Prebid 计划与 Google 互换位置,反转广告竞价流程
The Auction Flip: Prebid’s Plan To Swap Places With Google
- 1Prebid 峰会公布新方案,拟让 Google 把竞价接入 Prebid 并由其做决策
- 2背景是 Google 广告技术业务被裁定垄断,补救方案是整合而非拆分
- 3补救还要求发布商获得含赢标与输标的出价数据,Google 员工在会场沉默旁听
- 4Treasure AI 由 CDP 转型为智能体体验平台,媒体编辑称智能体蜜月期已过
事件实质是广告竞价决策权可能从 Google 体系转向 Prebid,反垄断补救以整合代替拆分。
- 影响:若 Prebid 方案落地,发布商与广告主可获得更完整的竞价和输赢数据,出价与流量策略有了重算空间;依赖 Google 广告技术栈的业务需评估渠道与对接安排的变化。
- 参考:可考虑让程序化广告团队梳理现有竞价链路对 Google 广告服务器和出价数据的依赖点,评估对接 Prebid 的工作量;营销团队值得复盘在用智能体工具的真实效果。
- 依据:补救要求出自垄断案裁决,Prebid 方案系该组织在峰会上公布,尚属提议;智能体缺陷的判断来自 AdExchanger 编辑的现场观察。
Thursday, October 15th, 2026 – 3:30 pm
SHARE:
Google’s ad tech business was ruled a monopoly. But the remedy isn’t a breakup; it’s an integration.
Google needs to do a few things: AdX must bid into other ad auctions, not just the ones in its ad server. Publishers must receive bid data, including winning and losing bids. And it is required to integrate its ad server with Prebid.
At the Prebid Summit this week, the group unveiled its plan for a new ad auction. The first shall be last, and the last shall be first: Instead of Prebid running an auction that’s fed into Google’s ad server, it’s proposing that Google feed its auction into Prebid, allowing Prebid to collect bids and handle the decision-making.
Does this mean Prebid wants to be an ad server?
That’s unclear, but the change ignited interest at the event. Google employees were in the meeting rooms, our editors note, but remained silent observers even during animated group discussions about Google’s own tech.
From CDP to AEP
Then, we turn to another conference AdExchanger attended last week: a Florida summit for Treasure AI, a CDP (customer data platform) turned AEP (agentic experience platform).
Marketers have used data to better understand their customers for ages, but AI is changing how they use that data. Treasure AI is one of a number of companies (Optimizely also built AI “colleagues,” AdExchanger Associate Editor Joanna Gerber observes) that are building agents with distinct areas of expertise. Just like at a company, this team of specialists works together to create a finished product.
But, as agents become more sophisticated, the humans working with these AI employees are gaining a stronger sense of what agents can – and can’t – do. “The honeymoon phase is over,” Gerber says. “Now we’re starting to see the flaws. And the question is, ‘Will the relationship last?’”
来源:AdExchanger · adexchanger.com