，不别扭才算过。",[16,80,81],{"id":81},"怎么测量",[12,83,84],{},"AI 引用的来源流量很难精确归因，因为很多引擎不传 referrer。",[12,86,87],{},"我们的做法是间接看三个数：被引用页面的总访问、这些页面的站内跳转率、品牌词搜索变化。",[12,89,90],{},"站内跳转率是最直接的指标。加了下一步之后，从文档页跳到产品页的比例应该上升。",[16,92,93],{"id":93},"一个态度",[12,95,96],{},"GEO 的终点不是被引用，是被引用之后有人愿意继续了解你。",[12,98,99],{},"引用是入口，承接是转化。只做引用不做承接，等于把流量引到一个没有门的房间。",[12,101,102],{},"这件事的投入很小，改几个页面的末尾就行。但它决定了 GEO 的投入能不能变成生意。",{"title":104,"searchDepth":105,"depth":105,"links":106},"",2,[107,108,109,110,111,112,113],{"id":18,"depth":105,"text":18},{"id":30,"depth":105,"text":30},{"id":42,"depth":105,"text":43},{"id":55,"depth":105,"text":56},{"id":68,"depth":105,"text":69},{"id":81,"depth":105,"text":81},{"id":93,"depth":105,"text":93},null,"SEO","2026-11-07","引用带来的点击通常落在被引用页面，而那一页往往没有转化入口。这篇写承接断在哪、怎么在不打扰阅读的前提下加下一步，以及怎么测量。",false,"md",{},true,"\u002Fblog\u002Fcitation-click-landing",{"title":6,"description":117},"blog\u002Fcitation-click-landing",[126,127,128],"GEO","转化路径","落地页","59VysvmXMIaHC1faOdEZ615TAhHXhzDH2Ig14Ey9ZV4",{"id":131,"title":132,"author":7,"body":133,"canonical":114,"category":115,"cover":114,"date":227,"description":228,"draft":118,"extension":119,"keywords":114,"meta":229,"navigation":121,"noindex":118,"ogImage":114,"path":230,"seo":231,"seoDescription":114,"seoTitle":114,"stem":232,"tags":233,"updated":114,"__hash__":237},"blog\u002Fblog\u002Fkeyword-library-maintenance.md","词库半年不维护就废了，一半的词已经没人搜",{"type":9,"value":134,"toc":216},[135,138,141,144,147,150,153,156,159,162,165,168,171,174,177,180,183,186,189,192,195,198,201,204,207,210,213],[12,136,137],{},"词库半年不维护就废了。2026 年 4 月我们把 2025 年做的词表翻出来重看，870 个词里，210 个已经写过，130 个连续 6 个月零曝光，60 个词的意思已经变了。",[16,139,140],{"id":140},"半年里会发生什么",[12,142,143],{},"三件事。新词出现，产品迭代、行业换说法都会带来新词。旧词消失，功能下线、用户改叫法。词义漂移，同一个词两年前是学习意图，现在是采购意图。",[12,145,146],{},"不维护的词库，一半是过期信息，用起来比没有更危险。",[16,148,149],{"id":149},"每月做什么",[12,151,152],{},"30 分钟，三件事：加新词、标已写的、删连续两月零曝光的。",[12,154,155],{},"新词来源固定三个：搜索后台的新查询、工单新问题、销售提的新叫法。",[16,157,158],{"id":158},"每季度做什么",[12,160,161],{},"重算一次优先级。竞争格局会变，原来好打的词可能被别人做掉了，原来难的可能空出来。",[12,163,164],{},"我们那次季度重算，120 个词里 31 个分数变化超过 20 分。",[16,166,167],{"id":167},"每年做什么",[12,169,170],{},"重建一次。把全年写过的内容倒推回去，看哪些词表里的词真的被写成了页面，哪些没写。",[12,172,173],{},"2026 年 4 月那次倒推发现，870 个词里只有 210 个变成了内容，占 24%。剩下 76% 是白挖的，说明挖词阶段的筛选太松。",[16,175,176],{"id":176},"谁来维护",[12,178,179],{},"写内容的人自己维护，别外包。词库是判断工具，不是资料库，交给别人维护，判断就断了。",[12,181,182],{},"我们现在的做法是每周五下班前 30 分钟，加每月第一个周一 30 分钟重算。",[16,184,185],{"id":185},"一个更狠的做法",[12,187,188],{},"每季度直接把词库里三个月没动过的词删掉。