[{"data":1,"prerenderedAt":599},["ShallowReactive",2],{"blog-seo-reporting-honest":3,"blog-prev-seo-reporting-honest":396,"blog-next-seo-reporting-honest":487},{"id":4,"title":5,"author":6,"body":7,"canonical":379,"category":380,"cover":379,"date":381,"description":382,"draft":383,"extension":384,"keywords":379,"meta":385,"navigation":386,"noindex":383,"ogImage":379,"path":387,"seo":388,"seoDescription":379,"seoTitle":379,"stem":389,"tags":390,"updated":379,"__hash__":395},"blog\u002Fblog\u002Fseo-reporting-honest.md","SEO 报告该怎么看:别被漂亮的图表骗了","元析智能",{"type":8,"value":9,"toc":365},"minimark",[10,14,19,22,35,42,46,49,52,57,61,64,67,72,76,79,82,93,101,105,108,116,121,125,128,133,136,139,144,147,152,155,160,163,180,183,188,191,196,199,204,207,211,214,219,222,227,230,235,238,243,246,251,254,259,262,265,268,274,313,318,350,353,356,359,362],[11,12,13],"p",{},"SEO 公司的周报月报,经常把不好的数据藏起来,只给你看好看的。客户看不懂,以为有效果,实际钱白花了。这篇文章写 SEO 报告常见的注水手法,以及真正该看的指标。我们自己也做 SEO 服务,这些是我们见过的同行操作,也是我们提醒客户警惕的事。",[15,16,18],"h2",{"id":17},"注水手法-1只挑好看的关键词","注水手法 1:只挑好看的关键词",[11,20,21],{},"最常见的注水手法。SEO 报告里放一张关键词排名表,20 个词全部上升,看起来效果不错。但你仔细看:",[23,24,25,29,32],"ul",{},[26,27,28],"li",{},"这 20 个词是从多少个词里挑出来的?(可能是 200 个里挑 20 个涨的)",[26,30,31],{},"这 20 个词的搜索量是多少?(可能全是月搜 10 的长尾)",[26,33,34],{},"排名从第 50 涨到第 30,这种\"上涨\"对流量贡献 0",[11,36,37,41],{},[38,39,40],"strong",{},"判断方法",":让 SEO 公司给完整的关键词排名表,不要\"精选\"。看全部词的分布,涨的多少、跌的多少、不动的多少。涨的比跌得多很多才算有效。",[15,43,45],{"id":44},"注水手法-2不显示搜索量","注水手法 2:不显示搜索量",[11,47,48],{},"只给关键词和排名,不给搜索量。客户不知道哪个词值钱哪个词不值钱。",[11,50,51],{},"举例:SEO 公司报告\"关键词 A 从第 20 涨到第 5\",看起来很好。但关键词 A 月搜索量 30,排到第 5 一个月带来 5 个点击。同期的关键词 B(月搜 3000)从第 8 掉到第 15,损失 200 个点击。整体是亏的,但报告只显示 A 涨了。",[11,53,54,56],{},[38,55,40],{},":报告必须带搜索量。每个关键词旁边标月搜索量。不带搜索量的关键词报告,基本是在藏数据。",[15,58,60],{"id":59},"注水手法-3用-brand-词充数","注水手法 3:用 brand 词充数",[11,62,63],{},"品牌词(你公司名、产品名)排名本来就第 1,且和 SEO 关系不大。SEO 公司把 brand 词放进报告充数,显得\"排名第一的关键词多\"。",[11,65,66],{},"举例:你公司叫\"元析智能\",SEO 公司报告\"元析智能\"排第 1,\"元析\"排第 1,\"元析智能 SEO\"排第 1。看起来 3 个第一,实际都是 brand 词,SEO 工作基本没贡献。",[11,68,69,71],{},[38,70,40],{},":报告要区分 brand 词和非 brand 词。非 brand 词的排名和流量才是 SEO 的真实效果。",[15,73,75],{"id":74},"注水手法-4只看点击不看转化","注水手法 4:只看点击不看转化",[11,77,78],{},"流量涨了,但转化没涨。