[{"data":1,"prerenderedAt":395},["ShallowReactive",2],{"blog-facebook-campaign-budget-optimization-zh-CN":3},{"id":4,"title":5,"excerpt":6,"content":7,"coverImage":336,"meta":344,"site":348,"status":369,"slug":370,"author":371,"category":381,"publishDate":19,"featured":130,"updatedAt":390,"createdAt":391,"contentHtml":392,"previewUrl":393,"localeSlugs":394},579,"Facebook 广告系列预算优化（CBO）：Meta 怎么分配预算，什么时候它在跟你作对","CBO 优化的是广告系列的总结果，不是各广告组之间的公平——这就是为什么一个广告组会把预算吃光，而且多数时候这是对的。分配机制、你仍然握着的控制手段、学习期陷阱，以及四种不该用 CBO 的场景。",{"root":8},{"children":9,"direction":19,"format":15,"indent":13,"type":335,"version":18},[10,21,29,35,39,58,62,66,70,74,80,88,96,100,104,110,116,135,141,145,149,153,174,178,182,196,200,227,231,235,270,274,278,283,287,291,295,299,303,307,311,315,319,323,327,331],{"children":11,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[12],{"detail":13,"format":13,"mode":14,"style":15,"text":16,"type":17,"version":18},0,"normal","","Facebook 广告系列预算优化（CBO）把预算从广告组挪到广告系列这一层，由 Meta 决定每个广告组分到多少。Meta 现在把它叫作 Advantage 广告系列预算，这次改名带来的困惑比这个功能本身还多——底下是同一套机制，只是换了个新标签。","text",1,null,"paragraph",{"children":22,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[23,25,27],{"detail":13,"format":13,"mode":14,"style":15,"text":24,"type":17,"version":18},"它之所以能引出这么多求助帖，是因为它严格按照它宣称的方式工作，而它宣称的方式并不是大多数投放者预期的：",{"detail":13,"format":18,"mode":14,"style":15,"text":26,"type":17,"version":18},"它优化的是整个广告系列的总结果，不是把你每个广告组都喂饱。",{"detail":13,"format":13,"mode":14,"style":15,"text":28,"type":17,"version":18}," 一旦接受这句话，几乎所有看起来奇怪的 CBO 行为都变得可以预测。",{"children":30,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":34},[31],{"detail":13,"format":13,"mode":14,"style":15,"text":32,"type":17,"version":18},"广告系列预算优化到底控制什么","heading","h2",{"children":36,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[37],{"detail":13,"format":13,"mode":14,"style":15,"text":38,"type":17,"version":18},"预算只可能待在两个位置：",{"children":40,"direction":19,"format":15,"indent":13,"type":55,"version":18,"listType":56,"start":18,"tag":57},[41,48],{"children":42,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":18},[43,45],{"detail":13,"format":18,"mode":14,"style":15,"text":44,"type":17,"version":18},"广告组预算（ABO）",{"detail":13,"format":13,"mode":14,"style":15,"text":46,"type":17,"version":18},"——每个广告组有自己的预算，各花各的。分配由你的结构决定。","listitem",{"children":49,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":54},[50,52],{"detail":13,"format":18,"mode":14,"style":15,"text":51,"type":17,"version":18},"广告系列预算（CBO / Advantage 广告系列预算）",{"detail":13,"format":13,"mode":14,"style":15,"text":53,"type":17,"version":18},"——整个广告系列一份预算，投放系统持续地在各广告组之间分配。",2,"list","bullet","ul",{"children":59,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[60],{"detail":13,"format":13,"mode":14,"style":15,"text":61,"type":17,"version":18},"差别仅此而已。其余的一切——目标、定向、出价、素材——两种模式下的工作方式完全一样。预算放在哪一层不是策略，而是一个关于「谁来决定分配」的决定：你，还是系统。",{"children":63,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[64],{"detail":13,"format":13,"mode":14,"style":15,"text":65,"type":17,"version":18},"这个分配不是每天开始时切一刀就完了，它是持续进行的。系统对每一次机会估计：哪个广告组更可能以什么成本产出一个结果，然后把花费送到当下预期回报最好的地方。