删了不心疼，说明它们本来也不重要。",[12,190,191],{},"我们试过一轮，删掉 300 多个词，后来真正被找回来的不到 20 个。",[16,193,194],{"id":194},"最后",[16,196,197],{"id":197},"一个维护中的坑",[12,199,200],{},"别让两个人都能改同一张表而不留记录。我们有过一次，两个人同时整理，把对方的标签覆盖了，一周的工作白做。",[12,202,203],{},"后来改成每周只有一个人有编辑权，其他人提交申请。",[16,205,206],{"id":206},"怎么判断该重算了",[12,208,209],{},"看三个信号：上季度的选题命中率、新词出现速度、有没有词从第 30 名外突然进前 10。",[12,211,212],{},"出现两个，就该重算，不用等季度末。",[12,214,215],{},"词库的价值不在大，在准。一个 200 词的库，每个词都有承接页，比 2000 个词的库有用得多。",{"title":104,"searchDepth":105,"depth":105,"links":217},[218,219,220,221,222,223,224,225,226],{"id":140,"depth":105,"text":140},{"id":149,"depth":105,"text":149},{"id":158,"depth":105,"text":158},{"id":167,"depth":105,"text":167},{"id":176,"depth":105,"text":176},{"id":185,"depth":105,"text":185},{"id":194,"depth":105,"text":194},{"id":197,"depth":105,"text":197},{"id":206,"depth":105,"text":206},"2026-11-06","2026 年 4 月我们重看 2025 年的词表，870 个词里 210 个已写、130 个零曝光、60 个词义变了。词库不是建一次就完事，我们固定每月小更新、每季度重算。",{},"\u002Fblog\u002Fkeyword-library-maintenance",{"title":132,"description":228},"blog\u002Fkeyword-library-maintenance",[234,235,236],"词库搭建","关键词研究","内容运营","G72W5tGDPZLVAHvuvMgoHBn-HNsd0T-zbQSXWK-RCtg",{"id":239,"title":240,"author":7,"body":241,"canonical":114,"category":236,"cover":114,"date":343,"description":344,"draft":118,"extension":119,"keywords":114,"meta":345,"navigation":121,"noindex":118,"ogImage":114,"path":346,"seo":347,"seoDescription":114,"seoTitle":114,"stem":348,"tags":349,"updated":114,"__hash__":353},"blog\u002Fblog\u002Fcontent-half-life.md","内容的半衰期：哪些文章三年后还在带量",{"type":9,"value":242,"toc":333},[243,246,249,252,255,258,261,264,267,270,273,276,279,282,285,288,291,294,297,300,303,306,309,312,315,318,321,324,327,330],[12,244,245],{},"内容有半衰期，有的三个月就没流量，有的三年还在涨。我们跟踪了 120 篇内容的流量曲线，分出三类：快消型、稳增型、长尾型。",[16,247,248],{"id":248},"什么是内容半衰期",[12,250,251],{},"一篇文章的月访问量掉到峰值的 50%，所花的时间，就是它的半衰期。",[12,253,254],{},"半衰期短的内容，需要持续补充；半衰期长的内容，是资产。",[16,256,257],{"id":257},"三类半衰期",[12,259,260],{},"快消型：半衰期 1 到 3 个月。新闻、热点分析、活动内容。写完就得推，过了时间就没用。",[12,262,263],{},"稳增型：半衰期 12 到 18 个月。教程、选型、对比。前期爬坡慢，稳定后有持续流量。",[12,265,266],{},"长尾型：半衰期 3 年以上。概念解释、踩坑记录、方法论。流量不大，但一直有。",[16,268,269],{"id":269},"我们跟踪的数据",[12,271,272],{},"2024 年 9 月开始，我们跟踪 120 篇文章 2 年：快消型 34 篇，平均半衰期 1.8 个月；稳增型 61 篇，平均 14 个月；长尾型 25 篇，两年后流量是峰值的 70%。",[12,274,275],{},"有意思的是，长尾型内容单篇流量最低，但两年累计流量最高，因为它在持续带量。",[16,277,278],{"id":278},"什么内容半衰期长",[12,280,281],{},"第一，回答\"不会变的问题\"。什么是 API、为什么需要限流，这些十年后还有人问。",[12,283,284],{},"第二，有独家信息的内容。别人抄不走，所以能长期占位。",[12,286,287],{},"第三，结构完整、可被引用的内容。