[{"data":1,"prerenderedAt":417},["ShallowReactive",2],{"blog-mobile-ad-attribution-guide-2026-zh-CN":3},{"id":4,"title":5,"excerpt":6,"content":7,"coverImage":367,"meta":375,"site":378,"status":390,"slug":391,"author":392,"category":403,"publishDate":19,"featured":303,"updatedAt":412,"createdAt":413,"contentHtml":414,"previewUrl":415,"localeSlugs":416},164,"移动广告归因：2026 衡量 App 安装的完整指南","移动广告归因把每次安装追溯回真正带来它的广告。本文讲清归因模型、SKAdNetwork、MMP，以及怎样搭一套你信得过的移动广告归因体系。",{"root":8},{"children":9,"direction":19,"format":15,"indent":13,"type":366,"version":18},[10,22,29,33,37,41,67,71,75,79,184,190,194,198,225,229,233,238,246,250,254,282,286,290,294,317,321,326,330,334,338,342,346,350,354,358],{"children":11,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":21},[12],{"detail":13,"format":13,"mode":14,"style":15,"text":16,"type":17,"version":18},0,"normal","","什么是移动广告归因？","text",1,null,"heading","h2",{"children":23,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":18,"textStyle":15},[24,26],{"detail":13,"format":18,"mode":14,"style":15,"text":25,"type":17,"version":18},"移动广告归因，就是把一次 App 安装、或一次应用内行为，追溯回真正带来它的那条广告、campaign、渠道或素材的过程。",{"detail":13,"format":13,"mode":14,"style":15,"text":27,"type":17,"version":18}," 没有归因，你只知道花了钱、也知道有安装发生——却分不清到底是哪条广告挣来的。有了归因，你能看清：昨天 40% 的安装来自某条 TikTok 视频，成本只有 Google campaign 的一半——于是据此调预算。","paragraph",{"children":30,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[31],{"detail":13,"format":13,"mode":14,"style":15,"text":32,"type":17,"version":18},"这篇指南讲清楚 2026 年移动广告归因怎么运作、背后的模型和标识符、隐私变化如何打破了旧玩法，以及怎样搭一套你真正信得过的度量体系。",{"children":34,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":21},[35],{"detail":13,"format":13,"mode":14,"style":15,"text":36,"type":17,"version":18},"为什么移动广告归因比网页归因更难",{"children":38,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[39],{"detail":13,"format":13,"mode":14,"style":15,"text":40,"type":17,"version":18},"在网页上，一个 cookie 加一个像素就能跟着用户从点击广告一路走到转化。移动端不一样，有三个结构性原因：",{"children":42,"direction":19,"format":15,"indent":13,"type":64,"version":18,"listType":65,"start":18,"tag":66},[43,50,57],{"children":44,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":18},[45,47],{"detail":13,"format":18,"mode":14,"style":15,"text":46,"type":17,"version":18},"点击和安装发生在不同地方。",{"detail":13,"format":13,"mode":14,"style":15,"text":48,"type":17,"version":18}," 用户在一个 App 里点广告，被送到 App Store 或 Google Play，之后再打开你的 App——这一跳之间没有共享的 cookie。","listitem",{"children":51,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":56},[52,54],{"detail":13,"format":18,"mode":14,"style":15,"text":53,"type":17,"version":18},"隐私框架限制了设备标识符。",{"detail":13,"format":13,"mode":14,"style":15,"text":55,"type":17,"version":18}," 苹果的 App 跟踪透明度（ATT）把 IDFA 变成需要用户主动同意，而多数用户会拒绝；安卓的 Privacy Sandbox 也在往同一方向走。",2,{"children":58,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":63},[59,61],{"detail":13,"format":18,"mode":14,"style":15,"text":60,"type":17,"version":18},"延迟深度链接（deferred deep linking）。",{"detail":13,"format":13,"mode":14,"style":15,"text":62,"type":17,"version":18}," 用户可能点了广告、两天后才安装，你仍然要把这次安装归因到那条广告——并把他导向正确的应用内页面——跨越网页从来不用处理的这段延迟。",3,"list","bullet","ul",{"children":68,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[69],{"detail":13,"format":13,"mode":14,"style":15,"text":70,"type":17,"version":18},"正是这些约束，让「一套专门的归因体系」而非「平台自报的原始数字」成为诚实度量的基础。",{"children":72,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":21},[73],{"detail":13,"format":13,"mode":14,"style":15,"text":74,"type":17,"version":18},"核心归因模型",{"children":76,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[77],{"detail":13,"format":13,"mode":14,"style":15,"text":78,"type":17,"version":18},"归因不是单一方法，而是「怎么分配功劳」的选择。