摘要
文章从用户用 ChatGPT 语音寻找海边宠物友好住宿的案例切入，指出消费者正在从多标签页搜索转向对话式 AI 的意图驱动型发现。ChatGPT 已拥有超 10 亿活跃用户、日均处理 25 亿次提示词，用户带着比较选项和做决策的目的而来，这与过去十年支撑广告预算的消费行为不同。品牌在搜索和产品发现等上、中漏斗已意识到这一变化，但要把新兴渠道变成媒体计划中的正式预算项，仍取决于 measurement。
决定 ChatGPT Ads 能否与 Meta、TikTok 和 Google 并列的核心，不是注意力或广告库存，而是广告主能否证明投放结果。点击和曝光只能说明用户与广告发生了互动，在当前预算压力下，这不足以向 CFO 和 CEO 说明花费产生了业务影响。因此，能否衡量后续转化和业务结果，是 ChatGPT 从实验性渠道走向严肃媒体渠道的关键。
传统营销技术栈在衡量对话式 AI 环境时面临两个结构性挑战。其一是打通 web 与 app 转化：Grubhub 自 OpenAI 2 月向广告主开放平台以来已在 ChatGPT 投放，积累约五个月 campaign 数据，并通过跨平台衡量在同一仪表盘中查看移动端与桌面端、app 与 web 的完整客户旅程，覆盖网站访问、购买、注册、app installs 和 in-app revenue 等指标。Brian Ryu 表示，这补全了无论发生在 web 还是 app、无论何种指标或互动场景的客户旅程全貌。其二是区分付费与自然推荐：ChatGPT 等 AI 引擎给品牌带来的流量大多并非付费流量，而是用户提问后因品牌出现在答案中被引导至 app 或网站；将独立衡量延伸至自然流量侧，才能获得一个诚实的判断，而不是依赖平台自报数据。付费或自然，纪律一致：无法诚实衡量，就无法优化。
更深层的问题在于信号质量。无论讨论 ChatGPT Ads、代表用户做研究的 AI agent，还是未来 18 个月出现的新平台，决定输出是否可信的，是输入的数据。模型、算法和优化引擎的效果，永远取决于输入数据的质量；如果信号混乱或有偏，信任就会崩塌，这不是技术失败，而是基础不牢。随着更多客户旅程迁移到 AI 界面，去重、无偏的衡量不是变得不重要，而是更加重要。广告主以往在 Meta、TikTok 和 Google 之间用同一标准比较，这一纪律现在需要延伸至 ChatGPT 及其后继平台。
对 CMO 而言，领先者不是用小额预算或 POC 对新渠道做短期试验的人，而是以衡量为核心做长期决策、追问业务影响如何被真实衡量的人。真正的机会不是一个新的广告投放位置，而是一个能被证明、能交付业务价值的新渠道。付费与自然 AI 流量都需要独立衡量，而非平台自报数字；决定广告主能否信任结果的，是信号质量，而非平台本身。
分析师点评
值得关注的是，这篇文章把新渠道能否进入媒体计划的门槛落在独立测量上，而不是流量规模；这与搜索、社交渠道过去的成熟路径一致：先有第三方归因，才有稳定预算。关键信号在于，对话式界面天然把付费投放与自然推荐混在一起，平台自报数据难以拆分，这实际上是UA（用户获取）与变现团队早已熟悉的“围墙花园”问题的延伸。在隐私限制收紧、信号质量普遍退化的背景下，去重且无偏的测量能力正从后台工具变成渠道准入的前置条件，而非投放后的补丁。
猜你喜欢
MetaA New Way for Businesses and Personal Agents to Work Together
Meta and Sierra are previewing an open Personal Agent Protocol with Genesys, NiCE, Decagon, Rocket, Shopify, Stripe, and Walmart to standardize how personal agents like Muse interact with businesses. For ad ops, the key shift: high-intent customers will arrive via agents, not browsers. Blocking agents risks losing those customers. The protocol aims to cover authentication, user identity, business steering, APIs/connectors, and agent-to-agent communication. Brands should prepare for agent-readable product, inventory, booking, and support flows, plus measurement and consent. Expect less page-based browsing, more structured API and agent traffic; adapt bot policies, attribution, and customer journey controls now.
MintegralThe Next (Agentic) Phase Of Media Buying Is Coming
Agentic AI is shifting media buying from manual execution to strategic oversight. With 91% adoption of Google PMax and 88% of Meta Advantage+, goal-based automation is already standard. The next phase uses AI agents to handle targeting, bidding, and creative optimization in real time, freeing buyers to focus on outcomes, incrementality tests, and strategic bets. Key considerations: ensure AI has direct supply, robust prediction models, and clear optimization goals (install, ROAS, CPE, CPL). The future lies in cross-platform coordination and behavioral targeting, connecting ad spend directly to business value.