SEO 公司报流量数据漂亮,转化数据藏起来。",[11,80,81],{},"原因可能:",[23,83,84,87,90],{},[26,85,86],{},"流量来源是宽泛词,意图不强(如搜\"SEO 是什么\"进来的人不会买服务)",[26,88,89],{},"承接页有问题,流量进来了不转化",[26,91,92],{},"排名涨的词和业务无关(如做 B2B 的公司,\"免费 SEO 工具\"排名涨了)",[11,94,95,97,100],{},[38,96,40],{},[98,99],"seo",{}," 报告必须带转化数据。流量和转化脱节,SEO 工作就是空的。",[15,102,104],{"id":103},"注水手法-5基准期不对","注水手法 5:基准期不对",[11,106,107],{},"SEO 报告经常用一个不公平的基准期对比。比如:",[23,109,110,113],{},[26,111,112],{},"基准期选淡季(如春节月),对比期是旺季,显得\"流量环比涨 200%\"",[26,114,115],{},"基准期选 SEO 接手前流量最低的一周,对比期是恢复后,显得\"涨了 3 倍\"",[11,117,118,120],{},[38,119,40],{},":基准期要选接手前 3 个月的平均值,且要对比去年同期(看季节性)。光看环比或光看同比都不够,两个一起看。",[15,122,124],{"id":123},"注水手法-6把-ppc-流量算进-seo","注水手法 6:把 PPC 流量算进 SEO",[11,126,127],{},"有些公司 PPC 和 SEO 一起做,报告里把 PPC 流量算进\"自然流量增长\"。客户分不清,以为是 SEO 起效。",[11,129,130,132],{},[38,131,40],{},":报告要分 SEO 和 PPC 流量。Google Analytics 里 source\u002Fmedium 是 organic \u002F search 的才是 SEO 流量,paid \u002F cpc 是 PPC。让 SEO 公司按这个维度分开报。",[15,134,135],{"id":135},"真正该看的指标",[11,137,138],{},"看完注水手法,说说真正该看的指标。一份诚实的 SEO 报告,应该包含:",[11,140,141],{},[38,142,143],{},"1. 自然流量总量",[11,145,146],{},"Google Analytics 的 organic \u002F search 维度。按月对比,看趋势。注意要排除 brand 流量,看非 brand 的自然流量。",[11,148,149],{},[38,150,151],{},"2. 非 brand 词排名",[11,153,154],{},"排名靠前的关键词里,有多少是 brand 词,多少是非 brand 词。非 brand 词的排名才是 SEO 效果。",[11,156,157],{},[38,158,159],{},"3. 关键词排名分布",[11,161,162],{},"不要只看 top 10 关键词。要看分布:",[23,164,165,168,171,174,177],{},[26,166,167],{},"第 1-3 名:几个",[26,169,170],{},"第 4-10 名:几个",[26,172,173],{},"第 11-20 名:几个",[26,175,176],{},"第 21-50 名:几个",[26,178,179],{},"第 50 名以后:几个",[11,181,182],{},"健康的 SEO 效果,应该是关键词从第 21-50 名往第 11-20 名移动,再到第 4-10 名。如果一直卡在第 11-20 名不动,说明有问题。",[11,184,185],{},[38,186,187],{},"4. 转化数据",[11,189,190],{},"通过 SEO 进来的流量,转化了多少。转化定义按业务定(注册、留资、购买)。流量没转化的 SEO 是空的。",[11,192,193],{},[38,194,195],{},"5. 收录情况",[11,197,198],{},"百度搜索资源平台和 Google Search Console 的收录数据。如果 SEO 做了 3 个月,收录页数没涨,内容建设有问题。",[11,200,201],{},[38,202,203],{},"6. 外链情况",[11,205,206],{},"外链总数、外链域名数、外链质量分布。不要只看数量,看质量。低质外链多了是负资产。",[15,208,210],{"id":209},"怎么判断-seo-公司在摸鱼","怎么判断 SEO 公司在摸鱼",[11,212,213],{},"几个明显的信号:",[11,215,216],{},[38,217,218],{},"1. 