主要模型：",{"children":80,"direction":19,"format":15,"indent":13,"type":183,"version":18},[81,103,123,143,163],{"children":82,"direction":19,"format":15,"indent":13,"type":102,"version":18},[83,90,96],{"children":84,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":18,"rowSpan":18},[85],{"children":86,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[87],{"detail":13,"format":13,"mode":14,"style":15,"text":88,"type":17,"version":18}," 模型 ","tablecell",{"children":91,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":18,"rowSpan":18},[92],{"children":93,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[94],{"detail":13,"format":13,"mode":14,"style":15,"text":95,"type":17,"version":18}," 功劳如何分配 ",{"children":97,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":18,"rowSpan":18},[98],{"children":99,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[100],{"detail":13,"format":13,"mode":14,"style":15,"text":101,"type":17,"version":18}," 最适合 ","tablerow",{"children":104,"direction":19,"format":15,"indent":13,"type":102,"version":18},[105,111,117],{"children":106,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[107],{"children":108,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[109],{"detail":13,"format":13,"mode":14,"style":15,"text":110,"type":17,"version":18}," 末次点击 ",{"children":112,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[113],{"children":114,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[115],{"detail":13,"format":13,"mode":14,"style":15,"text":116,"type":17,"version":18}," 100% 给安装前最后一个触点 ",{"children":118,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[119],{"children":120,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[121],{"detail":13,"format":13,"mode":14,"style":15,"text":122,"type":17,"version":18}," 简单漏斗、直效响应 campaign ",{"children":124,"direction":19,"format":15,"indent":13,"type":102,"version":18},[125,131,137],{"children":126,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[127],{"children":128,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[129],{"detail":13,"format":13,"mode":14,"style":15,"text":130,"type":17,"version":18}," 首次点击 ",{"children":132,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[133],{"children":134,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[135],{"detail":13,"format":13,"mode":14,"style":15,"text":136,"type":17,"version":18}," 100% 给第一个触点 ",{"children":138,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[139],{"children":140,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[141],{"detail":13,"format":13,"mode":14,"style":15,"text":142,"type":17,"version":18}," 理解漏斗顶端的发现路径 ",{"children":144,"direction":19,"format":15,"indent":13,"type":102,"version":18},[145,151,157],{"children":146,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[147],{"children":148,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[149],{"detail":13,"format":13,"mode":14,"style":15,"text":150,"type":17,"version":18}," 