这个估计随数据变化而更新——这就是为什么你中午看到的分配比例，和半夜看到的可能完全不像。",{"children":67,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":34},[68],{"detail":13,"format":13,"mode":14,"style":15,"text":69,"type":17,"version":18},"为什么一个广告组把预算吃光了",{"children":71,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[72],{"detail":13,"format":13,"mode":14,"style":15,"text":73,"type":17,"version":18},"这是关于 CBO 最常见的抱怨，而多数情况下，系统正在做对的事，只是看起来像坏了。",{"children":75,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[76,78],{"detail":13,"format":13,"mode":14,"style":15,"text":77,"type":17,"version":18},"如果广告组 A 能以比 B 更低的成本产出结果，那么把钱花在 A 上，整个广告系列每块钱能拿到更多结果。把大部分预算送给 A 不是 bug——那正是你选的目标。",{"detail":13,"format":18,"mode":14,"style":15,"text":79,"type":17,"version":18},"分配不均本身就是优化结果。",{"children":81,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[82,84,86],{"detail":13,"format":13,"mode":14,"style":15,"text":83,"type":17,"version":18},"它在一种特定情形下才变成真问题：当你的各个广告组",{"detail":13,"format":18,"mode":14,"style":15,"text":85,"type":17,"version":18},"对生意的价值并不相等",{"detail":13,"format":13,"mode":14,"style":15,"text":87,"type":17,"version":18},"时。系统优化的是你让它优化的那个指标，它并不知道那个便宜的受众之所以便宜，是因为它转化的是低客单产品；也不知道那个贵的受众是你明年的战略市场。如果各广告组的单位经济不同，而这个差异没有体现在优化事件里，CBO 会非常自信地把钱分到与你的利益相反的方向上。",{"children":89,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[90,92,94],{"detail":13,"format":13,"mode":14,"style":15,"text":91,"type":17,"version":18},"由此得到一条好用的判据：",{"detail":13,"format":18,"mode":14,"style":15,"text":93,"type":17,"version":18},"当广告系列里所有广告组「每个结果对你的价值相同」时，CBO 才合适。",{"detail":13,"format":13,"mode":14,"style":15,"text":95,"type":17,"version":18}," 不相同时，要么把它们拆到不同广告系列，要么把价值差异表达出来让系统看得见。",{"children":97,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":34},[98],{"detail":13,"format":13,"mode":14,"style":15,"text":99,"type":17,"version":18},"你仍然握着的控制手段",{"children":101,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[102],{"detail":13,"format":13,"mode":14,"style":15,"text":103,"type":17,"version":18},"把分配权交出去，不等于把一切都交出去。",{"children":105,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":18,"textStyle":15},[106,108],{"detail":13,"format":18,"mode":14,"style":15,"text":107,"type":17,"version":18},"广告组花费上下限。",{"detail":13,"format":13,"mode":14,"style":15,"text":109,"type":17,"version":18}," 你可以在 CBO 广告系列内给单个广告组设置最低和最高花费。这是「某个广告组被饿死」的直接解法，也是最容易被用过头的工具——把足够多的广告组卡得足够死，你就等于用更多步骤重建了 ABO，同时还保留着 CBO 的缺点。",{"children":111,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":18,"textStyle":15},[112,114],{"detail":13,"format":18,"mode":14,"style":15,"text":113,"type":17,"version":18},"排期。",{"detail":13,"format":13,"mode":14,"style":15,"text":115,"type":17,"version":18}," 广告组层面的排期依然生效，它约束一个广告组在什么时间才有资格拿到花费。",{"children":117,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":18,"textStyle":15},[118,120,122,133],{"detail":13,"format":18,"mode":14,"style":15,"text":119,"type":17,"version":18},"出价策略。",{"detail":13,"format":13,"mode":14,"style":15,"text":121,"type":17,"version":18}," 它设在广告系列层，并直接与分配相互作用。成本目标或出价上限会改变系统愿意购买哪些机会，从而改变 CBO 产出的分配形态。