Chris Vanderhook

@dan_balis @joe_zappa Yes! That was our AI planning product within ViantAI. 30% of customers use that product which is free but helps them create more impactful campaigns. Also, AI planning is a star player within our Outcomes product. Outcomes is powered by ViantAI.

@cvanderhook @joe_zappa Congrats! Genuinely curious because I haven’t kept up — IIRC you had a v cool LLM agent buying demo quite a while back focused on media planning. Is that among the drivers or still too early in the product lifecycle?

Viant $DSP jumps 16% after smashing Q4 expectations! Revenue up 22% to $110M, adj. EBITDA surges 45% to $25M. Real AI-powered ad tech is winning big. 🚀 #ViantTech #EarningsBeat

9/ What are your thoughts on $DSP Q4 earnings?

8/ We are having a lot of fun. I love this shit. Only the Vanderhooks would blast the Lattice Brain soundtrack to open your earnings call 🔥🔥🔥

@101Programmatic @Seth_Ulinski @efroestad Vlad. It’s time. Been enjoying your content for years. You need to be on Viant. We service the shit out of our customers. No tickets or chat bots.

@Seth_Ulinski @efroestad Maybe. Our account is US based. And I just asked pol on reddit and there are a lot of frustrated people sharing the same experience…

Haven’t seen this discussed anywhere - TTD has quietly changed how they service agencies. No more allocated AM team. Now you submit tickets, just like every other big DSP. That was IMO one of TTD’s biggest strengths. And yes, ticket-based support is as painful as you’d expect. Frustrating.

@tkawaja Love it!!

BAD BANKER: I gave a presentation outlining 10 predictions about AI and Advertising to a packed house at the Marketecture Live conference in NY today. As always, my objective is to capture audience attention and deliver insights in an entertaining way to make it memorable. So I dressed up as Bad (Bunny) Banker. Based on the reaction and feedback, it seemed to work. Congratulations to @aripap and the whole Marketecture Media team on a terrific event!

@adamheimlich @ChrisHarihar I think the point is that the LLM’s are commoditized between each other because they are all largely trained off of the same data. But agreed Claude has separated itself by focusing on the application within the Enterprise where OpenAI is chasing everything.

@ChrisHarihar I just don't think there's a scenario where private data lock-in drives the market for AI tech

"Models alone are commoditized; proprietary data... are the real moat." Yup. You're also seeing TTD start to use this same language around data differentiation.

Nom, Nom, Nom... (Tim Rowe for the win!)

@TreyTitone Absolutely. Too many ad tech bros selling the sexy complexity of agentic when they should just focus on what the customer wants…outcomes.

This is probably the most accurate take on how most people will perceive "agentic" integration into the software they use.

@TreyTitone Pretty much, this way you avoid having to say that Ventura isn’t going to work. And honestly, you can earn a lot more respect from investors when you can just call a spade a spade and move on. But maybe selling magic tricks makes more $ 🤷🏻‍♂️

@cvanderhook But seriously, it might be. If you are a CTV OS and vibe to OpenPath and UID2 - are you “on Ventura”?

So TTD Ventura is now an SDK and not an operating system?

@samwoods How exactly did they prove audience overlap and frequency by users? Was not aware Meta allowed for user level data

DTC brand spending $180K/month on Meta. ROAS had been declining for four straight months. The media buyer kept launching new creatives, testing new hooks, adjusting audiences. Nothing moved the needle for more than a few days. The founder exported the full ad account history. Every creative, every audience, every metric, every day for the past year. Gave the Agent swarm one job: "What's actually happening to our ad performance and why?" One agent analyzed creative-level performance decay curves. Another mapped audience overlap and frequency across every active campaign. A third modeled the relationship between landing page behavior and ad engagement. The creative wasn't fatiguing the way everyone assumed. The decay agent found that new creatives were actually performing well for the first 72 hours, then flatlining. But old creatives weren't dying of exhaustion either. They were being cannibalized. The audience overlap agent found the real problem. Twelve active campaigns were targeting audiences with 60-80% overlap. The same people were seeing the same brand six to nine times a day across different campaigns. The frequency wasn't visible at the campaign level because each campaign showed a reasonable frequency of two to three. But at the person level, it was absurd. The landing page agent added a layer nobody expected. Visitors arriving from the highest-frequency exposures had a bounce rate of 78%, compared to 31% for first-exposure visitors. The retargeting campaigns were actively training the algorithm to optimize for people who would never buy. They consolidated to three campaigns with hard exclusions between audiences. Cut spend to $140K. ROAS went from 1.4x to 3.2x in six weeks. The media buyer said it was the first time in a year he wasn't launching new creative every three days.

We just launched Outcomes, the first fully autonomous ad product for the open internet. This innovation is made possible because of our new AI Lattice Brain. The soundtrack is live and it’s 🔥🔥🔥 youtu.be/5wGvLm5T3fo?si…