[{"data":1,"prerenderedAt":438},["ShallowReactive",2],{"blog-cross-engine-citation-gap":3,"blog-prev-cross-engine-citation-gap":240,"blog-next-cross-engine-citation-gap":352},{"id":4,"title":5,"author":6,"body":7,"canonical":223,"category":224,"cover":223,"date":225,"description":226,"draft":227,"extension":228,"keywords":223,"meta":229,"navigation":230,"noindex":227,"ogImage":223,"path":231,"seo":232,"seoDescription":223,"seoTitle":223,"stem":233,"tags":234,"updated":223,"__hash__":239},"blog\u002Fblog\u002Fcross-engine-citation-gap.md","同一篇文章在 6 个 AI 引擎里的引用差异：一张对照表","元析智能",{"type":8,"value":9,"toc":214},"minimark",[10,14,18,21,24,134,137,144,150,156,159,162,168,174,180,183,186,202,205,208,211],[11,12,13],"p",{},"同一篇文章，在六个引擎里表现完全不同。2026 年 3 月我们拿一个客户的 20 篇内容跑 60 个查询，记录被引来源。ChatGPT 偏爱文档，Perplexity 偏爱对比表，AI Overviews 偏表格和定义段，豆包偏知乎，元宝偏公众号，DeepSeek 偏技术博客。",[15,16,17],"h2",{"id":17},"测试怎么做的",[11,19,20],{},"20 篇内容覆盖三类：8 篇文档页、6 篇对比页、6 篇博客。60 个查询，每类各 20 个。六个引擎各跑一遍，记录回答里出现的目标内容，以及挂出来的来源。",[11,22,23],{},"样本不算大，但差异明显到不需要统计检验。",[25,26,27,46],"table",{},[28,29,30],"thead",{},[31,32,33,37,40,43],"tr",{},[34,35,36],"th",{},"引擎",[34,38,39],{},"最常引的来源",[34,41,42],{},"官网占比",[34,44,45],{},"特点",[47,48,49,64,78,92,106,120],"tbody",{},[31,50,51,55,58,61],{},[52,53,54],"td",{},"ChatGPT",[52,56,57],{},"文档页、报错页",[52,59,60],{},"41%",[52,62,63],{},"按句挂源，读得深",[31,65,66,69,72,75],{},[52,67,68],{},"Perplexity",[52,70,71],{},"对比页、第三方评测",[52,73,74],{},"34%",[52,76,77],{},"多源交叉，来源多",[31,79,80,83,86,89],{},[52,81,82],{},"AI Overviews",[52,84,85],{},"表格、定义段",[52,87,88],{},"46%",[52,90,91],{},"只挂少数链接",[31,93,94,97,100,103],{},[52,95,96],{},"豆包",[52,98,99],{},"知乎、CSDN",[52,101,102],{},"23%",[52,104,105],{},"中文平台权重高",[31,107,108,111,114,117],{},[52,109,110],{},"元宝",[52,112,113],{},"公众号、门户",[52,115,116],{},"19%",[52,118,119],{},"平台内容池",[31,121,122,125,128,131],{},[52,123,124],{},"DeepSeek",[52,126,127],{},"技术博客、GitHub",[52,129,130],{},"27%",[52,132,133],{},"技术细节偏好强",[15,135,136],{"id":136},"三个反常识",[11,138,139,143],{},[140,141,142],"strong",{},"一、官网占比高的引擎，不一定是好消息。"," AI Overviews 官网占比 46%，但它挂的链接最少，一条回答平均 2.1 个。占比高是因为它几乎不引第三方，不是因为它特别喜欢你。",[11,145,146,149],{},[140,147,148],{},"二、Perplexity 的来源数量在下降。"," 我们 2025 年底测的时候平均 6.4 个，2026 年 3 月是 5.2 个。