Yesterday, new filings for the remedies phase of the U.S. antitrust case against Google were finally made available after months of waiting. Google won’t have to divest Google Ad Manager or its ad exchange. But the company will have to adhere to yet-to-be-finalized behavioral remedies.
From yesterday’s filing:
“ORDERED that Plaintiffs’ proposals for structural remedies in the form of the divestiture of AdX, the open-sourcing of DFP’s final auction logic, and the contingent divestiture of DFP Remainder, be and are REJECTED. ORDERED that most of the parties’ proposed behavioral remedies, as modified by this Court, be and are ACCEPTED.”
Judge Leonie Brinkema’s opinion will be made publicly available after both parties are allowed to redact confidential business information. The parties have 30 days to finalize the behavioral remedies.
Read: “United States v. Google LLC — Document #1857” (September 2) – Court Listener
The WSJ reports that the judge “had voiced concern during the trial about the practicality of a breakup, saying it would take years because Google would appeal her ruling. She also expressed uneasiness with not knowing who would buy Google’s ad exchange and operate it in the future.”
More: Google Avoids Breakup of Dominant Ad Tech Business (Sept 2) – The Wall Street Journal (subscription)
From tipsheet: The key question remains whether the final order merely constrains conduct or creates technically enforceable access and interoperability that makes rival publisher infrastructure more valuable.
TECH
Google ad tech trial: Reaction
Reaction was swift after the latest Google ad tech monopoly trial filings became public yesterday.
Google on X (since deleted): “‘We’re very pleased the Court rejected the DOJ’s proposal to break apart tools that help small businesses reach new customers and grow.’ — Lee-Anne Mulholland, VP Regulatory Affairs.”
In a statement, PubMatic CEO Rajeev Goel said in part: “The Court’s focus on behavioral remedies will likely provide a more expeditious path to address the ongoing competitive harms from Google’s illegal monopoly. The market opportunity is meaningful: we estimate that each percentage point of market share that shifts to PubMatic will generate approximately $50M-$75M in incremental revenue with the vast majority of it increasing profitability.”
Markecture’s Ari Paparo on X: 1) “Not unexpected, the AdX spin out seemed complicated and unlikely to make a big difference.” 2) “The DOJ screwed this up by asking for overly complicated remedies. The judge is in her 80s and likely to retire. She didn’t want a decade long resolution.”
TECH
Opinions
- Hightouch v. LiveRamp: “I was hoping to ease into my first post at Hightouch, but some findings are just too big to hold back.” (Sept 2) – Hightouch’s Jitendra Kumar on LinkedIn
- DSP challengers: “The Counterintuitive Comeback of the DSP” (Sept 1) – CSO Donnie Williams on Media Consulting Group’s blog
- RMNs: “Retail Media Can Build Brands—Including Ways Other Media Can’t” (Sept 1) – Analyst Andrew Lipsman on Adweek
FINANCIALS
Is Snowflake an AI company?
Snowflake’s AI business is starting to show up materially in its financial results. Product revenue for the cloud data platform grew 37% year over year in its fiscal second quarter, and CEO Sridhar Ramaswamy said AI contributed “approximately half” of the company’s recent growth acceleration. The company’s stock jumped 23% in after-hours trading.
There’s an advertising implication, too: Snowflake executives say customers adopting its AI products also consume more of its underlying data platform. AI isn’t replacing the data layer. It’s making it more valuable.
That recalls what Snowflake CMO Denise Persson told tipsheet last year: “Think of it as a ‘self-driving’ marketing platform in the future.” She described Snowflake as the infrastructure underneath that system, with applications handling functions including media planning, buying, segmentation, content creation and personalization.
On yesterday’s earnings call, Ramaswamy, a former advertising leader at Google, described a “flywheel effect” as AI creates new workloads and drives more platform consumption.
Read:
- Snowflake Reports Financial Results for the Second Quarter of Fiscal 2027 (Sept 2) – Snowflake
- Snowflake Q2 2027 Earnings Call: Complete Transcript (Sept 2) – Benzinga
- Sridhar Ramaswamy comments on the quarter (Sept 2) – See his thread on X
From tipsheet: A self-driving marketing system requires somewhere to learn. Snowflake supplies the proprietary context. Advertising can supply the feedback loop.
PLATFORMS
Forrester: Selecting a CDP for the utility era
Forrester VP and analyst Joe Stanhope outlines what marketers should prioritize in a CDP alongside the firm’s new “Customer Data Platforms For B2C, Q3 2026” Wave.