报告一直只有排名表",[11,220,221],{},"3 个月了报告还只有关键词排名表,没有流量数据、没有转化数据、没有收录数据。基本是在用排名充数。",[11,223,224],{},[38,225,226],{},"2. 流量涨了转化没涨",[11,228,229],{},"3 个月流量涨 30%,转化没动。要么是宽泛词流量,要么是承接页有问题。SEO 公司两个都不解决,就是摸鱼。",[11,231,232],{},[38,233,234],{},"3. 关键词全是 brand",[11,236,237],{},"报告里的关键词 80% 是品牌词。SEO 工作基本没贡献。",[11,239,240],{},[38,241,242],{},"4. 收录页数没涨",[11,244,245],{},"3 个月收录页数没变化,说明没做内容建设。光做技术调整和发外链,效果天花板很低。",[11,247,248],{},[38,249,250],{},"5. 外链来源可疑",[11,252,253],{},"外链来自一堆陌生站点,且站点质量低。可能是买的外链或 PBN。2025 年百度对这种外链识别很严,风险高。",[11,255,256],{},[38,257,258],{},"6. 周报内容千篇一律",[11,260,261],{},"每周报告内容模板化,复制粘贴改几个数字。说明没在做具体分析,在走流程。",[15,263,264],{"id":264},"我们自己的报告怎么写",[11,266,267],{},"我们做 SEO 服务,报告格式写在这里供参考:",[11,269,270,273],{},[38,271,272],{},"月报包含",":",[275,276,277,283,289,295,301,307],"ol",{},[26,278,279,282],{},[38,280,281],{},"核心指标摘要","(1 页):自然流量(brand 和非 brand 分开)、转化数、收录数、外链数,全部环比和同比",[26,284,285,288],{},[38,286,287],{},"关键词排名分布","(1 页):全量关键词按排名区间分布,涨跌词数",[26,290,291,294],{},[38,292,293],{},"Top 涨跌词","(1 页):涨幅最大的 10 个词、跌幅最大的 10 个词,带搜索量",[26,296,297,300],{},[38,298,299],{},"转化漏斗","(1 页):SEO 流量 → 留资 → 成交,各环节转化率",[26,302,303,306],{},[38,304,305],{},"本月工作总结","(1 页):做了什么、效果如何、下月计划",[26,308,309,312],{},[38,310,311],{},"问题和风险","(1 页):技术上发现的问题、算法变化的影响、需要客户配合的事",[11,314,315,273],{},[38,316,317],{},"报告原则",[23,319,320,326,332,338,344],{},[26,321,322,325],{},[38,323,324],{},"数据全透明",":所有数据可查,不挑不藏",[26,327,328,331],{},[38,329,330],{},"brand 和非 brand 分开",":不让 brand 词充数",[26,333,334,337],{},[38,335,336],{},"涨跌都报",":不只报涨的",[26,339,340,343],{},[38,341,342],{},"转化优先",":流量数据放次要位置,转化数据放主要位置",[26,345,346,349],{},[38,347,348],{},"诚实说问题",":发现效果不好就直说,不掩饰",[11,351,352],{},"我们这么做不是因为高尚,是因为客户能看懂。客户看不懂的报告能糊弄一时,糊弄不了一直。SEO 是长期合作,透明是基础。",[15,354,355],{"id":355},"总结",[11,357,358],{},"看 SEO 报告别被图表骗。常见的注水手法:挑好看的关键词、不显示搜索量、用 brand 充数、只看点击不看转化、基准期不对、把 PPC 算进 SEO。真正该看的指标:非 brand 自然流量、关键词排名分布、转化数据、收录情况、外链质量。判断 SEO 公司是否摸鱼:报告是否只有排名表、流量和转化是否脱节、关键词是否全是 brand、收录是否没涨、外链来源是否可疑。",[11,360,361],{},"找 SEO 公司,先看他们的报告样本。