AI 搜索和同行引用会持续带来流量。",[16,289,290],{"id":290},"怎么用半衰期做排期",[12,292,293],{},"快消型：控制在 20% 以内。做多了，团队会很累，而且不积累。",[12,295,296],{},"稳增型：主力，占 50%。这是搜索流量的基本盘。",[12,298,299],{},"长尾型：占 30%。短期看不到效果，但每年都在积累。",[12,301,302],{},"我们的季度排期一般按 2:5:3 分，和流量贡献的节奏匹配。",[16,304,305],{"id":305},"一个提醒",[12,307,308],{},"半衰期会变长。一篇教程如果每年更新，半衰期可以从 14 个月拉到 3 年。",[12,310,311],{},"所以\"更新\"不只是维护，还是延长资产寿命。",[16,313,314],{"id":314},"两个陷阱",[12,316,317],{},"第一个陷阱，把快消内容当资产。热点文章写完就完了，别指望它持续带量，也别为它做大量内链。",[12,319,320],{},"第二个陷阱，用快消数据判断团队。如果团队这个月写了两篇热点，流量好看，但三个月后归零。考核要看长尾型内容的存量。",[12,322,323],{},"我们的做法是分两个指标：当月流量看新内容，年度流量看资产内容。",[16,325,326],{"id":326},"一句判断",[12,328,329],{},"做内容不能只看当月流量。快消型内容让你这个月好看，长尾型内容让你三年后还有饭吃。",[12,331,332],{},"两者都要有，但比例别搞反。我们的比例是 2:5:3，跑了两年，比 5:3:2 稳。这个比例每年调一次，看当年的数据再定。",{"title":104,"searchDepth":105,"depth":105,"links":334},[335,336,337,338,339,340,341,342],{"id":248,"depth":105,"text":248},{"id":257,"depth":105,"text":257},{"id":269,"depth":105,"text":269},{"id":278,"depth":105,"text":278},{"id":290,"depth":105,"text":290},{"id":305,"depth":105,"text":305},{"id":314,"depth":105,"text":314},{"id":326,"depth":105,"text":326},"2026-11-05","内容有半衰期，有的三个月就没流量，有的三年还在涨。我们跟踪了 120 篇内容的流量曲线，分出三类：快消型、稳增型、长尾型。半衰期怎么判断、怎么用它做排期都在这篇，数据来自我们对 120 篇内容的两年跟踪。",{},"\u002Fblog\u002Fcontent-half-life",{"title":240,"description":344},"blog\u002Fcontent-half-life",[350,351,352],"内容评估","内容策略","数据","q-rbxJmVri6Tw5XjEVR2NSeOO82TWzmDEap-WFPvTak",[355,467,674],{"id":356,"title":357,"body":358,"canonical":114,"category":458,"date":227,"description":459,"draft":118,"extension":119,"keywords":460,"meta":461,"navigation":121,"noindex":118,"ogImage":114,"path":462,"seo":463,"seoDescription":114,"seoTitle":114,"source":464,"stem":465,"__hash__":466},"news\u002Fnews\u002Findustry-digest.md","搜索与 AI 行业动态速览（每日自动更新）",{"type":9,"value":359,"toc":453},[360,373,377,414,417,447,450],[361,362,363,366],"blockquote",{},[12,364,365],{},"本页由官网脚本每天自动汇总外部权威信源，只记录标题、来源、发布日期和原文链接，不改写、不评论原文；内容以链接原文为准。",[12,367,368,369,372],{},"原文标题保持原语言（多为英文）。最近更新时间：",[370,371,227],"code",{},"（UTC+8，页面每天定时刷新）。",[374,375,227],"h3",{"id":376},"_2026-11-06",[378,379,380,391,399,406],"ul",{},[381,382,383,390],"li",{},[384,385,389],"a",{"href":386,"rel":387},"https:\u002F\u002Fwww.searchenginejournal.com\u002Fgoogle-search-queries-hit-a-5-year-high-and-search-revenue-didnt-follow\u002F595485\u002F",[388],"nofollow","Google Search Queries Hit A 5-Year High, And Search Revenue Didn’t Follow"," · Search Engine