多触点 ",{"children":152,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[153],{"children":154,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[155],{"detail":13,"format":13,"mode":14,"style":15,"text":156,"type":17,"version":18}," 在所有触点之间分摊 ",{"children":158,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[159],{"children":160,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[161],{"detail":13,"format":13,"mode":14,"style":15,"text":162,"type":17,"version":18}," 路径长的多渠道 campaign ",{"children":164,"direction":19,"format":15,"indent":13,"type":102,"version":18},[165,171,177],{"children":166,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[167],{"children":168,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[169],{"detail":13,"format":13,"mode":14,"style":15,"text":170,"type":17,"version":18}," 数据驱动 ",{"children":172,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[173],{"children":174,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[175],{"detail":13,"format":13,"mode":14,"style":15,"text":176,"type":17,"version":18}," 按贡献算法加权 ",{"children":178,"direction":19,"format":15,"indent":13,"type":89,"version":18,"backgroundColor":19,"colSpan":18,"headerState":13,"rowSpan":18},[179],{"children":180,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[181],{"detail":13,"format":13,"mode":14,"style":15,"text":182,"type":17,"version":18}," 数据量足够大的账户 ","table",{"children":185,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":18,"textStyle":15},[186,188],{"detail":13,"format":18,"mode":14,"style":15,"text":187,"type":17,"version":18},"实操建议：",{"detail":13,"format":13,"mode":14,"style":15,"text":189,"type":17,"version":18}," 多数投放团队从末次点击起步，因为它无歧义，也和广告平台向你计费的口径一致。等量级够大、能看清渠道之间是互相助攻而非互相抢功时，再叠加多触点或数据驱动模型。",{"children":191,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":21},[192],{"detail":13,"format":13,"mode":14,"style":15,"text":193,"type":17,"version":18},"2026 年的标识符与框架",{"children":195,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[196],{"detail":13,"format":13,"mode":14,"style":15,"text":197,"type":17,"version":18},"进入隐私时代，移动广告归因依赖的是一组信号的组合，而不是某个通用 ID：",{"children":199,"direction":19,"format":15,"indent":13,"type":64,"version":18,"listType":65,"start":18,"tag":66},[200,206,212,218],{"children":201,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":18},[202,204],{"detail":13,"format":18,"mode":14,"style":15,"text":203,"type":17,"version":18},"SKAdNetwork（SKAN）与 AdAttributionKit",{"detail":13,"format":13,"mode":14,"style":15,"text":205,"type":17,"version":18},"——苹果的隐私保护框架以聚合方式上报转化，带有刻意的延迟和有限的粒度。你拿到的是 campaign 级的真相，而非用户级。",{"children":207,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":56},[208,210],{"detail":13,"format":18,"mode":14,"style":15,"text":209,"type":17,"version":18},"Google Play 安装来源（Install Referrer）",{"detail":13,"format":13,"mode":14,"style":15,"text":211,"type":17,"version":18},"——在安卓上传递来源数据，把一次安装追溯回那次点击。",{"children":213,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":63},[214,216],{"detail":13,"format":18,"mode":14,"style":15,"text":215,"type":17,"version":18},"确定性匹配",{"detail":13,"format":13,"mode":14,"style":15,"text":217,"type":17,"version":18},"——当有第一方标识符（已登录用户、邮箱）时，可以确定性地匹配。",{"children":219,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":224},[220,222],{"detail":13,"format":18,"mode":14,"style":15,"text":221,"type":17,"version":18},"概率建模",{"detail":13,"format":13,"mode":14,"style":15,"text":223,"type":17,"version":18},"——在缺少确定性信号处，用统计推断在各平台规则内补齐缺口。",4,{"children":226,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[227],{"detail":13,"format":13,"mode":14,"style":15,"text":228,"type":17,"version":18},"结论是：你不再能对每个用户都拿到一条从点击到安装的干净确定性链路。