报告注水的,合作了也是踩坑。",[11,363,364],{},"想看一份诚实的 SEO 报告长什么样,联系我们,发一份我们做过的真实月报样本(脱敏)给你参考。",{"title":366,"searchDepth":367,"depth":367,"links":368},"",2,[369,370,371,372,373,374,375,376,377,378],{"id":17,"depth":367,"text":18},{"id":44,"depth":367,"text":45},{"id":59,"depth":367,"text":60},{"id":74,"depth":367,"text":75},{"id":103,"depth":367,"text":104},{"id":123,"depth":367,"text":124},{"id":135,"depth":367,"text":135},{"id":209,"depth":367,"text":210},{"id":264,"depth":367,"text":264},{"id":355,"depth":367,"text":355},null,"SEO","2026-01-20","SEO 公司的报告经常注水:挑好看的关键词、不显示搜索量、用品牌词充数。真正该看的指标和判断 SEO 公司是否摸鱼的方法,以及我们自己的报告怎么写。",false,"md",{},true,"\u002Fblog\u002Fseo-reporting-honest",{"title":5,"description":382},"blog\u002Fseo-reporting-honest",[391,392,393,394],"SEO报告","SEO效果评估","SEO公司选择","数据指标","w4THNozibBUoVIC77bsh5BFaLVtzoePHxHfgS64AoBs",{"id":397,"title":398,"author":6,"body":399,"canonical":379,"category":475,"cover":379,"date":476,"description":477,"draft":383,"extension":384,"keywords":379,"meta":478,"navigation":386,"noindex":383,"ogImage":379,"path":479,"seo":480,"seoDescription":379,"seoTitle":379,"stem":481,"tags":482,"updated":379,"__hash__":486},"blog\u002Fblog\u002Fwebp-avif-image-compression.md","把站内 4200 张图转成 WebP 和 AVIF，我们记下了三个数",{"type":8,"value":400,"toc":467},[401,404,407,410,413,417,420,423,426,429,432,440,443,446,449,452,455,458,461,464],[11,402,403],{},"把站内 4200 张图批量转格式能省多少？我们的实测：PNG 转 WebP 平均小 62%，JPEG 转 WebP 平均小 34%，WebP 再转 AVIF 还能再小 18%，但编码时间多五倍。这是去年 11 月一个设计工具站的项目，数据可以直接拿去估自己的站。",[15,405,406],{"id":406},"先算账再动手",[11,408,409],{},"这个站媒体库 4200 张图：PNG 1400 张，主要是界面截图和图标；JPEG 2800 张，博客配图和案例图。转之前总体积 3.1GB，转成 WebP 之后 1.4GB，平均压掉 55%。",[11,411,412],{},"PNG 那批收益最大，因为里面很多是界面截图，色块多、渐变少，WebP 对这类图特别友好。JPEG 那批收益小一些，因为原图本来就压过一轮。",[15,414,416],{"id":415},"avif-值不值","AVIF 值不值",[11,418,419],{},"AVIF 压缩率确实更好。我们抽 200 张图对比，WebP 平均 96KB，AVIF 平均 79KB。但代价是编码慢：同一台机器转 WebP 用 12 分钟，转 AVIF 要 1 小时，CI 里根本跑不动。",[11,421,422],{},"我的判断是：首屏大图的 AVIF 值得，剩下 95% 的图用 WebP。为了 18% 的体积把构建时间拉长五倍，工程上不划算。",[15,424,425],{"id":425},"三个必踩的坑",[11,427,428],{},"第一个坑是质量参数直接照搬。WebP 的 q=75 和 AVIF 的 q=50 视觉质量接近，硬套同一个数不是太糊就是太大，得按格式分别调。",[11,430,431],{},"第二个坑是透明背景。PNG 转 WebP 支持透明，但有些老工具链会丢掉 alpha 通道，图标变黑底。