Journal",[381,392,393,398],{},[384,394,397],{"href":395,"rel":396},"https:\u002F\u002Fwww.seroundtable.com\u002Fopenai-updates-chatgpt-ads-bidding-options-42470.html",[388],"OpenAI Updates ChatGPT Ads Bidding Options For Max Bid or Max CPC"," · Search Engine Roundtable",[381,400,401,390],{},[384,402,405],{"href":403,"rel":404},"https:\u002F\u002Fwww.searchenginejournal.com\u002Fnovember-click-spikes-in-google-search-console-are-bot-traffic-how-to-filter-your-reports\u002F595834\u002F",[388],"November Click Spikes In Google Search Console Are Bot Traffic – How To Filter Your Reports",[381,407,408,413],{},[384,409,412],{"href":410,"rel":411},"https:\u002F\u002Fblog.google\u002Finnovation-and-ai\u002Ftechnology\u002Fdevelopers-tools\u002Fintroducing-gemini-omni-1-2-flash\u002F",[388],"Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips"," · Google AI 博客",[374,415,343],{"id":416},"_2026-11-05",[378,418,419,426,433,440],{},[381,420,421,390],{},[384,422,425],{"href":423,"rel":424},"https:\u002F\u002Fwww.searchenginejournal.com\u002Fgoogle-ai-answer-ads-apparel-electronics-report\u002F595984\u002F",[388],"Ads In Google AI Answers Concentrate In Apparel & Electronics",[381,427,428,398],{},[384,429,432],{"href":430,"rel":431},"https:\u002F\u002Fwww.seroundtable.com\u002Ffacebook-answers-google-42467.html",[388],"Facebook Answers AI-Generated 6 Million Pages Is Google Scaled Content Abuse",[381,434,435,390],{},[384,436,439],{"href":437,"rel":438},"https:\u002F\u002Fwww.searchenginejournal.com\u002Fgeo-isnt-a-distraction-referral-data-misses-where-ai-shapes-buying-decisions\u002F595730\u002F",[388],"GEO Isn’t A Distraction – Referral Data Misses Where AI Shapes Buying Decisions",[381,441,442,398],{},[384,443,446],{"href":444,"rel":445},"https:\u002F\u002Fwww.seroundtable.com\u002Fgoogle-ads-expands-ai-content-label-disclosure-42461.html",[388],"Google Ads Expands AI Content Disclosure Label To More Regions",[448,449],"hr",{},[12,451,452],{},"信源简介：Search Engine Roundtable、Search Engine Journal 是海外 SEO 行业媒体；Google AI 博客、OpenAI 官方博客提供 AI 引擎动态。以上链接均指向原文，不转载全文。",{"title":104,"searchDepth":105,"depth":105,"links":454},[455,457],{"id":376,"depth":456,"text":227},3,{"id":416,"depth":456,"text":343},"行业动态","每天自动汇总 Search Engine Roundtable、Search Engine Journal、Google AI 博客、OpenAI 官方博客的搜索与 AI 引擎动态。