现代归因是「能确定则确定、需聚合与建模则聚合与建模」的混合——而一个好的度量伙伴，会把这些信号调和成一个统一视图。",{"children":230,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":21},[231],{"detail":13,"format":13,"mode":14,"style":15,"text":232,"type":17,"version":18},"移动度量伙伴（MMP）",{"children":234,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":18,"textStyle":15},[235,236],{"detail":13,"format":18,"mode":14,"style":15,"text":232,"type":17,"version":18},{"detail":13,"format":13,"mode":14,"style":15,"text":237,"type":17,"version":18}," 是一个中立的第三方，把每个广告网络的归因信号收在一处，套用一致的规则。它的价值在于中立：TikTok、Google、Meta 各自在自己的看板里认领同一次安装，数字永远对不上。MMP 就是那个去重这些互相抢功声明的「单一事实来源」。",{"children":239,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[240,242,244],{"detail":13,"format":13,"mode":14,"style":15,"text":241,"type":17,"version":18},"无论你用完整的 MMP，还是更轻的第一方方案，原则相同：",{"detail":13,"format":18,"mode":14,"style":15,"text":243,"type":17,"version":18},"在一个中立系统里度量，而不是在每个广告平台自报的看板里度量。",{"detail":13,"format":13,"mode":14,"style":15,"text":245,"type":17,"version":18}," 平台自报数字永远高估，因为每个平台都有动机去认领那次转化。",{"children":247,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":21},[248],{"detail":13,"format":13,"mode":14,"style":15,"text":249,"type":17,"version":18},"搭一套你信得过的归因体系",{"children":251,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[252],{"detail":13,"format":13,"mode":14,"style":15,"text":253,"type":17,"version":18},"一套可靠的移动广告归因栈有四层：",{"children":255,"direction":19,"format":15,"indent":13,"type":64,"version":18,"listType":280,"start":18,"tag":281},[256,262,268,274],{"children":257,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":18},[258,260],{"detail":13,"format":18,"mode":14,"style":15,"text":259,"type":17,"version":18},"SDK 或服务端对接（S2S）",{"detail":13,"format":13,"mode":14,"style":15,"text":261,"type":17,"version":18},"——干净地采集安装和应用内事件，走 SDK 或服务端连接皆可。",{"children":263,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":56},[264,266],{"detail":13,"format":18,"mode":14,"style":15,"text":265,"type":17,"version":18},"一致的事件口径",{"detail":13,"format":13,"mode":14,"style":15,"text":267,"type":17,"version":18},"——把关键事件（安装、注册、首购、再触达）定义一次，并在各处用同一套命名。",{"children":269,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":63},[270,272],{"detail":13,"format":18,"mode":14,"style":15,"text":271,"type":17,"version":18},"去重与对账",{"detail":13,"format":13,"mode":14,"style":15,"text":273,"type":17,"version":18},"——把互相冲突的平台声明，收敛成每次转化一条去重后的记录。",{"children":275,"direction":19,"format":15,"indent":13,"type":49,"version":18,"value":224},[276,278],{"detail":13,"format":18,"mode":14,"style":15,"text":277,"type":17,"version":18},"看漏斗下游价值，而非只看安装",{"detail":13,"format":13,"mode":14,"style":15,"text":279,"type":17,"version":18},"——归因到收入和留存，而不只是安装，这样你优化的是「留下来的用户」而非「只是来过的用户」。","number","ol",{"children":283,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[284],{"detail":13,"format":13,"mode":14,"style":15,"text":285,"type":17,"version":18},"第四层，正是大多数账户把钱丢在桌上的地方。