转完必须抽一批 PNG 图标肉眼过一遍。",[11,433,434,435,439],{},"第三个坑是兼容兜底。AVIF 的浏览器支持已经不错，但旧设备还在用，用 ",[436,437,438],"code",{},"\u003Cpicture>"," 加 WebP 兜底，别只给一个 src。",[15,441,442],{"id":442},"什么图不该转",[11,444,445],{},"三类图建议保留原格式。图标类的小 PNG 和 SVG 本来就几 KB，转格式的收益还不如多一次请求；带大面积纯色文字的截图，WebP 有损模式会把字边缘压糊，我们用无损或者 q=90 才过关；GIF 动图别转 WebP 动图，转成 MP4 更小，实在不行就保留。",[11,447,448],{},"这个站转完第一版有 60 多张博客长图文字发虚，编辑在群里骂了一轮。后来用无损 WebP 重转那批，体积还是比 PNG 小 30%，问题解决。",[15,450,451],{"id":451},"落地方式",[11,453,454],{},"我们没手工转，改的是上传流程：编辑器上传时自动生成 WebP，原图留档。历史图片用脚本批量跑，跑完 grep 一遍 HTML 和数据库里的引用路径。这一步最花时间，因为总有几十张图是硬编码在正文里的。",[11,456,457],{},"这个站改完，移动端首屏图片传输从 1.9MB 降到 620KB。但要说清楚：格式只能解决体积，解决不了\"该不该出现在首屏\"。有些站首屏放三张装饰图，转成 AVIF 也救不回来。",[15,459,460],{"id":460},"转完之后要立个规矩",[11,462,463],{},"批量转换是一次性的，不立规矩三个月就反弹。我们给这个站加了两条：编辑器上传时超过 500KB 的图自动压缩并生成 WebP；每周跑一次巡检，把超过 300KB 的图列出来给内容团队。",[11,465,466],{},"巡检脚本不到 40 行，按页面路径输出图片体积排名，谁传了大图一目了然。这条规矩执行三个月，媒体库平均体积又降了 11%。",{"title":366,"searchDepth":367,"depth":367,"links":468},[469,470,471,472,473,474],{"id":406,"depth":367,"text":406},{"id":415,"depth":367,"text":416},{"id":425,"depth":367,"text":425},{"id":442,"depth":367,"text":442},{"id":451,"depth":367,"text":451},{"id":460,"depth":367,"text":460},"技术SEO","2026-01-19","实测 PNG 转 WebP 小 62%、JPEG 转 WebP 小 34%、再转 AVIF 只多小 18% 但编码慢五倍。三个必踩的坑：质量参数照搬、透明通道丢失、兼容兜底。",{},"\u002Fblog\u002Fwebp-avif-image-compression",{"title":398,"description":477},"blog\u002Fwebp-avif-image-compression",[483,484,485],"图片优化","WebP","AVIF","Rj3KS2SZmZe52cWKT5Gq2KLaetP2WoN9_Kjbt3QBzR0",{"id":488,"title":489,"author":6,"body":490,"canonical":379,"category":587,"cover":379,"date":588,"description":589,"draft":383,"extension":384,"keywords":379,"meta":590,"navigation":386,"noindex":383,"ogImage":379,"path":591,"seo":592,"seoDescription":379,"seoTitle":379,"stem":593,"tags":594,"updated":379,"__hash__":598},"blog\u002Fblog\u002Ftopic-backlog-grooming.md","选题池每周维护 30 分钟，超过 60 条就该冻结",{"type":8,"value":491,"toc":576},[492,495,499,502,505,508,511,514,518,521,524,527,530,533,536,539,542,545,548,551,554,557,561,564,567,570,573],[11,493,494],{},"选题池不是待办清单，是会腐烂的库存。我们固定每周五下午花 30 分钟维护，只做三件事：加新词、补状态、删过期。