只记录标题、来源、发布日期和原文链接，不改写原文；内容以链接原文为准。","SEO动态,GEO动态,搜索引擎算法更新,AI搜索动态",{},"\u002Fnews\u002Findustry-digest",{"title":357,"description":459},"外部信源自动汇总","news\u002Findustry-digest","OB6v_dirD5pvO3L9MMZeg2T5QQw_-UpmCd2_Kx8RgnE",{"id":468,"title":469,"body":470,"canonical":114,"category":665,"date":666,"description":667,"draft":118,"extension":119,"keywords":114,"meta":668,"navigation":121,"noindex":118,"ogImage":114,"path":669,"seo":670,"seoDescription":114,"seoTitle":114,"source":671,"stem":672,"__hash__":673},"news\u002Fnews\u002Fkimi-web-search.md","Kimi 升级联网搜索,国内 AI 搜索竞争加剧",{"type":9,"value":471,"toc":658},[472,475,479,482,496,499,503,506,588,591,594,597,603,608,613,618,622,625,628,634,640,646,649,652,655],[12,473,474],{},"Kimi 1 月升级了联网搜索,响应速度和信源覆盖都明显改善。结论先给:国内 AI 搜索市场四家(文心一言、豆包、Kimi、通义)的格局基本成型,但各家抓取偏好差异大,做国内 GEO 不能一套策略打全部。",[16,476,478],{"id":477},"kimi-这次升级了什么","Kimi 这次升级了什么",[12,480,481],{},"1 月下旬,月之暗面对 Kimi 的联网搜索做了升级。我们测下来几个变化:",[378,483,484,487,490,493],{},[381,485,486],{},"响应速度比之前快,复杂问题 3-5 秒出结果(之前 8-10 秒)",[381,488,489],{},"信源覆盖更广,知乎、微信公众号、新闻源都进来了",[381,491,492],{},"引用来源比之前多,平均一个回答引用 4-6 个信源(之前 2-3 个)",[381,494,495],{},"多轮追问稳定性提升",[12,497,498],{},"这不是小修小补,是 Kimi 在补联网搜索这块的短板。之前 Kimi 强在长文本理解,弱在实时搜索,这次升级明显是想把搜索这条腿也补上。",[16,500,502],{"id":501},"国内-ai-搜索四家格局","国内 AI 搜索四家格局",[12,504,505],{},"2025 年下半年到 2026 年初,国内 AI 搜索市场基本形成四家分立的局面:",[507,508,509,528],"table",{},[510,511,512],"thead",{},[513,514,515,519,522,525],"tr",{},[516,517,518],"th",{},"厂商",[516,520,521],{},"强项",[516,523,524],{},"用户特征",[516,526,527],{},"抓取偏好",[529,530,531,546,560,574],"tbody",{},[513,532,533,537,540,543],{},[534,535,536],"td",{},"文心一言",[534,538,539],{},"百度系资源、资讯全",[534,541,542],{},"大众用户、信息查询",[534,544,545],{},"百度系内容(百家号、百度百科)、新闻源",[513,547,548,551,554,557],{},[534,549,550],{},"豆包",[534,552,553],{},"字节系资源、响应快",[534,555,556],{},"年轻用户、生活查询",[534,558,559],{},"抖音、今日头条、新闻源",[513,561,562,565,568,571],{},[534,563,564],{},"Kimi",[534,566,567],{},"长文本、深度分析",[534,569,570],{},"知识工作者、开发者",[534,572,573],{},"知乎、公众号、技术社区",[513,575,576,579,582,585],{},[534,577,578],{},"通义",[534,580,581],{},"阿里系资源、电商场景",[534,583,584],{},"商务用户、B 端",[534,586,587],{},"淘宝、1688、行业资讯",[12,589,590],{},"四家各有基本盘,短期看格局稳定。Kimi 这次升级后,在\"深度搜索\"这个细分场景上,和文心一言、通义的差距在缩小。",[16,592,593],{"id":593},"各家抓取偏好差异",[12,595,596],{},"这是做国内 GEO 最要关注的点。四家 AI 引擎的抓取偏好差别很大,同一套内容,在四家被引用的概率差很多。",[12,598,599,602],{},[600,601,536],"strong",{},"重百度系内容。百家号、百度百科、百度知道在文心一言里权重高。如果你想被文心一言引用,百度系内容布局是必须的。",[12,604,605,607],{},[600,606,550],{},"重字节系。抖音短视频描述、今日头条文章是主要信源。豆包用户偏生活场景,B 端软件公司在这家能拿到的流量有限。",[12,609,610,612],{},[600,611,564],{},"重深度内容。知乎高赞回答、公众号长文、技术社区(掘金、CSDN)是主要信源。Kimi 用户里知识工作者和开发者占比高,软件公司在这家值得重点投入。",[12,614,615,617],{},[600,616,578],{},"重 B 端和电商。