安装是成本；一个被留住、被再触达的用户，才是回报。",{"children":287,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":21},[288],{"detail":13,"format":13,"mode":14,"style":15,"text":289,"type":17,"version":18},"从归因到再触达",{"children":291,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[292],{"detail":13,"format":13,"mode":14,"style":15,"text":293,"type":17,"version":18},"归因告诉你哪个渠道挣来了一个用户。下一个问题是：怎么把这个用户召回，而不必再付一次拉新的点击费。这正是度量与再触达的连接点：一旦你能归因安装、并识别出哪些用户流失，就能通过自有渠道把他们赢回来，而不是买第二遍。",{"children":295,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[296,306,308,315],{"children":297,"direction":19,"format":15,"indent":13,"type":300,"version":63,"fields":301,"id":305},[298],{"detail":13,"format":13,"mode":14,"style":15,"text":299,"type":17,"version":18},"DeepClick 的再触达产品","link",{"linkType":302,"newTab":303,"url":304},"custom",false,"/product/re-engagement","6a4c656e70dbbc00c8a63a15",{"detail":13,"format":13,"mode":14,"style":15,"text":307,"type":17,"version":18}," 正是为此而建——用 Web 推送和自有渠道 campaign，把已归因、已安装的用户变成被再触达、会回来的用户。而如果你的拉新漏斗跑在可安装的网页体验上，",{"children":309,"direction":19,"format":15,"indent":13,"type":300,"version":63,"fields":312,"id":314},[310],{"detail":13,"format":13,"mode":14,"style":15,"text":311,"type":17,"version":18},"PWA 安装产品",{"linkType":302,"newTab":303,"url":313},"/product/pwa-install","6a4c656e70dbbc00c8a63a16",{"detail":13,"format":13,"mode":14,"style":15,"text":316,"type":17,"version":18}," 会从漏斗顶端喂给这同一个闭环。归因证明什么有效，再触达把它复利放大。",{"children":318,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":21},[319],{"detail":13,"format":13,"mode":14,"style":15,"text":320,"type":17,"version":18},"常见问题",{"children":322,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":325},[323],{"detail":13,"format":13,"mode":14,"style":15,"text":324,"type":17,"version":18},"移动广告归因和 MMP 有什么区别？","h3",{"children":327,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[328],{"detail":13,"format":13,"mode":14,"style":15,"text":329,"type":17,"version":18},"移动广告归因是概念——把一次转化追溯回造成它的那条广告。移动度量伙伴（MMP）是工具——一个中立的第三方平台，把每个网络的归因信号收齐并套用一致的去重规则，让你拿到一个可信数字，而不是每个平台各自夸大的自报。",{"children":331,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":325},[332],{"detail":13,"format":13,"mode":14,"style":15,"text":333,"type":17,"version":18},"ATT 和 SKAdNetwork 是不是让移动归因失效了？",{"children":335,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[336],{"detail":13,"format":13,"mode":14,"style":15,"text":337,"type":17,"version":18},"没有，但它们改变了归因。苹果的 ATT 把 IDFA 变成需主动同意，SKAdNetwork 把 iOS 归因转向聚合、延迟、campaign 级的上报。你在 iOS 上失去了用户级粒度，但通过组合 SKAN、第一方数据和建模信号，仍能可靠地衡量 campaign 表现。",{"children":339,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":325},[340],{"detail":13,"format":13,"mode":14,"style":15,"text":341,"type":17,"version":18},"广告平台看板里的安装数字该信吗？",{"children":343,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[344],{"detail":13,"format":13,"mode":14,"style":15,"text":345,"type":17,"version":18},"当成被夸大的看。每个广告平台都有动机认领同一次安装，所以 TikTok、Google、Meta 各自报的转化都会比实际发生的多。