每种选项到底约束了什么，见",{"children":123,"direction":19,"format":15,"indent":13,"type":126,"version":127,"fields":128,"id":132},[124],{"detail":13,"format":13,"mode":14,"style":15,"text":125,"type":17,"version":18},"Facebook 广告出价策略怎么选","link",3,{"linkType":129,"newTab":130,"url":131},"custom",false,"https://deepclick.com/resources/blog/facebook-ads-bidding-strategy/","6a994bfa1b5dab00c8405751",{"detail":13,"format":13,"mode":14,"style":15,"text":134,"type":17,"version":18},"。",{"children":136,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":18,"textStyle":15},[137,139],{"detail":13,"format":18,"mode":14,"style":15,"text":138,"type":17,"version":18},"结构。",{"detail":13,"format":13,"mode":14,"style":15,"text":140,"type":17,"version":18}," 广告系列里放了什么，是最强的控制手段——不存在的广告组拿不到预算，而分在两个广告系列里的广告组不会争同一个池子。",{"children":142,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":34},[143],{"detail":13,"format":13,"mode":14,"style":15,"text":144,"type":17,"version":18},"与学习期的相互作用",{"children":146,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[147],{"detail":13,"format":13,"mode":14,"style":15,"text":148,"type":17,"version":18},"这是 CBO 与账户结构交汇的地方，也是支持它的最有力论据。",{"children":150,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[151],{"detail":13,"format":13,"mode":14,"style":15,"text":152,"type":17,"version":18},"集中有助于广告组积累到离开学习期所需的事件量。把同样的预算和同样的受众切成很多个小广告组，是让它们全部长期吃不饱的经典做法——每个都在永远学习，没有一个稳定下来。一个只有少数几个广告组的 CBO 广告系列倾向于集中花费，这会帮助跑赢的那些更快走出学习期。",{"children":154,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[155,157,164,166,173],{"detail":13,"format":13,"mode":14,"style":15,"text":156,"type":17,"version":18},"陷阱是相反的结构：CBO 配上很多个窄广告组。系统把花费集中到其中一两个，其余的贴着零，你拿到的是两头的坏处——大部分结构上没有数据，而且哪些被饿死也不由你决定。想测很多东西，要么给每个测试足够有意义的预算，要么串行地测。整体结构见",{"children":158,"direction":19,"format":15,"indent":13,"type":126,"version":127,"fields":161,"id":163},[159],{"detail":13,"format":13,"mode":14,"style":15,"text":160,"type":17,"version":18},"Facebook 广告账户结构",{"linkType":129,"newTab":130,"url":162},"https://deepclick.com/resources/blog/facebook-ads-account-structure/","6a994bfa1b5dab00c8405752",{"detail":13,"format":13,"mode":14,"style":15,"text":165,"type":17,"version":18},"，到底什么会重置学习期见",{"children":167,"direction":19,"format":15,"indent":13,"type":126,"version":127,"fields":170,"id":172},[168],{"detail":13,"format":13,"mode":14,"style":15,"text":169,"type":17,"version":18},"Meta 广告学习期",{"linkType":129,"newTab":130,"url":171},"https://deepclick.com/resources/blog/meta-ads-learning-phase/","6a994bfa1b5dab00c8405753",{"detail":13,"format":13,"mode":14,"style":15,"text":134,"type":17,"version":18},{"children":175,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":34},[176],{"detail":13,"format":13,"mode":14,"style":15,"text":177,"type":17,"version":18},"改预算的代价",{"children":179,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[180],{"detail":13,"format":13,"mode":14,"style":15,"text":181,"type":17,"version":18},"改预算不是免费的。改动广告系列预算，就改变了系统一直在据以规划的花费节奏，幅度大的改动可能把广告组打回学习期。