它开始收敛，意味着单条来源的权重变高，进去更难，进去之后更值钱。",[11,151,152,155],{},[140,153,154],{},"三、同一引擎中英文表现能差一倍。"," ChatGPT 英文查询引官网的比例 48%，中文查询 31%。中文查询它更多去找中文平台，因为你的中文内容密度不够。",[15,157,158],{"id":158},"怎么用这个差异",[11,160,161],{},"别写六版内容，那样成本撑不住。三种做法：",[11,163,164,167],{},[140,165,166],{},"底层事实句通用。"," 带数字、带条件、带时间的那句话，在所有引擎里都好使。这部分写一份，全平台复用。",[11,169,170,173],{},[140,171,172],{},"形态按引擎调。"," 同一批事实，在官网做成表格，在知乎写成长回答，在公众号写成场景故事,在 GitHub 写进 README。不是复制粘贴，是同一事实的四种容器。",[11,175,176,179],{},[140,177,178],{},"平台投放看引擎。"," 如果你的客户主要在中文引擎里做调研，就把资源压到知乎和公众号；如果客户是海外技术团队，官网文档和 GitHub 优先。",[15,181,182],{"id":182},"一个具体的分配",[11,184,185],{},"我们给一个客户排 2026 年 Q2 的内容，是这么分的：",[187,188,189,193,196,199],"ul",{},[190,191,192],"li",{},"官网：12 篇文档页改造 + 4 篇对比页",[190,194,195],{},"知乎：8 个高赞问题下写回答，每篇带一个可引用表格",[190,197,198],{},"公众号：6 篇场景文，讲客户现场",[190,200,201],{},"GitHub：README 重写 + 4 个常见报错的 Issue 模板",[11,203,204],{},"四块内容讲的是一批事实，形态完全不同。三个月后六个引擎的覆盖率从平均 12% 到 24%。",[15,206,207],{"id":207},"一句实在话",[11,209,210],{},"跨引擎差异这件事，知道就好，别过度优化。真正决定你被不被引用的是内容里有没有可搬走的句子，这件工作在哪个引擎都一样。",[11,212,213],{},"引擎的偏好会变，去年和今年就不一样。你追着它改，永远慢一步。",{"title":215,"searchDepth":216,"depth":216,"links":217},"",2,[218,219,220,221,222],{"id":17,"depth":216,"text":17},{"id":136,"depth":216,"text":136},{"id":158,"depth":216,"text":158},{"id":182,"depth":216,"text":182},{"id":207,"depth":216,"text":207},null,"行业观察","2026-03-07","20 篇内容、60 个查询、6 个引擎。ChatGPT 引文档，Perplexity 引对比，AI Overviews 引表格，豆包引知乎，元宝引公众号。一份内容通吃是幻觉。",false,"md",{},true,"\u002Fblog\u002Fcross-engine-citation-gap",{"title":5,"description":226},"blog\u002Fcross-engine-citation-gap",[235,236,237,238],"跨引擎","AI引用","GEO","平台差异","vA62lvr-AdPl0Ra4CThbC775hIOAXeYVtZG4nuHo2MQ",{"id":241,"title":242,"author":6,"body":243,"canonical":223,"category":340,"cover":223,"date":341,"description":342,"draft":227,"extension":228,"keywords":223,"meta":343,"navigation":230,"noindex":227,"ogImage":223,"path":344,"seo":345,"seoDescription":223,"seoTitle":223,"stem":346,"tags":347,"updated":223,"__hash__":351},"blog\u002Fblog\u002Fdelete-post-301-or-410.md","删掉的文章该 301 还是 410，我们的判断只有三条",{"type":8,"value":244,"toc":330},[245,248,252,255,258,262,265,268,272,275,278,281,284,287,290,293,296,299,302,305,309,312,315,318,321,324,327],[11,246,247],{},"删文章的时候，301 和 410 的选择比大多数人想的更有讲究。我们的判断只有三条：这个页面上有没有外链、有没有排名、有没有历史流量。