池子上限 60 条，超了就冻结新增。",[15,496,498],{"id":497},"三件事各-10-分钟","三件事各 10 分钟",[11,500,501],{},"前 10 分钟把本周从搜索后台、工单、销售记录里冒出来的词加进去。中间 10 分钟补状态，把卡住的题标出来。最后 10 分钟从池子里捞三条进下周的排期。",[11,503,504],{},"30 分钟之外不做任何整理。整理得越漂亮，越说明没人真用它写东西。",[15,506,507],{"id":507},"状态只有五种",[11,509,510],{},"待定、待资料、可写、写作中、已发。",[11,512,513],{},"一个题卡在“待资料”超过三周，就要么降级、要么删掉。资源拿不到就是拿不到，挂着只会让人以为还有事没做。",[15,515,517],{"id":516},"池子超过-60-条就该冻结","池子超过 60 条就该冻结",[11,519,520],{},"2025 年 9 月我们的池子涨到 210 条，两个月里没人打开，因为打开也不知道从哪条开始。",[11,522,523],{},"后来砍到 45 条，只留能在一个季度内写完的。第一个月就有人主动从里面捞题，这是之前没发生过的。",[15,525,526],{"id":526},"什么时候删",[11,528,529],{},"三个月没人认领的删。产品功能已下线的删。竞品已经把这件事做透、我们赢不了的删。跟核心词不沾边的删。",[11,531,532],{},"删掉不等于丢失。我们留一个归档表，只写词和删除原因，翻起来比在池子里混着找快。",[15,534,535],{"id":535},"谁维护",[11,537,538],{},"一个人。别搞轮流，轮流等于没人负责。",[11,540,541],{},"我们现在是内容岗的人自己维护，每周五下班前 30 分钟，写在日历上，不做其他事。",[15,543,544],{"id":544},"一个配套的小约定",[11,546,547],{},"新题进池子必须带一句“这篇的独家信息是什么”。写不出来就不进池子。",[15,549,550],{"id":550},"池子和排期的关系",[11,552,553],{},"池子是候选，排期是承诺。只有进了排期的题才需要定发布日，池子里的题不需要日期。",[11,555,556],{},"我们以前给池子里每条都写了计划发布日期，结果日期一路往后推，三个月后整张表全是过期的日期，看着就丧气。",[15,558,560],{"id":559},"三个来源每周各看一遍","三个来源，每周各看一遍",[11,562,563],{},"搜索后台的新查询，工单里的新问题，销售在群里转的客户原话。三个来源各花 3 分钟，扫一眼就行，不用逐条记录。",[11,565,566],{},"记的时候只记原话，不记你理解后的版本。原话才是用户在搜索框里会打的字。",[15,568,569],{"id":569},"一个小细节",[11,571,572],{},"池子里的题按意图排序，不按时间排序。写的人打开就知道自己这周该写哪一类，不用再问。",[11,574,575],{},"这句话逼着我们在选题阶段就想清楚能不能写，省掉了后面一半的返工。",{"title":366,"searchDepth":367,"depth":367,"links":577},[578,579,580,581,582,583,584,585,586],{"id":497,"depth":367,"text":498},{"id":507,"depth":367,"text":507},{"id":516,"depth":367,"text":517},{"id":526,"depth":367,"text":526},{"id":535,"depth":367,"text":535},{"id":544,"depth":367,"text":544},{"id":550,"depth":367,"text":550},{"id":559,"depth":367,"text":560},{"id":569,"depth":367,"text":569},"内容运营","2026-01-21","选题池不是待办清单，是会腐烂的库存。我们固定每周五花 30 分钟做三件事：加新词、补状态、删过期。池子从 210 条砍到 45 条之后，才真正有人用它。",{},"\u002Fblog\u002Ftopic-backlog-grooming",{"title":489,"description":589},"blog\u002Ftopic-backlog-grooming",[595,596,597],"选题机制","内容排期","工作流","DeVBh5sv2s5KU1Z0a4O7Apl8DnlS4JKMetQFkywqbWU",1794088109054]