He identifies three distinguishing traits:
“Excel at data ingestion and profile assembly. Top solutions support data-handling capabilities like warehouse-native access, advanced data management and governance features, and multiple matching and identity resolution options for profile assembly.
Extend audience generation into customer decisioning. Top solutions support interaction decisioning including out-of-the-box models, real-time triggers, and journey orchestration for multichannel and multi-interaction workflows.
Align with evolving AI strategies. Top solutions support multiple AI delivery methods including AI agents, MCP server access, and data platform capabilities like managing a data context layer.”
Read: “B2C CDPs Enter Their Utility Era: Data Is The Foundation, But Outcomes Drive Investment” (September 1) – Forrester Research blog
From tipsheet: The CDP evaluation is moving beyond can it unify customer data? toward can that data drive a decision across the customer journey?
Related: Forrester on agentic AI for B2C marketers (September 2025) – tipsheet
AGENTS
Index’s ‘take’: ARTF
VP of Product Michael Richardson on Index Exchange’s website:
- “AI and automation are reshaping programmatic, bringing real-time intelligence directly into the auction to power faster, smarter decisioning. But realizing that potential at scale requires shared infrastructure and standards the industry can rely on. Enter the Agentic Real-Time Framework (ARTF), an IAB Tech Lab initiative designed to power interoperable, high-performance agentic workflows within an auction.”
Read: “What Is the Agentic Real-Time Framework (ARTF)?” (Sept 2) – Index
CREATIVE
Audience data veteran says creative comes first
Omar Tawakol, co-founder and CEO of AI advertising startup Rembrand and former CEO of BlueKai (acquired by Oracle in 2014), argues that digital advertising has had the relationship between audience and creative backward.
He outlines his hypothesis:
“Audience data is important, but what if we had part of this backward? What if, instead of starting with a hypothesized audience and serving all creatives to that fixed audience, we started with the creative and used it to determine who to serve to? In this formulation, the creative itself is an incredibly rich source of data that drives the targeting and our understanding of the audience.”
That model is already taking hold inside Meta and TikTok, he writes. The next step is extending it to the open web and CTV.
Read: “Creative Is the New Targeting” (August 31) – Rembrand blog
From tipsheet: As machines get better at finding the audience, creative starts doing some of the work audience data used to do. The marketer’s advantage may move upstream: what ideas do I give the machine to find demand?
LLMs & CHATBOTS
Developments
- Introducing Gemini 3.8 Flash and 3.8 Flash Cyber (September 2) – Google
- Path to Astra: critical capabilities and frontier safeguards (September 1) – OpenAI
- What’s new in Claude Fable 5.1 – Claude Platform Docs
MEASUREMENT
Incrementality moves into the optimization loop
Online resale marketplace ThredUp nearly tripled its Pinterest spending after a year of sequential incrementality experiments with measurement company Haus, according to a new case study.
After an initial holdout test showed Pinterest acquiring incremental customers at 40% below ThredUp’s target cost, increasing spend hurt efficiency. ThredUp, Haus and Pinterest then ran a multi-cell experiment to determine the optimal campaign-objective mix.
The winner: an “80% Catalog sales, 20% Conversion objective split that kept growth strong and efficient.” ThredUp subsequently doubled spending at that mix, ultimately producing a cost per incremental acquisition 45% below its goal.
Read: ThredUp surpasses growth goals on Pinterest with a test-and-learn approach – Pinterest Business
From tipsheet: Incrementality usually answers did the advertising work? Here it’s starting to answer how should the advertising run?
And as AI systems such as Pinterest’s Performance+ automate more of the execution, incrementality could help determine the objectives those systems should optimize against.
PEOPLE MOVES
Now hiring
- “I’m happy to share that I’m starting a new position as SVP, Head of Revenue – North America at InMobi!” (Sept 1) – Former OpenX CEO and sales leader Matt Sattel on LinkedIn
MORE
- AdSense: Google AdSense Begin-to-Render Impression Measurement Coming (Sept 2) – SE Roundtable
- Momentum: Big Screens, News, Reality Shows Lead Programmatic Ad Gains (Sept 2) – MediaPost
- Measurement: NielsenIQ and Similarweb Advance Agentic Commerce Measurement for the AI Shopping Era (Sept 2) – press release
- Ad tech map: “CartographAI Launched To Help Advertisers Pick The Right Tech Vendors. Now, It’s Helping Vendors Market Themselves, Too” (Sept 2) – AdExchanger