三条都是否，才走 410，其余一律 301。",[15,249,251],{"id":250},"_410-适合什么页面","410 适合什么页面",[11,253,254],{},"从来没被收录、从来没曝光、没有任何外链、内容本身没有价值的页面。测试页、空标签页、批量生成的垃圾页，这类走 410。",[11,256,257],{},"410 的意思是“永久删除，别再来了”。它比 404 更明确，爬虫放弃得更快。",[15,259,261],{"id":260},"_301-适合什么页面","301 适合什么页面",[11,263,264],{},"有外链的，有排名的，有过流量的，被外部引用过的。哪怕现在没流量，只要有一个外链，就 301。",[11,266,267],{},"2025 年那次审计，88 篇里只有 12 篇走了 410，剩下 76 篇都 301 到最相关的一篇。判断标准就是这个。",[15,269,271],{"id":270},"_301-指向哪","301 指向哪",[11,273,274],{},"指向内容最接近的那篇，不是首页，不是分类页。",[11,276,277],{},"实在找不到相关页面的，指向同主题的支柱页。全站指向首页的做法，等于把一堆不相干的权重混在一起，搜索引擎会给这种改法打折。",[15,279,280],{"id":280},"别一次性批量删",[11,282,283],{},"我们按批次删，每批 20 篇以内，隔两周看一次抓取和收录。",[11,285,286],{},"一次性删 80 篇，日志里会出现一大批 404，抓取预算会被这些错误占掉。2024 年有个客户这么干过，收录恢复花了两个多月。",[15,288,289],{"id":289},"删之前做三件事",[11,291,292],{},"导出 URL 清单。查每个 URL 的外链数。在搜索后台提交移除请求，加速下架。",[11,294,295],{},"这三件事做完再动手，顺序反了容易漏掉有外链的页面。",[15,297,298],{"id":298},"一个我们踩过的坑",[11,300,301],{},"2025 年 8 月，我们把一个 410 用在了有 3 个外链的页面上。当时判断是“内容没用”，忽略了外链。",[11,303,304],{},"两个月后客户发现有个行业目录站还在链那个地址，权重全落空。后来补了 301，但延迟已经造成了。",[15,306,308],{"id":307},"_404-和-410-的实际差别","404 和 410 的实际差别",[11,310,311],{},"对搜索引擎来说，两者都表示页面不存在。410 更明确，一般下架更快；404 会被反复重试，有时候要几周才彻底移除。",[11,313,314],{},"但对用户来说，410 和 404 长得一样。所以这个选择的技术意义大于体验意义，别在这上面纠结太久。",[15,316,317],{"id":317},"一个可以偷懒的做法",[11,319,320],{},"拿不准的时候，用 301。301 至少有把权重转给别的页面的机会，410 什么都没有。",[11,322,323],{},"我们内部有个更简单的判断：如果这篇内容在退订邮件、销售资料、渠道目录里可能被提到过，就别用 410。链接一旦出现在别人的文档里，你就控制不了它什么时候生效。",[11,325,326],{},"我们给客户删页面时的默认选项就是 301，只有确认边缘价值为零，才改成 410。",[11,328,329],{},"应该是：先查外链，再判断内容价值，最后决定 301 还是 410。",{"title":215,"searchDepth":216,"depth":216,"links":331},[332,333,334,335,336,337,338,339],{"id":250,"depth":216,"text":251},{"id":260,"depth":216,"text":261},{"id":270,"depth":216,"text":271},{"id":280,"depth":216,"text":280},{"id":289,"depth":216,"text":289},{"id":298,"depth":216,"text":298},{"id":307,"depth":216,"text":308},{"id":317,"depth":216,"text":317},"技术SEO","2026-03-06","2025 年我们给客户删了 88 篇文章，其中只有 12 篇走 410，其余全部 301。判断依据就三条：有没有外链、有没有排名、有没有历史流量。