The Trade Desk beat Wall Street’s lowered expectations late yesterday, reporting essentially flat Q3 2026 revenue while outlining how agents, measurement and data could help restore growth.
The Trade Desk
-
- Markets: The Trade Desk stock was up 2.4% in after-hours trading yesterday.
- Stat: Q3 revenue was $736 million, down 0.5% year over year and approximately $85 million above Wall Street expectations. The company expects Q4 revenue of at least $755 million.
- Context: Results included a nonrecurring partner contribution worth approximately three percentage points of revenue growth. Political advertising also helped, while a change in how certain costs are recognized affected reported revenue and expenses; management did not quantify that impact.
- Results: “The Trade Desk Reports Third Quarter 2026 Financial Results” (Nov. 9) – The Trade Desk
- Transcript: “Trade Desk Q3 2026 Earnings Call Transcript” (Nov. 9) – Benzinga
Among the highlights from CEO Jeff Green’s comments on the call with Wall Street analysts:
-
- Agents selecting audiences: The Trade Desk is using Model Context Protocol (MCP) to connect its AI, Koa, with participating data providers’ AI systems, which can respond to advertisers’ objectives with custom audience segments. “Koa can then evaluate those options and use a segment only when it believes it is likely to improve performance.”
- LLM ads, with conditions: Green sees opportunity in chatbot advertising but wants enough metadata to evaluate inventory and enable price discovery. “It’s no longer about getting the first access. It is about getting the right access.”
- Measurement: The Trade Desk plans to launch Open Measurement later this quarter, bringing attribution and incrementality into the platform. Its argument for premium media: “Branding will always suffer and be underfunded when last-touch attribution is the default because the heavy lifting of branding is done in the premium media.”
- Campaign creation: A campaign creation agent in beta can read a client’s media plan and set up campaign structure, ad groups, budgets and flights “in a matter of minutes,” Green said, automating manual work that can take hours.
- Pricing: Green defended The Trade Desk’s percentage-of-spend model and criticized outcome-based platforms that control customers’ data and report results using their own metrics: “You’re essentially renting your own data back and you’re dependent forever on those platforms to produce those results.”
- Leaner teams: Green explicitly connected agents with staffing: “Work that has historically required many people and significant manual effort can increasingly be assisted or executed by agents, allowing leaner teams to operate even more sophisticated programs.”
Wall Street reacts
Dan Salmon, lead analyst, New Street Research: “The product story is improving, but there is nothing we see as game changing and much of it is still in beta or pre-launch.” Salmon, who reiterated a Sell rating after the call, warned that agents could also weaken DSPs’ position, citing Olyzon’s agentic orchestration across platforms, including The Trade Desk: “That’s the kind of layer sitting above the DSP that we flagged as a risk to the model.”
More: “The Trade Desk Clears Wall Street’s Low Expectations as Revenue Stalls” (Nov. 9) – Adweek (subscription)
LLMs & CHATBOTS
OpenAI draws a line on chat data
OpenAI still won’t share what users tell ChatGPT with advertising partners, according to an OpenAI executive at LUMA Partners’ Digital Media Summit yesterday.
Previewing her Adweek article, Kendra Barnett reported on LinkedIn (Nov. 9):
“OpenAI monetization chief Benji Shomair says the company’s commitment not to share what people tell ChatGPT means there are things that advertising partners want that OpenAI “just can’t” provide… he also attempted to distance himself from the company’s reported $100 billion ad revenue target, saying the lab has ‘made no statements about the direction of the [ads] business and where it’s trying to go.’”
Read: “OpenAI’s Monetization Chief to Ad Partners: User Trust Means Some Things ‘We Just Can’t’ Do” (Nov. 9) – Adweek (subscription)
At DMS: “We have a pretty clear rubric [from OpenAI]. First and foremost, we are optimizing for user trust, after that, our next priority is to create User value and only after that we focus on creating Advertiser value.” – Marketing executive Lou Paskalis on LinkedIn
From tipsheet: Trust matters. But what if a consumer could choose to share a specific shopping conversation with advertisers? ‘Reach me with a good offer.’