从机制而不是从某个具体阈值出发，可以得到这几条做法：",{"children":183,"direction":19,"format":15,"indent":13,"type":55,"version":18,"listType":56,"start":18,"tag":57},[184,188,192],{"children":185,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":18},[186],{"detail":13,"format":13,"mode":14,"style":15,"text":187,"type":17,"version":18},"预算改动分步走，不要一次跳很大；每一步给它时间被观测到，再去判断。",{"children":189,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":54},[190],{"detail":13,"format":13,"mode":14,"style":15,"text":191,"type":17,"version":18},"不要在同一次编辑里既改预算又改结构——效果变化了你也分不清是哪一个造成的。",{"children":193,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":127},[194],{"detail":13,"format":13,"mode":14,"style":15,"text":195,"type":17,"version":18},"判断一次改动，要看一个覆盖完整转化周期的窗口，而不是改完之后那几个小时。",{"children":197,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":34},[198],{"detail":13,"format":13,"mode":14,"style":15,"text":199,"type":17,"version":18},"什么时候不该用 CBO",{"children":201,"direction":19,"format":15,"indent":13,"type":55,"version":18,"listType":56,"start":18,"tag":57},[202,208,214,220],{"children":203,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":18},[204,206],{"detail":13,"format":18,"mode":14,"style":15,"text":205,"type":17,"version":18},"各广告组「每个结果的业务价值」不同时。",{"detail":13,"format":13,"mode":14,"style":15,"text":207,"type":17,"version":18}," 见上文，这是最主要的一条。",{"children":209,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":54},[210,212],{"detail":13,"format":18,"mode":14,"style":15,"text":211,"type":17,"version":18},"需要控制分配的对照测试。",{"detail":13,"format":13,"mode":14,"style":15,"text":213,"type":17,"version":18}," 如果测试的意义就在于每个组拿到可比的花费，那预算层面的控制就是实验本身，CBO 不会替你保住它。",{"children":215,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":127},[216,218],{"detail":13,"format":18,"mode":14,"style":15,"text":217,"type":17,"version":18},"一个广告系列里混着差异很大的漏斗阶段。",{"detail":13,"format":13,"mode":14,"style":15,"text":219,"type":17,"version":18}," 拉新和再营销的每结果成本结构性地不同，系统会可预测地偏向再营销——而再营销自己是长不大的，因为它依赖的受众正是被你饿死的拉新在构建。",{"children":221,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":226},[222,224],{"detail":13,"format":18,"mode":14,"style":15,"text":223,"type":17,"version":18},"预算小到「集中」就意味着永远只有一个广告组在跑。",{"detail":13,"format":13,"mode":14,"style":15,"text":225,"type":17,"version":18}," 到这一步，你实际上是在跑一个单广告组的广告系列，外加一层多余结构，不如直接照那样搭。",4,{"children":228,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":34},[229],{"detail":13,"format":13,"mode":14,"style":15,"text":230,"type":17,"version":18},"一个看起来不对的 CBO 广告系列怎么诊断",{"children":232,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[233],{"detail":13,"format":13,"mode":14,"style":15,"text":234,"type":17,"version":18},"按顺序问：",{"children":236,"direction":19,"format":15,"indent":13,"type":55,"version":18,"listType":268,"start":18,"tag":269},[237,243,249,255,261],{"children":238,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":18},[239,241],{"detail":13,"format":18,"mode":14,"style":15,"text":240,"type":17,"version":18},"这些广告组真的彼此不同吗？",{"detail":13,"format":13,"mode":14,"style":15,"text":242,"type":17,"version":18}," CBO 广告系列内部受众重叠时，竞争关系并不干净；分配看起来任意，可能是因为从系统视角看，这几个广告组几乎是同一个东西。",{"children":244,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":54},[245,247],{"detail":13,"format":18,"mode":14,"style":15,"text":246,"type":17,"version":18},"那个没花钱的广告组，究竟有没有投放资格？",{"detail":13,"format":13,"mode":14,"style":15,"text":248,"type":17,"version":18}," 查排期、受众规模、版位限制，以及素材是否通过审核。