判断错了就是白丢权重。",{},"\u002Fblog\u002Fdelete-post-301-or-410",{"title":242,"description":342},"blog\u002Fdelete-post-301-or-410",[348,349,350],"删除页面","301重定向","内容审计","ZPRFj_HA8gjqu4Wl_HHu11YwU3l-mQ7paseRzo43nMs",{"id":353,"title":354,"author":6,"body":355,"canonical":223,"category":426,"cover":223,"date":427,"description":428,"draft":227,"extension":228,"keywords":223,"meta":429,"navigation":230,"noindex":227,"ogImage":223,"path":430,"seo":431,"seoDescription":223,"seoTitle":223,"stem":432,"tags":433,"updated":223,"__hash__":437},"blog\u002Fblog\u002Fprivacy-policy-minimum-software.md","软件站隐私政策的最小写法：这六段必须写清",{"type":8,"value":356,"toc":421},[357,360,363,369,375,381,387,393,399,403,406,409,412,415,418],[11,358,359],{},"隐私政策抄模板，最大的问题不是法律风险，是采购方和搜索引擎都能看出你没认真写。2025 年 10 月我们给一个客户做站点评审，他的隐私政策是从同行那里复制来的，里面还留着对方公司的产品名。",[15,361,362],{"id":362},"六段必须自己写的内容",[11,364,365,368],{},[140,366,367],{},"收集了什么数据。"," 表单字段、Cookie、服务器日志、支付信息，逐项列出来。别写\"我们可能收集您在使用过程中的相关信息\"，这种话等于什么都没说。",[11,370,371,374],{},[140,372,373],{},"每条数据为什么收集。"," 收集邮箱是为了发试用账号，收集手机号是为了商务联系，收集 IP 是为了防刷。目的写清楚，用户才知道该不该给。",[11,376,377,380],{},[140,378,379],{},"保存多久。"," 写具体时长，比如\"试用期数据保留 6 个月\"\"日志保留 90 天\"。写\"在法律允许的最长期限内保留\"是偷懒。",[11,382,383,386],{},[140,384,385],{},"共享给谁。"," 云服务商、统计分析工具、支付通道、邮件服务，按类别列出。B2B 客户的法务会专门看这一段。",[11,388,389,392],{},[140,390,391],{},"用户怎么行使权利。"," 查询、更正、删除、导出，每一种给出具体路径和一个邮箱。只写\"您可以通过本政策所述方式联系我们\"是不够的。",[11,394,395,398],{},[140,396,397],{},"Cookie 怎么用。"," 分必要、统计、营销三类，说明各自的用途，并给出拒绝统计和营销 Cookie 的方式。",[15,400,402],{"id":401},"为什么-b2b-软件更要注意","为什么 B2B 软件更要注意",[11,404,405],{},"B2B 采购流程里，隐私政策是必查项。尤其是客户是金融、医疗、政企的时候，对方会问数据存在哪、谁能访问、出境不出境。这些答不上来，单子就卡住。",[11,407,408],{},"海外客户还会看你的政策里有没有数据处理者条款、有没有跨境传输说明。我们有个客户做海外生意，隐私页被对方打回两次，第二次是因为没有写清楚子处理者名单。",[15,410,411],{"id":411},"我们的写法建议",[11,413,414],{},"用表格列数据项，比大段文字好读。页面顶部放一个更新日期和联系人邮箱。全文用短句，别用法律套话堆满。",[11,416,417],{},"页面底部留一个版本记录，写清每次修改的日期和改了哪一段。我们给客户的模板里这一栏是必填的，因为用户和法务都会问\"什么时候改的、改了什么\"。没有版本记录的隐私页，看起来像三年没维护过。",[11,419,420],{},"隐私政策是信任文件。客户愿意把数据交给你，靠的不是条款写得多严密，是你写得让他看得懂。",{"title":215,"searchDepth":216,"depth":216,"links":422},[423,424,425],{"id":362,"depth":216,"text":362},{"id":401,"depth":216,"text":402},{"id":411,"depth":216,"text":411},"内容运营","2026-03-08","隐私政策抄模板，问题不在法律风险，在用户和搜索引擎都能看出你没认真写。这篇给六段必须自己写的内容：收集什么、为什么、存多久、给谁、怎么删、Cookie 怎么管。",{},"\u002Fblog\u002Fprivacy-policy-minimum-software",{"title":354,"description":428},"blog\u002Fprivacy-policy-minimum-software",[434,435,436],"隐私政策","信任页面","合规","hxwhfmNnP4f452NI21sTd_dm1jUHPF4BRS9sZKSC7oI",1794088107818]