TikTokIntroducing TikTok Agentic Hub: AI Skills to streamline your ad workflows | TikTok For Business Blog
TikTok launches Agentic Hub, a marketplace for AI-powered advertising solutions built on TikTok for Business MCP. It connects AI agents to advertisers' tools, enabling automated campaign creation, management, analysis, and optimization. The ecosystem includes first-party and third-party AI skills from partners like HubSpot and Wix. Advertisers can reduce manual work, gain insights, and make data-driven decisions. A limited promotion offers ad credits for new SMB accounts spending $100-$1500 within 30 days, with restrictions on eligibility.
TikTokIntroducing Symphony Agent | TikTok For Business Blog
TikTok's Symphony Agent is an AI-powered creative engine that helps advertisers produce trend-driven ads at scale. It powers Symphony Creative Studio for video generation from prompts, Content Suite for AI search of relevant creator videos, and TikTok One for streamlined creator matching and outreach. Key benefits include leveraging platform signals to generate authentic content, reducing manual effort, and enabling fast A/B testing. A limited offer provides ad credits for new SMB advertisers spending $100-$1500.
GoogleEnhanced Local Services Ads for Home Listings bring homebuyers and local agents together.
Local Services Ads (LSAs) for real estate are expanding across all 50 U.S. states, offering richer home listings with property details from HouseCanary. This allows buyers to call, message, or book appointments directly from ads. Existing LSA agents automatically appear, while new agents can sign up for high-intent leads. The initiative aims to bridge buyers and agents, with portal partners enrolling agents via the LSA managed partner program. For ad ops, this means increased lead quality and a new channel for generating qualified real estate leads.
AdjustAgentic growth management at Adjust | Adjust
Adjust introduces 'Agentic Growth Management' at MAU Vegas 2026, a goal-driven AI system that automates the campaign optimization loop. Instead of manual intervention, it sets a business objective (e.g., increase ROAS), analyzes performance, generates actionable recommendations (e.g., budget reallocation, creative refresh), and executes with marketer oversight. The system tracks impact and iterates, allowing teams to shift from execution to supervision. This approach aims to reduce operational burden while accelerating progress toward revenue targets, with automation scaling as AI confidence grows.
更多来自 AppsFlyer
CTV Measurement: Attribution to Contribution
文章核心观点是：CTV 广告在 last-touch 归因体系下看起来“失效”，根本原因是测量方式不匹配，而非渠道本身无效。CTV 几乎没有点击信号，转化又发生在其他设备和其他渠道上，因此归因模型必须从“分配功劳”转向“证明贡献”，也就是把...
Why attribution fails: How to fix it
营销归因失效并非数据导出或技术层面的问题，而是结构性的利益冲突：售卖广告的平台同时充当效果裁判，没有任何一方有动力与其他平台对账。同一笔转化常被三到四个平台同时计功，使归因总收入达到实际收入的2到4倍；在ATT与cookie退场的双重挤压下...
Creative optimization: Why your best video ad fails as a static image
AppsFlyer 基于120万条广告素材的分析指出，视频广告与静态图片广告所奖励的创意属性几乎相反，把两者视为同一套创意策略，会让广告主错失已经拥有数据验证的优胜素材。头部广告主的素材测试量正快速拉开差距：季度 UA 支出700万美元以上...
Cracking the 2026 Holiday Season: A Data-Backed Playbook for Marketing Leaders
AppsFlyer 基于 2025 年旺季八大市场 1.65 亿次安装与 2.1 亿次应用内购买数据的分析显示，Q4 大促的购买转化高度压缩在单日甚至单个小时爆发，购买提升幅度普遍是安装提升幅度的 2 至 3 倍，Black Friday ...
Omnichannel Analytics Explained
全渠道分析（omnichannel analytics）要解决的核心问题不是“哪个仪表盘的数字更准”，而是如何确认散落在移动端、Web、CTV、AI 搜索等各触点上的三次曝光其实属于同一个人，并在同一套归因引擎下完成统一衡量。文章引用数据指...
Web Gaming: The untapped path to lower CPI and boost ROAS
移动端 CPI 连年攀升、应用商店最高 30% 的抽成叠加，使游戏厂商原有 ROAS 模型难以为继，而桌面 Web 游戏已从"阉割版"进化为 Unity 到 HTML5 的完整移植，成为重启增长的新渠道。关键数据显示，Web 投放的用户获取...