在一个中立的归因系统里对账，才能拿到去重后的真相。",{"children":347,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":325},[348],{"detail":13,"format":13,"mode":14,"style":15,"text":349,"type":17,"version":18},"什么是延迟深度链接（deferred deep linking）？",{"children":351,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[352],{"detail":13,"format":13,"mode":14,"style":15,"text":353,"type":17,"version":18},"延迟深度链接会归因一次「延迟发生」的安装——用户点了广告、之后才安装，首次打开时被导向广告所推广的那个具体应用内内容，同时这次安装仍被正确归功于那条广告。",{"children":355,"direction":19,"format":15,"indent":13,"type":20,"version":18,"tag":21},[356],{"detail":13,"format":13,"mode":14,"style":15,"text":357,"type":17,"version":18},"关键要点",{"children":359,"direction":19,"format":15,"indent":13,"type":28,"version":18,"textFormat":13,"textStyle":15},[360,362,364],{"detail":13,"format":13,"mode":14,"style":15,"text":361,"type":17,"version":18},"2026 年的移动广告归因，讲的是在隐私约束下的诚实：",{"detail":13,"format":18,"mode":14,"style":15,"text":363,"type":17,"version":18},"在一个中立系统里度量，混合确定性与建模信号，归因漏斗下游价值而非原始安装。",{"detail":13,"format":13,"mode":14,"style":15,"text":365,"type":17,"version":18}," 选一个和你漏斗匹配的归因模型，把平台自报数字当成被夸大的，并通过把已归因用户变成被再触达的用户来闭环。这就是广告花费变成可度量、站得住脚的增长的方式。","root",{"id":368,"alt":369,"updatedAt":370,"createdAt":370,"url":371,"thumbnailURL":19,"filename":372,"mimeType":373,"filesize":374,"width":19,"height":19},326,"Abstract illustration of mobile ad attribution: flow lines connecting an ad to an app install, funnel and analytics motifs","2026-07-07T02:31:44.221Z","https://cms-r2.deepclick.com/gpt_1783391484754_0-7b6f4394df43.png","gpt_1783391484754_0-7b6f4394df43.png","application/octet-stream",1545760,{"title":5,"description":376,"image":377},"移动广告归因怎么做？2026 讲清归因模型、SKAdNetwork 与 ATT、MMP，以及如何在一个中立系统里衡量安装，拿到你信得过的去重真相。",{"id":368,"alt":369,"updatedAt":370,"createdAt":370,"url":371,"thumbnailURL":19,"filename":372,"mimeType":373,"filesize":374,"width":19,"height":19},{"id":18,"key":379,"name":380,"prodHost":381,"testHost":382,"blogPath":383,"docPath":384,"zhPrefix":385,"deployHookTest":386,"deployHookProd":387,"enabled":388,"updatedAt":389,"createdAt":389},"deepclick","DeepClick","https://deepclick.com","https://www-test-deepclick.qiliangjia.one","/resources/blog/{slug}","/docs/{slug}","/zh-CN","https://api.cloudflare.com/client/v4/pages/webhooks/deploy_hooks/60a9adef-153b-4c89-8d07-7118e91e9522","https://api.cloudflare.com/client/v4/pages/webhooks/deploy_hooks/05323321-f694-4ce8-a5af-173c507b8bae",true,"2026-07-14T06:26:38.962Z","published","mobile-ad-attribution-guide-2026",{"id":56,"name":380,"avatar":393,"updatedAt":401,"createdAt":402},{"id":394,"alt":380,"updatedAt":395,"createdAt":395,"url":396,"thumbnailURL":19,"filename":397,"mimeType":398,"filesize":399,"width":400,"height":400},25,"2026-04-22T08:09:22.606Z","https://cms-r2.deepclick.com/头像-白.png","头像-白.png","image/png",26626,1024,"2026-04-22T08:09:35.299Z","2026-04-22T06:42:49.116Z",{"id":404,"site":405,"titleZh":406,"titleEn":407,"slug":408,"order":409,"updatedAt":410,"createdAt":411},7,{"id":18,"key":379,"name":380,"prodHost":381,"testHost":382,"blogPath":383,"docPath":384,"zhPrefix":385,"deployHookTest":386,"deployHookProd":387,"enabled":388,"updatedAt":389,"createdAt":389},"技术导航","Tech Guides","tech-guides",5,"2026-04-27T08:37:10.576Z","2026-04-23T02:59:13.436Z","2026-07-07T02:33:24.603Z","2026-07-07T02:33:07.123Z","\u003Cdiv class=\"payload-richtext\">\u003Ch2>什么是移动广告归因？\u003C/h2>\u003Cp>\u003Cstrong>移动广告归因，就是把一次 App 安装、或一次应用内行为，追溯回真正带来它的那条广告、campaign、渠道或素材的过程。\u003C/strong> 没有归因，你只知道花了钱、也知道有安装发生——却分不清到底是哪条广告挣来的。有了归因，你能看清：昨天 40% 的安装来自某条 TikTok 视频，成本只有 Google campaign 的一半——于是据此调预算。\u003C/p>\u003Cp>这篇指南讲清楚 2026 年移动广告归因怎么运作、背后的模型和标识符、隐私变化如何打破了旧玩法，以及怎样搭一套你真正信得过的度量体系。