没有资格的广告组，经常被误读成 CBO 的分配决定。",{"children":250,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":127},[251,253],{"detail":13,"format":18,"mode":14,"style":15,"text":252,"type":17,"version":18},"跑赢的那个广告组，赢在你真正在意的指标上吗？",{"detail":13,"format":13,"mode":14,"style":15,"text":254,"type":17,"version":18}," 如果它赢在每结果成本、却输在每结果价值，那 CBO 在正常工作，错的是你的优化事件。",{"children":256,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":226},[257,259],{"detail":13,"format":18,"mode":14,"style":15,"text":258,"type":17,"version":18},"最近有没有什么被重置了？",{"detail":13,"format":13,"mode":14,"style":15,"text":260,"type":17,"version":18}," 一次显著编辑可能让整个广告系列重新学习，而早期的分配比稳定后的分配噪声大得多。",{"children":262,"direction":19,"format":15,"indent":13,"type":47,"version":18,"value":267},[263,265],{"detail":13,"format":18,"mode":14,"style":15,"text":264,"type":17,"version":18},"是不是出价策略在压制投放？",{"detail":13,"format":13,"mode":14,"style":15,"text":266,"type":17,"version":18}," 卡得过紧的上限会以看起来像分配决定的方式抑制花费。",5,"number","ol",{"children":271,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[272],{"detail":13,"format":13,"mode":14,"style":15,"text":273,"type":17,"version":18},"把这几条走完，再去考虑花费上下限。",{"children":275,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":34},[276],{"detail":13,"format":13,"mode":14,"style":15,"text":277,"type":17,"version":18},"常见问题",{"children":279,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":282},[280],{"detail":13,"format":13,"mode":14,"style":15,"text":281,"type":17,"version":18},"CBO 和 Advantage 广告系列预算有什么区别？","h3",{"children":284,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[285],{"detail":13,"format":13,"mode":14,"style":15,"text":286,"type":17,"version":18},"是同一个东西。Advantage 广告系列预算是广告系列预算优化的现用名称，改名时机制没有变化。",{"children":288,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":282},[289],{"detail":13,"format":13,"mode":14,"style":15,"text":290,"type":17,"version":18},"为什么一个广告组把预算全花了？",{"children":292,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[293],{"detail":13,"format":13,"mode":14,"style":15,"text":294,"type":17,"version":18},"因为系统估计它当下产出结果的效率最高。这是设计中的行为。只有当你的各广告组「每个结果的业务价值」不相等时它才是问题，而这时该改的是结构，不是某个开关。",{"children":296,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":282},[297],{"detail":13,"format":13,"mode":14,"style":15,"text":298,"type":17,"version":18},"怎么强制把花费喂给某个指定广告组？",{"children":300,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[301],{"detail":13,"format":13,"mode":14,"style":15,"text":302,"type":17,"version":18},"给那个广告组设最低花费下限，或者把它挪进自己的广告系列、配独立预算。第二种更可靠，因为它是消除了竞争，而不是约束竞争。",{"children":304,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":282},[305],{"detail":13,"format":13,"mode":14,"style":15,"text":306,"type":17,"version":18},"预算很小的时候 CBO 还能用吗？",{"children":308,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[309],{"detail":13,"format":13,"mode":14,"style":15,"text":310,"type":17,"version":18},"能用，但集中效应意味着你实际上可能只在跑一个广告组。预算小的时候，广告组更少、控制更明确，通常比一个你影响不了的分配更清楚。",{"children":312,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":282},[313],{"detail":13,"format":13,"mode":14,"style":15,"text":314,"type":17,"version":18},"改广告系列预算会重置学习期吗？",{"children":316,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[317],{"detail":13,"format":13,"mode":14,"style":15,"text":318,"type":17,"version":18},"幅度大的改动可能会。