PLATFORMS
More ads for Google AI Mode — now live
Google has extended advertising deeper into AI Mode conversations, with ads appearing in responses to follow-up questions.
Search Engine Roundtable’s Barry Schwartz spotted Google’s previously announced “multiturn ads” (Oct. 28) during testing on Friday, including ads in responses to third and fourth follow-up queries. The placements are labeled “Sponsored.”
Schwartz concludes, “I guess Google will have better earnings in its next quarterly report because we are seeing more ads in AI mode?”
Read — and see screenshots: “Google AI Mode Multiturn Ads Are Live – Follow-Up Query Ads” (Nov. 9) – Search Engine Roundtable
CONNECTED TV
Viant buys Waymark for AI video ads
AI ad technology platform Viant is acquiring AI video creation company Waymark, the companies said Monday.
Waymark will retain its brand and platform, with existing customer contracts and pricing unchanged, according to Waymark CEO Alex Persky-Stern on Waymark’s website. (Nov. 9)
The goal is to connect generative creative with Viant’s media buying, attention measurement and AI decisioning. Viant CEO Tim Vanderhook described the ambition on LinkedIn: “Generate, activate, measure attention, and optimize — all inside the Viant platform.” Read it.
Earnings: See Viant’s results below.
PLATFORMS
Former TTD-ers launch “incrementality DSP”
Former The Trade Desk executive Eric Holt announced Artifact yesterday, describing the new demand-side platform (DSP) as “the incrementality DSP, built to replace platforms built to extract.”
Holt outlined the pitch to agencies and brands on LinkedIn: “We do it in a few ways. One clear fee. Spend that follows what the ads caused. Every move visible and provable. Simple workflows, with agents and machine learning doing the heavy lifting.”
The team includes former The Trade Desk employees.
Read Holt’s announcement. (Nov. 9)
AGENCIES
Provenance: WPP agents enable celebrity IP
Ogilvy launched ECHO, an AI creative platform powered by WPP’s agentic marketing platform WPP Open, through a partnership between WPP and talent agency Creative Artists Agency (CAA).
The platform connects verified celebrity scans from CAA Vault with AI production, machine-readable usage restrictions and mandatory talent approval of every asset before publication. Models are trained exclusively on the Vault’s scans. Consumer-facing ads carry a visible watermark: “[Talent Name] Approved | Recreated with AI.”
“By embedding talent consent, machine-readable contracts, and creative craft directly into the platform, we’ve created a system where the biggest stars can license verified digital replicas safely,” said Ogilvy global innovation chief Carol Reed.
Ogilvy says ECHO can expand a traditional shoot’s eight to 10 assets into approximately 30 to 50, potentially reducing average cost per asset by 65% to 75%. It is commercially available across the WPP network.
Read: “Ogilvy Launches ECHO, the Industry’s First Talent Consent-Based Creative AI Platform” (Nov. 9) – Ogilvy
More: “A.I.-Generated Celebrity Ads Get an Assist From Terry Crews” (Nov. 9) – The New York Times (subscription)
From tipsheet: Provenance is becoming a commercial capability inside WPP Open. That gives WPP, across its agencies, something concrete to sell alongside faster production: a documented basis for using a celebrity’s digital replica.
CREATIVE
Provenance: Disney’s AI starts with Disney’s IP
Adobe and The Walt Disney Company announced a multi-year AI alliance yesterday, extending a Disney Imagineering collaboration announced in June to The Walt Disney Studios’ StudioLAB and ESPN Edge. Adobe will provide creative tools plus dedicated engineering and research resources.
The models are “customized using Disney intellectual property, including unique world knowledge like character attributes, environments, and visual styles,” Adobe says. The company describes the resulting outputs as “commercially safe” and production-ready.