1688 商品页、行业资讯站是主要信源。软件公司做 B 端选型内容,通义是个该关注的渠道。",[16,619,621],{"id":620},"对国内-geo-策略的影响","对国内 GEO 策略的影响",[12,623,624],{},"国内 GEO 和海外 GEO 最大的不同,是引擎多、偏好分散。海外基本就是 ChatGPT + Perplexity 两家,加 Google AI Overviews 三家。国内四家各有偏好,一套内容打不过来。",[12,626,627],{},"我们建议的国内 GEO 策略分三步:",[12,629,630,633],{},[600,631,632],{},"第一步,按目标用户选主战场。"," 你的用户是开发者,重点投 Kimi(知乎、技术社区);你的用户是企业采购,重点投通义和文心一言(行业资讯、百度系);你的用户是大众,豆包不能漏。",[12,635,636,639],{},[600,637,638],{},"第二步,内容按引擎偏好分发。"," 同一个产品的对比内容,知乎版写深一点(给 Kimi),公众号版写全一点(给文心一言),行业媒体版写专业一点(给通义)。不是一稿多发,是按引擎偏好调整。",[12,641,642,645],{},[600,643,644],{},"第三步,做多引擎可见度监测。"," 国内四家加上 ChatGPT(国内开发者也在用),五个引擎每周跑一批关键词,记录引用情况。没有现成工具,得自己搭脚本。",[16,647,648],{"id":648},"我们怎么看",[12,650,651],{},"Kimi 这次升级是 2026 年国内 AI 搜索竞争的第一枪。后面几个月,文心一言、豆包、通义都会跟进升级,四家的差距会进一步缩小,竞争会更激烈。",[12,653,654],{},"对做 GEO 的人是好事——引擎竞争意味着各家都会优化抓取和引用机制,内容质量好的站被引用的概率会上升。但难度也在涨——四家偏好分散,一套策略打不过来,必须做差异化。",[12,656,657],{},"我们的判断:2026 年国内 GEO 会从\"可选项\"变成软件公司的标配。但和海外不同,国内没有一家独大的 AI 引擎,做多引擎差异化布局,是这一年软件公司要解决的核心问题。早动手的,能在四家都占住位置;晚动手的,只能挤一家。",{"title":104,"searchDepth":105,"depth":105,"links":659},[660,661,662,663,664],{"id":477,"depth":105,"text":478},{"id":501,"depth":105,"text":502},{"id":593,"depth":105,"text":593},{"id":620,"depth":105,"text":621},{"id":648,"depth":105,"text":648},"行业资讯","2026-01-22","Kimi 1 月升级联网搜索,国内 AI 搜索市场四家(文心一言、豆包、Kimi、通义)格局基本成型。各家抓取偏好不同,国内 GEO 策略要做差异化。",{},"\u002Fnews\u002Fkimi-web-search",{"title":469,"description":667},"月之暗面","news\u002Fkimi-web-search","EvJbWJF2o7NodVp2fzKNDvn8iYmtfo-Pdw3neJQWgDc",{"id":675,"title":676,"body":677,"canonical":114,"category":665,"date":810,"description":811,"draft":118,"extension":119,"keywords":114,"meta":812,"navigation":121,"noindex":118,"ogImage":114,"path":813,"seo":814,"seoDescription":114,"seoTitle":114,"source":815,"stem":816,"__hash__":817},"news\u002Fnews\u002Fgoogle-core-update.md","Google 2025 年第三次核心算法更新:低质内容站继续被清洗",{"type":9,"value":678,"toc":803},[679,682,685,688,691,705,708,711,714,720,726,732,738,742,745,759,762,765,768,774,780,786,792,794,797,800],[12,680,681],{},"Google 2025 年第三次核心算法更新 1 月初完成滚动,SearchEngineLand 的跟踪数据显示,这次更新对低质内容站的打击比前两次都重。结论先给:对认真做内容的软件公司官网是利好,流量会有回升;但靠 AI 批量生成内容和堆外链的站,这次掉得厉害。",[16,683,684],{"id":684},"这次更新是什么",[12,686,687],{},"Google 核心更新一年通常发 3-4 次,2025 年第三次在 12 月底开始,1 月初完成。