\u003C/p>\u003Ch2>为什么移动广告归因比网页归因更难\u003C/h2>\u003Cp>在网页上，一个 cookie 加一个像素就能跟着用户从点击广告一路走到转化。移动端不一样，有三个结构性原因：\u003C/p>\u003Cul class=\"list-bullet\">\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"1\"\n        >\u003Cstrong>点击和安装发生在不同地方。\u003C/strong> 用户在一个 App 里点广告，被送到 App Store 或 Google Play，之后再打开你的 App——这一跳之间没有共享的 cookie。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"2\"\n        >\u003Cstrong>隐私框架限制了设备标识符。\u003C/strong> 苹果的 App 跟踪透明度（ATT）把 IDFA 变成需要用户主动同意，而多数用户会拒绝；安卓的 Privacy Sandbox 也在往同一方向走。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"3\"\n        >\u003Cstrong>延迟深度链接（deferred deep linking）。\u003C/strong> 用户可能点了广告、两天后才安装，你仍然要把这次安装归因到那条广告——并把他导向正确的应用内页面——跨越网页从来不用处理的这段延迟。\u003C/li>\u003C/ul>\u003Cp>正是这些约束，让「一套专门的归因体系」而非「平台自报的原始数字」成为诚实度量的基础。\u003C/p>\u003Ch2>核心归因模型\u003C/h2>\u003Cp>归因不是单一方法，而是「怎么分配功劳」的选择。主要模型：\u003C/p>\u003Cdiv class=\"lexical-table-container\">\n        \u003Ctable class=\"lexical-table\" style=\"border-collapse: collapse;\">\n          \u003Ctbody>\u003Ctr class=\"lexical-table-row\">\n        \u003Cth\n        class=\"lexical-table-cell lexical-table-cell-header-1\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 模型 \u003C/p>\n      \u003C/th>\n    \u003Cth\n        class=\"lexical-table-cell lexical-table-cell-header-1\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 功劳如何分配 \u003C/p>\n      \u003C/th>\n    \u003Cth\n        class=\"lexical-table-cell lexical-table-cell-header-1\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 最适合 \u003C/p>\n      \u003C/th>\n    \n      \u003C/tr>\u003Ctr class=\"lexical-table-row\">\n        \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 末次点击 \u003C/p>\n      \u003C/td>\n    \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 100% 给安装前最后一个触点 \u003C/p>\n      \u003C/td>\n    \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 简单漏斗、直效响应 campaign \u003C/p>\n      \u003C/td>\n    \n      \u003C/tr>\u003Ctr class=\"lexical-table-row\">\n        \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 首次点击 \u003C/p>\n      \u003C/td>\n    \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 100% 给第一个触点 \u003C/p>\n      \u003C/td>\n    \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 理解漏斗顶端的发现路径 \u003C/p>\n      \u003C/td>\n    \n      \u003C/tr>\u003Ctr class=\"lexical-table-row\">\n        \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 多触点 \u003C/p>\n      \u003C/td>\n    \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 在所有触点之间分摊 \u003C/p>\n      \u003C/td>\n    \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 路径长的多渠道 campaign \u003C/p>\n      \u003C/td>\n    \n      \u003C/tr>\u003Ctr class=\"lexical-table-row\">\n        \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 数据驱动 \u003C/p>\n      \u003C/td>\n    \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 按贡献算法加权 \u003C/p>\n      \u003C/td>\n    \u003Ctd\n        class=\"lexical-table-cell lexical-table-cell-header-0\"\n        \n        \n        style=\"border: 1px solid #ccc; padding: 8px;\"\n      >\n        \u003Cp> 数据量足够大的账户 \u003C/p>\n      \u003C/td>\n    \n      \u003C/tr>\u003C/tbody>\n        \u003C/table>\n      \u003C/div>\u003Cp>\u003Cstrong>实操建议：\u003C/strong> 多数投放团队从末次点击起步，因为它无歧义，也和广告平台向你计费的口径一致。等量级够大、能看清渠道之间是互相助攻而非互相抢功时，再叠加多触点或数据驱动模型。