Initial StudioLAB projects will explore custom AI models for post-production and visual effects. ESPN will explore applications for sports highlights, documentaries and its archive. The alliance also includes joint marketing activations and an expanded Adobe presence across Disney’s media platforms.
Read: “Adobe and The Walt Disney Company team up to bring new creative possibilities to life with AI.” (Nov. 9) – Adobe
From tipsheet: For brands evaluating AI production, the source material and permission to use it are part of the product’s value.
SELL-SIDE
Time keeps people doing the selling
Following its September rollout of Kana AI agents, Time is keeping advertiser relationships in human hands. Its agents help prepare proposals and manage campaigns, but the publisher isn’t planning to have them negotiate deals directly with buyers’ agents, AdExchanger reports.
Read: “Time’s AI Agents Do The Coordinating, And Its Sellers Still Do The Selling” (Nov. 9) – AdExchanger
Related: “Kana CEO Tom Chavez on what comes after the martech stack” (Aug. 20) – tipsheet
COMMERCE MEDIA
Who’s buying: the shopper or the agent?
Shopify VP Lauren Nemeth discusses AI commerce in Shopify-sponsored content published by Retail Dive. The article spans three distinct AI shopping behaviors:
- AI-assisted discovery: Ask an assistant, click through and buy.
- Checkout inside an AI app: Select and approve a purchase in the conversation.
- Delegated purchasing: Let an agent choose and buy within your instructions.
Shopify says AI-referred orders grew 3.3× year over year in Q3. Those are orders attributed to AI referrals, not a count of autonomous purchases.
Read: “How enterprise brands are retooling commerce platforms to capture AI shoppers” (Nov. 9) – Retail Dive
Related: Meta’s Muse offers a middle ground: the agent finds products, compares options and builds a cart, while the shopper approves the purchase. Shopify describes the process here.
FINANCIALS
Earnings reports
The latest earnings reports from across AI, advertising and marketing:
Zeta Global
-
- Markets: Zeta Global stock was up 1.7% in after-hours trading yesterday.
- Stat: “Delivered revenue of $493 million for the third quarter, an increase of 46% Y/Y or 30% Y/Y excluding Marigold and political candidate revenue.”
- Results: Zeta Global announces Third Quarter 2026 Financial Results (Nov. 9) – Zeta Global
PubMatic
-
- Markets: PubMatic stock was down 2.2% in after-hours trading yesterday.
- Stat: “Revenue in the third quarter of 2026 was $79.6 million, up 17% compared to the same period of 2025…Revenues in Q3 2026 from CTV, mobile app and emerging revenues represented approximately 64% of total revenue, up from approximately 60% in Q2 2026.”
- Results: PubMatic Announces Third Quarter 2026 Financial Results (Nov. 9) – PubMatic
Viant
-
- Markets: Viant stock was down 5.7% in after-hours trading yesterday.
- Stat: “Connected TV advertiser spend increased nearly 40% representing over 50% of total advertiser spend on the platform and reached a record high in the third quarter.”
- Results: Viant Announces Third Quarter 2026 Results (Nov. 9) – Viant
LLMs & CHATBOTS
Developments
MORE
-
- It’s about outcomes: “AI-Powered Video Outcomes: Agentic Audience Discovery for Video” (Nov. 9) – IAB
- AI visibility informs paid search: Microsoft recommends using AI Visibility’s grounding queries to identify gaps in paid search coverage. (Nov. 9) – Microsoft Advertising on LinkedIn
- About that Instacart acquisition: “DoorDash Shareholders Aren’t Convinced the Company Can Take On Amazon” (Nov. 8) – The Wall Street Journal (subscription)
- Noise: “Coca-Cola wouldn’t make a third AI Christmas ad if online backlash mattered enough” (Nov. 9) – Andrew Tindall on The Drum
- Retail media: “Why McDonald’s Media Network’s $1 Billion Ambition Isn’t So Crazy” (Nov. 6) – Andrew Lipsman on his Media, Ads + Commerce Substack (subscription)