核心更新不是针对某个作弊手法,是 Google 对\"什么算高质量内容\"的整体判断调整。",[12,689,690],{},"这次更新的特点,根据 SearchEngineLand 和几家 SEO 工具( Semrush、Sistrix)的数据:",[378,692,693,696,699,702],{},[381,694,695],{},"受影响最大的是内容站,尤其是 AI 生成内容占比高的站",[381,697,698],{},"联盟营销站(affiliate)受打击明显",[381,700,701],{},"老牌媒体和品牌官网多数稳定或上升",[381,703,704],{},"论坛和 UGC 站点(Reddit、Quora 等)继续涨",[12,706,707],{},"Sistrix 的数据显示,这次更新期间,排名前 100 的内容站里有约 18% 出现明显波动,其中下跌的占多数。",[16,709,710],{"id":710},"低质内容站被洗的标准",[12,712,713],{},"Google 没公开\"低质\"的精确判断标准,但从被洗的站反推,几个特征很明显:",[12,715,716,719],{},[600,717,718],{},"第一,内容信息密度低。"," 一篇文章 800 字,删掉一半意思不变。这类内容在 AI 生成站里最常见。Google 的 Helpful Content 系统对这种内容的识别更准了。",[12,721,722,725],{},[600,723,724],{},"第二,内容主题和站点定位不匹配。"," 一个软件公司官网发\"如何选牙刷\"\"十大旅游景点\",这是典型为流量而内容,被识别后整站权重受影响。",[12,727,728,731],{},[600,729,730],{},"第三,E-E-A-T 信号弱。"," Experience(经验)、Expertise(专业)、Authoritativeness(权威)、Trust(信任)。没有作者署名、没有作者专业背景、没有第一手经验的内容,E-E-A-T 分低,这次更新里掉得明显。",[12,733,734,737],{},[600,735,736],{},"第四,外链质量差。"," 大量低质外链、买卖外链的站,这次继续被清。",[16,739,741],{"id":740},"对软件公司官网的影响偏正面","对软件公司官网的影响(偏正面)",[12,743,744],{},"软件公司官网这次多数是受益的。我们监测的几个客户站,1 月初自然流量平均涨了 8%-15%。原因不复杂:",[378,746,747,750,753,756],{},[381,748,749],{},"软件公司官网内容主题集中,围绕产品和技术,匹配度高",[381,751,752],{},"多数有真实作者(技术团队),E-E-A-T 信号强",[381,754,755],{},"外链来源相对真实(行业媒体、社区、合作方)",[381,757,758],{},"受打击的联盟站和 AI 内容站掉下去后,排名位置让出来了",[12,760,761],{},"不是所有软件公司都涨。涨得明显的,是过去半年一直在做内容建设的;没动作的,基本持平。这印证了一件事:核心更新对认真做内容的站是奖励,不是惩罚。",[16,763,764],{"id":764},"怎么做才能不被洗",[12,766,767],{},"如果你想让自己的软件公司官网在下一波核心更新里也站稳,关注几件事:",[12,769,770,773],{},[600,771,772],{},"把作者署名和专业背景做出来。"," 技术博客每篇要有真实作者,作者页面要有专业背景介绍。这件事成本低,但 E-E-A-T 信号提升明显。",[12,775,776,779],{},[600,777,778],{},"内容只写自己懂的领域。"," 软件公司官网就写产品、技术、行业。别为了 SEO 去写泛流量话题。Google 现在对主题集中度的判断比一年前强。",[12,781,782,785],{},[600,783,784],{},"删掉低质旧内容。"," 官网上有两年前写的、现在看信息密度低的内容,该删就删,该重写就重写。留着会拉低整站质量评分。",[12,787,788,791],{},[600,789,790],{},"外链只做真实的。"," 行业媒体投稿、社区讨论、合作方互链,这些是真实外链。买外链这条路 2026 年彻底走不通了。",[16,793,648],{"id":648},[12,795,796],{},"Google 这几次核心更新的方向很一致:奖励认真做内容的,洗掉投机取巧的。对软件公司是好事——本来就在认真做内容,之前被低质站挤掉的排名位置,现在开始往回还。",[12,798,799],{},"但别误以为\"不做 SEO 也会涨\"。涨的站都是过去半年在内容上花了力气的。如果去年你什么都没做,这次更新和你关系不大。核心更新是放大器,放大你做对的事,也放大你做错的事。",[12,801,802],{},"下一波核心更新估计在 2026 年 3-4 月。这段时间把作者署名、内容主题集中度、旧内容清理这三件事做了,下一波还能继续受益。",{"title":104,"searchDepth":105,"depth":105,"links":804},[805,806,807,808,809],{"id":684,"depth":105,"text":684},{"id":710,"depth":105,"text":710},{"id":740,"depth":105,"text":741},{"id":764,"depth":105,"text":764},{"id":648,"depth":105,"text":648},"2026-01-10","Google 2026 年 1 月核心更新完成,低质内容站继续被清洗。对认真做内容的软件公司官网是利好,但有几点要注意。",{},"\u002Fnews\u002Fgoogle-core-update",{"title":676,"description":811},"SearchEngineLand","news\u002Fgoogle-core-update","tMfYP8p4-KAj0R9KD6i3XyI3a9wqkHNXEjTfIzpY8Lw",1794088098156]