小步改动干扰更小。每次只改一件事，这样投放发生变化时你知道是什么推动的。",{"children":320,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":282},[321],{"detail":13,"format":13,"mode":14,"style":15,"text":322,"type":17,"version":18},"拉新和再营销该放在同一个 CBO 广告系列里吗？",{"children":324,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[325],{"detail":13,"format":13,"mode":14,"style":15,"text":326,"type":17,"version":18},"通常不该。再营销的每结果成本几乎总是更低，于是它吸走预算，而负责给再营销受众蓄水的拉新被饿死。拆成两个广告系列能让两边都拿到钱。",{"children":328,"direction":19,"format":15,"indent":13,"type":33,"version":18,"tag":34},[329],{"detail":13,"format":13,"mode":14,"style":15,"text":330,"type":17,"version":18},"一句话版本",{"children":332,"direction":19,"format":15,"indent":13,"type":20,"version":18,"textFormat":13,"textStyle":15},[333],{"detail":13,"format":13,"mode":14,"style":15,"text":334,"type":17,"version":18},"广告系列预算优化把分配权从你手里交给投放系统，而系统优化的是广告系列的总结果，不是各广告组之间的公平。花费不均通常是功能在正常工作。它出问题的场景是：各广告组每结果的业务价值不同、很多窄广告组被饿进永久学习期、或者拉新与再营销共用一个池子。动用花费上下限之前，先诊断投放资格、受众重叠和出价约束，并且每次只改一件事，这样你才读得懂结果。","root",{"id":337,"alt":338,"updatedAt":339,"createdAt":339,"url":340,"thumbnailURL":19,"filename":341,"mimeType":342,"filesize":343,"width":19,"height":19},1060,"Illustration of one campaign budget being reallocated unevenly across three ad sets","2026-09-03T10:28:48.241Z","https://cms-r2.deepclick.com/807-cover-f1f1a061e75e.jpg","807-cover-f1f1a061e75e.jpg","application/octet-stream",108626,{"title":345,"description":346,"image":347},"Facebook 广告系列预算优化 CBO：分配机制与四种不该用的场景","为什么一个广告组花光了 CBO 预算、Advantage 广告系列预算如何持续分配、你还能用哪些花费上下限与出价手段，以及什么时候该拆成两个广告系列。",{"id":337,"alt":338,"updatedAt":339,"createdAt":339,"url":340,"thumbnailURL":19,"filename":341,"mimeType":342,"filesize":343,"width":19,"height":19},{"id":18,"key":349,"name":350,"prodHost":351,"testHost":352,"blogPath":353,"docPath":354,"zhPrefix":355,"deployHookTest":356,"deployHookProd":357,"brandAuthor":358,"brandKit":362,"enabled":366,"updatedAt":367,"createdAt":368},"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",{"id":54,"name":350,"avatar":359,"updatedAt":360,"createdAt":361},25,"2026-04-22T08:09:35.299Z","2026-04-22T06:42:49.116Z",{"logoLight":19,"logoDark":19,"primaryColor":363,"accentColor":364,"tagline":365,"wechatName":19,"xhsHandle":19,"wechatQr":19},"#1a73e8","#ff6600","一次点击，多重价值",true,"2026-08-25T06:40:57.663Z","2026-07-14T06:26:38.962Z","published","facebook-campaign-budget-optimization",{"id":226,"name":372,"avatar":373,"updatedAt":380,"createdAt":380},"Ethan Cole",{"id":374,"alt":375,"updatedAt":376,"createdAt":376,"url":377,"thumbnailURL":19,"filename":378,"mimeType":342,"filesize":379,"width":19,"height":19},922,"Ethan Cole — growth & ad-tech editor","2026-07-27T07:17:05.290Z","https://cms-r2.deepclick.com/gpt_1785136218181_0-d1b8139c1927.png","gpt_1785136218181_0-d1b8139c1927.png",1965798,"2026-07-27T07:17:07.920Z",{"id":382,"site":383,"titleZh":385,"titleEn":386,"slug":387,"order":267,"updatedAt":388,"createdAt":389},7,{"id":18,"key":349,"name":350,"prodHost":351,"testHost":352,"blogPath":353,"docPath":354,"zhPrefix":355,"deployHookTest":356,"deployHookProd":357,"brandAuthor":54,"brandKit":384,"enabled":366,"updatedAt":367,"createdAt":368},{"logoLight":19,"logoDark":19,"primaryColor":363,"accentColor":364,"tagline":365,"wechatName":19,"xhsHandle":19,"wechatQr":19},"技术导航","Tech Guides","tech-guides","2026-04-27T08:37:10.576Z","2026-04-23T02:59:13.436Z","2026-09-03T10:29:17.329Z","2026-09-03T10:29:01.319Z","\u003Cdiv class=\"payload-richtext\">\u003Cp>Facebook 广告系列预算优化（CBO）把预算从广告组挪到广告系列这一层，由 Meta 决定每个广告组分到多少。