\u003C/p>\u003Ch2>2026 年的标识符与框架\u003C/h2>\u003Cp>进入隐私时代，移动广告归因依赖的是一组信号的组合，而不是某个通用 ID：\u003C/p>\u003Cul class=\"list-bullet\">\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"1\"\n        >\u003Cstrong>SKAdNetwork（SKAN）与 AdAttributionKit\u003C/strong>——苹果的隐私保护框架以聚合方式上报转化，带有刻意的延迟和有限的粒度。你拿到的是 campaign 级的真相，而非用户级。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"2\"\n        >\u003Cstrong>Google Play 安装来源（Install Referrer）\u003C/strong>——在安卓上传递来源数据，把一次安装追溯回那次点击。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"3\"\n        >\u003Cstrong>确定性匹配\u003C/strong>——当有第一方标识符（已登录用户、邮箱）时，可以确定性地匹配。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"4\"\n        >\u003Cstrong>概率建模\u003C/strong>——在缺少确定性信号处，用统计推断在各平台规则内补齐缺口。\u003C/li>\u003C/ul>\u003Cp>结论是：你不再能对每个用户都拿到一条从点击到安装的干净确定性链路。现代归因是「能确定则确定、需聚合与建模则聚合与建模」的混合——而一个好的度量伙伴，会把这些信号调和成一个统一视图。\u003C/p>\u003Ch2>移动度量伙伴（MMP）\u003C/h2>\u003Cp>\u003Cstrong>移动度量伙伴（MMP）\u003C/strong> 是一个中立的第三方，把每个广告网络的归因信号收在一处，套用一致的规则。它的价值在于中立：TikTok、Google、Meta 各自在自己的看板里认领同一次安装，数字永远对不上。MMP 就是那个去重这些互相抢功声明的「单一事实来源」。\u003C/p>\u003Cp>无论你用完整的 MMP，还是更轻的第一方方案，原则相同：\u003Cstrong>在一个中立系统里度量，而不是在每个广告平台自报的看板里度量。\u003C/strong> 平台自报数字永远高估，因为每个平台都有动机去认领那次转化。\u003C/p>\u003Ch2>搭一套你信得过的归因体系\u003C/h2>\u003Cp>一套可靠的移动广告归因栈有四层：\u003C/p>\u003Col class=\"list-number\">\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"1\"\n        >\u003Cstrong>SDK 或服务端对接（S2S）\u003C/strong>——干净地采集安装和应用内事件，走 SDK 或服务端连接皆可。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"2\"\n        >\u003Cstrong>一致的事件口径\u003C/strong>——把关键事件（安装、注册、首购、再触达）定义一次，并在各处用同一套命名。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"3\"\n        >\u003Cstrong>去重与对账\u003C/strong>——把互相冲突的平台声明，收敛成每次转化一条去重后的记录。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"4\"\n        >\u003Cstrong>看漏斗下游价值，而非只看安装\u003C/strong>——归因到收入和留存，而不只是安装，这样你优化的是「留下来的用户」而非「只是来过的用户」。\u003C/li>\u003C/ol>\u003Cp>第四层，正是大多数账户把钱丢在桌上的地方。安装是成本；一个被留住、被再触达的用户，才是回报。\u003C/p>\u003Ch2>从归因到再触达\u003C/h2>\u003Cp>归因告诉你哪个渠道挣来了一个用户。下一个问题是：怎么把这个用户召回，而不必再付一次拉新的点击费。这正是度量与再触达的连接点：一旦你能归因安装、并识别出哪些用户流失，就能通过自有渠道把他们赢回来，而不是买第二遍。\u003C/p>\u003Cp>\u003Ca href=\"/product/re-engagement\">DeepClick 的再触达产品\u003C/a> 正是为此而建——用 Web 推送和自有渠道 campaign，把已归因、已安装的用户变成被再触达、会回来的用户。而如果你的拉新漏斗跑在可安装的网页体验上，\u003Ca href=\"/product/pwa-install\">PWA 安装产品\u003C/a> 会从漏斗顶端喂给这同一个闭环。归因证明什么有效，再触达把它复利放大。\u003C/p>\u003Ch2>常见问题\u003C/h2>\u003Ch3>移动广告归因和 MMP 有什么区别？\u003C/h3>\u003Cp>移动广告归因是概念——把一次转化追溯回造成它的那条广告。移动度量伙伴（MMP）是工具——一个中立的第三方平台，把每个网络的归因信号收齐并套用一致的去重规则，让你拿到一个可信数字，而不是每个平台各自夸大的自报。\u003C/p>\u003Ch3>ATT 和 SKAdNetwork 是不是让移动归因失效了？\u003C/h3>\u003Cp>没有，但它们改变了归因。苹果的 ATT 把 IDFA 变成需主动同意，SKAdNetwork 把 iOS 归因转向聚合、延迟、campaign 级的上报。你在 iOS 上失去了用户级粒度，但通过组合 SKAN、第一方数据和建模信号，仍能可靠地衡量 campaign 表现。\u003C/p>\u003Ch3>广告平台看板里的安装数字该信吗？\u003C/h3>\u003Cp>当成被夸大的看。每个广告平台都有动机认领同一次安装，所以 TikTok、Google、Meta 各自报的转化都会比实际发生的多。在一个中立的归因系统里对账，才能拿到去重后的真相。\u003C/p>\u003Ch3>什么是延迟深度链接（deferred deep linking）？\u003C/h3>\u003Cp>延迟深度链接会归因一次「延迟发生」的安装——用户点了广告、之后才安装，首次打开时被导向广告所推广的那个具体应用内内容，同时这次安装仍被正确归功于那条广告。\u003C/p>\u003Ch2>关键要点\u003C/h2>\u003Cp>2026 年的移动广告归因，讲的是在隐私约束下的诚实：\u003Cstrong>在一个中立系统里度量，混合确定性与建模信号，归因漏斗下游价值而非原始安装。\u003C/strong> 选一个和你漏斗匹配的归因模型，把平台自报数字当成被夸大的，并通过把已归因用户变成被再触达的用户来闭环。这就是广告花费变成可度量、站得住脚的增长的方式。\u003C/p>\u003C/div>","https://deepclick.com/zh-CN/resources/blog/mobile-ad-attribution-guide-2026",{"zh-CN":391,"en":391},1785922851037]