Meta 现在把它叫作 Advantage 广告系列预算，这次改名带来的困惑比这个功能本身还多——底下是同一套机制，只是换了个新标签。\u003C/p>\u003Cp>它之所以能引出这么多求助帖，是因为它严格按照它宣称的方式工作，而它宣称的方式并不是大多数投放者预期的：\u003Cstrong>它优化的是整个广告系列的总结果，不是把你每个广告组都喂饱。\u003C/strong> 一旦接受这句话，几乎所有看起来奇怪的 CBO 行为都变得可以预测。\u003C/p>\u003Ch2>广告系列预算优化到底控制什么\u003C/h2>\u003Cp>预算只可能待在两个位置：\u003C/p>\u003Cul class=\"list-bullet\">\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"1\"\n        >\u003Cstrong>广告组预算（ABO）\u003C/strong>——每个广告组有自己的预算，各花各的。分配由你的结构决定。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"2\"\n        >\u003Cstrong>广告系列预算（CBO / Advantage 广告系列预算）\u003C/strong>——整个广告系列一份预算，投放系统持续地在各广告组之间分配。\u003C/li>\u003C/ul>\u003Cp>差别仅此而已。其余的一切——目标、定向、出价、素材——两种模式下的工作方式完全一样。预算放在哪一层不是策略，而是一个关于「谁来决定分配」的决定：你，还是系统。\u003C/p>\u003Cp>这个分配不是每天开始时切一刀就完了，它是持续进行的。系统对每一次机会估计：哪个广告组更可能以什么成本产出一个结果，然后把花费送到当下预期回报最好的地方。这个估计随数据变化而更新——这就是为什么你中午看到的分配比例，和半夜看到的可能完全不像。\u003C/p>\u003Ch2>为什么一个广告组把预算吃光了\u003C/h2>\u003Cp>这是关于 CBO 最常见的抱怨，而多数情况下，系统正在做对的事，只是看起来像坏了。\u003C/p>\u003Cp>如果广告组 A 能以比 B 更低的成本产出结果，那么把钱花在 A 上，整个广告系列每块钱能拿到更多结果。把大部分预算送给 A 不是 bug——那正是你选的目标。\u003Cstrong>分配不均本身就是优化结果。\u003C/strong>\u003C/p>\u003Cp>它在一种特定情形下才变成真问题：当你的各个广告组\u003Cstrong>对生意的价值并不相等\u003C/strong>时。系统优化的是你让它优化的那个指标，它并不知道那个便宜的受众之所以便宜，是因为它转化的是低客单产品；也不知道那个贵的受众是你明年的战略市场。如果各广告组的单位经济不同，而这个差异没有体现在优化事件里，CBO 会非常自信地把钱分到与你的利益相反的方向上。\u003C/p>\u003Cp>由此得到一条好用的判据：\u003Cstrong>当广告系列里所有广告组「每个结果对你的价值相同」时，CBO 才合适。\u003C/strong> 不相同时，要么把它们拆到不同广告系列，要么把价值差异表达出来让系统看得见。\u003C/p>\u003Ch2>你仍然握着的控制手段\u003C/h2>\u003Cp>把分配权交出去，不等于把一切都交出去。\u003C/p>\u003Cp>\u003Cstrong>广告组花费上下限。\u003C/strong> 你可以在 CBO 广告系列内给单个广告组设置最低和最高花费。这是「某个广告组被饿死」的直接解法，也是最容易被用过头的工具——把足够多的广告组卡得足够死，你就等于用更多步骤重建了 ABO，同时还保留着 CBO 的缺点。\u003C/p>\u003Cp>\u003Cstrong>排期。\u003C/strong> 广告组层面的排期依然生效，它约束一个广告组在什么时间才有资格拿到花费。\u003C/p>\u003Cp>\u003Cstrong>出价策略。\u003C/strong> 它设在广告系列层，并直接与分配相互作用。成本目标或出价上限会改变系统愿意购买哪些机会，从而改变 CBO 产出的分配形态。每种选项到底约束了什么，见\u003Ca href=\"https://deepclick.com/resources/blog/facebook-ads-bidding-strategy/\">Facebook 广告出价策略怎么选\u003C/a>。\u003C/p>\u003Cp>\u003Cstrong>结构。\u003C/strong> 广告系列里放了什么，是最强的控制手段——不存在的广告组拿不到预算，而分在两个广告系列里的广告组不会争同一个池子。\u003C/p>\u003Ch2>与学习期的相互作用\u003C/h2>\u003Cp>这是 CBO 与账户结构交汇的地方，也是支持它的最有力论据。\u003C/p>\u003Cp>集中有助于广告组积累到离开学习期所需的事件量。把同样的预算和同样的受众切成很多个小广告组，是让它们全部长期吃不饱的经典做法——每个都在永远学习，没有一个稳定下来。一个只有少数几个广告组的 CBO 广告系列倾向于集中花费，这会帮助跑赢的那些更快走出学习期。\u003C/p>\u003Cp>陷阱是相反的结构：CBO 配上很多个窄广告组。系统把花费集中到其中一两个，其余的贴着零，你拿到的是两头的坏处——大部分结构上没有数据，而且哪些被饿死也不由你决定。想测很多东西，要么给每个测试足够有意义的预算，要么串行地测。整体结构见\u003Ca href=\"https://deepclick.com/resources/blog/facebook-ads-account-structure/\">Facebook 广告账户结构\u003C/a>，到底什么会重置学习期见\u003Ca href=\"https://deepclick.com/resources/blog/meta-ads-learning-phase/\">Meta 广告学习期\u003C/a>。\u003C/p>\u003Ch2>改预算的代价\u003C/h2>\u003Cp>改预算不是免费的。改动广告系列预算，就改变了系统一直在据以规划的花费节奏，幅度大的改动可能把广告组打回学习期。从机制而不是从某个具体阈值出发，可以得到这几条做法：\u003C/p>\u003Cul class=\"list-bullet\">\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"1\"\n        >预算改动分步走，不要一次跳很大；每一步给它时间被观测到，再去判断。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"2\"\n        >不要在同一次编辑里既改预算又改结构——效果变化了你也分不清是哪一个造成的。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"3\"\n        >判断一次改动，要看一个覆盖完整转化周期的窗口，而不是改完之后那几个小时。\u003C/li>\u003C/ul>\u003Ch2>什么时候不该用 CBO\u003C/h2>\u003Cul class=\"list-bullet\">\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"1\"\n        >\u003Cstrong>各广告组「每个结果的业务价值」不同时。\u003C/strong> 见上文，这是最主要的一条。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"2\"\n        >\u003Cstrong>需要控制分配的对照测试。\u003C/strong> 如果测试的意义就在于每个组拿到可比的花费，那预算层面的控制就是实验本身，CBO 不会替你保住它。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"3\"\n        >\u003Cstrong>一个广告系列里混着差异很大的漏斗阶段。\u003C/strong> 拉新和再营销的每结果成本结构性地不同，系统会可预测地偏向再营销——而再营销自己是长不大的，因为它依赖的受众正是被你饿死的拉新在构建。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"4\"\n        >\u003Cstrong>预算小到「集中」就意味着永远只有一个广告组在跑。\u003C/strong> 到这一步，你实际上是在跑一个单广告组的广告系列，外加一层多余结构，不如直接照那样搭。\u003C/li>\u003C/ul>\u003Ch2>一个看起来不对的 CBO 广告系列怎么诊断\u003C/h2>\u003Cp>按顺序问：\u003C/p>\u003Col class=\"list-number\">\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"1\"\n        >\u003Cstrong>这些广告组真的彼此不同吗？\u003C/strong> CBO 广告系列内部受众重叠时，竞争关系并不干净；分配看起来任意，可能是因为从系统视角看，这几个广告组几乎是同一个东西。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"2\"\n        >\u003Cstrong>那个没花钱的广告组，究竟有没有投放资格？\u003C/strong> 查排期、受众规模、版位限制，以及素材是否通过审核。没有资格的广告组，经常被误读成 CBO 的分配决定。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"3\"\n        >\u003Cstrong>跑赢的那个广告组，赢在你真正在意的指标上吗？\u003C/strong> 如果它赢在每结果成本、却输在每结果价值，那 CBO 在正常工作，错的是你的优化事件。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"4\"\n        >\u003Cstrong>最近有没有什么被重置了？\u003C/strong> 一次显著编辑可能让整个广告系列重新学习，而早期的分配比稳定后的分配噪声大得多。\u003C/li>\u003Cli\n          class=\"\"\n          style=\"\"\n          value=\"5\"\n        >\u003Cstrong>是不是出价策略在压制投放？\u003C/strong> 卡得过紧的上限会以看起来像分配决定的方式抑制花费。\u003C/li>\u003C/ol>\u003Cp>把这几条走完，再去考虑花费上下限。\u003C/p>\u003Ch2>常见问题\u003C/h2>\u003Ch3>CBO 和 Advantage 广告系列预算有什么区别？\u003C/h3>\u003Cp>是同一个东西。Advantage 广告系列预算是广告系列预算优化的现用名称，改名时机制没有变化。\u003C/p>\u003Ch3>为什么一个广告组把预算全花了？\u003C/h3>\u003Cp>因为系统估计它当下产出结果的效率最高。这是设计中的行为。只有当你的各广告组「每个结果的业务价值」不相等时它才是问题，而这时该改的是结构，不是某个开关。\u003C/p>\u003Ch3>怎么强制把花费喂给某个指定广告组？\u003C/h3>\u003Cp>给那个广告组设最低花费下限，或者把它挪进自己的广告系列、配独立预算。第二种更可靠，因为它是消除了竞争，而不是约束竞争。\u003C/p>\u003Ch3>预算很小的时候 CBO 还能用吗？\u003C/h3>\u003Cp>能用，但集中效应意味着你实际上可能只在跑一个广告组。预算小的时候，广告组更少、控制更明确，通常比一个你影响不了的分配更清楚。\u003C/p>\u003Ch3>改广告系列预算会重置学习期吗？\u003C/h3>\u003Cp>幅度大的改动可能会。小步改动干扰更小。每次只改一件事，这样投放发生变化时你知道是什么推动的。\u003C/p>\u003Ch3>拉新和再营销该放在同一个 CBO 广告系列里吗？\u003C/h3>\u003Cp>通常不该。再营销的每结果成本几乎总是更低，于是它吸走预算，而负责给再营销受众蓄水的拉新被饿死。拆成两个广告系列能让两边都拿到钱。\u003C/p>\u003Ch2>一句话版本\u003C/h2>\u003Cp>广告系列预算优化把分配权从你手里交给投放系统，而系统优化的是广告系列的总结果，不是各广告组之间的公平。花费不均通常是功能在正常工作。它出问题的场景是：各广告组每结果的业务价值不同、很多窄广告组被饿进永久学习期、或者拉新与再营销共用一个池子。动用花费上下限之前，先诊断投放资格、受众重叠和出价约束，并且每次只改一件事，这样你才读得懂结果。\u003C/p>\u003C/div>","https://deepclick.com/zh-CN/resources/blog/facebook-campaign-budget-optimization",{"en":370,"zh-CN":370},1788431727864]