AdTech AdTalk

Adam and Gareth

Buy Side Meets Sell Side. Live Q&A with CEOs telling the truth about where Ad Tech is headed.

  1. 1 day ago

    Rock Bottom = Good News

    Fresh off vacation, Adam Heimlich and Gareth Glaser return for an unfiltered look at digital advertising’s shifting landscape. The show opens with an honest discussion on leadership friction, founder existential dread, and the contrasting "publisher dread" currently gripping the open web. Gareth likens the current state of digital publishing to being stuck on a 150-year-old sinking ocean liner, exploring how publishers are navigating traffic declines, Google AI Overviews, and the transition toward lifetime value (LTV) tracking and direct audience relationships. On the buy side, Adam reveals a hilarious yet frustrating trend: major agency holding companies are sending Chalice RFIs that verbatim repeat Chalice’s own modular AI pitch back to him after seven years. The hosts dissect why agency data aggregation models inherently output average performance for enterprise brands, while analyzing the broader macroeconomic shift as traditional consumer ad spend contracts and migrates into growth sectors like pharma, GLP-1 weight loss drugs, sports betting, and gaming. The episode climaxes with a deep dive into Taboola’s game-changing deal to manage NBC Universal’s programmatic sales. Gareth and Adam evaluate Taboola’s closed-loop performance model, its massive SMB advertiser scale, and why Wall Street is severely hypocritical in its valuation of Taboola compared to trade DSPs. They close with a spirited debate on agentic ad tech protocols, Beeswax history, and why AI will inevitably accelerate the in-housing of $500M+ enterprise ad budgets. KEY TOPICS Publisher Existential Dread & The Traffic Plateau: Gareth described open web traffic decline as a sinking 150-year-old liner, noting that while traffic has hit rock bottom, bright spots exist in local news adopting LTV tracking and gaming sites with direct type-in audiences. Agency RFIs & The Flaw of Data Aggregation: Adam shared that agency holding companies are sending him RFIs repeating Chalice's exact modular AI pitch back to him, noting that generic agency stacks aggregate data to output average results rather than custom enterprise optimization. The Rise of Pharma & Shifting Consumer Economies: Adam and Gareth noted that traditional consumer brand categories like beer and snacks are contracting in a slowing economy, shifting ad spend into growth sectors like pharma, GLP-1s, sports betting, and gaming. Taboola’s NBC Deal & The Closed-Loop Performance Model: Gareth highlighted Taboola taking over NBC Universal's programmatic sales, arguing that Taboola's closed-loop 65/35 revenue split and SMB onboarding infrastructure make it a massively undervalued performance powerhouse. Agentic Ad Tech & Accelerating Brand In-Housing: The hosts discussed how AI agents will streamline campaign execution, predicting an accelerated wave of brand in-housing that will challenge traditional agency fee structures. CHAPTER TIMESTAMPS 00:00 Golf Shop Pre-Roll & Show Intro 01:10 Leadership Team Meltdowns & Founder Dread 05:43 State of the Open Web: Sinking Liners & Traffic Drops 07:32 Publisher Survival: LTV Tracking & Gaming Sites 12:13 Google AI Overviews & The Review Site Dilemma 17:30 Yahoo Finance, Weather & Black-Hat SEO Traffic 19:25 State of the Buy Side: Agencies Quoting Chalice's Pitch 22:22 Why Aggregated Agency Data Outputs Average Results 24:17 Contracting Economies vs. Pharma & GLP-1 Ad Spend 28:06 Free Trade, Inflation & TV Price Deflation 31:33 DSP Custom Algorithms & Beeswax History 37:47 Taboola Takes Over NBC Universal's Programmatic Sales 42:04 Why Taboola is Massively Undervalued vs. DSPs 49:01 The Open Web vs. Taboola's Publisher Guarantees 54:03 Agentic Ad Tech, ADCP & Swivel Debates 59:21 Reaching Rock Bottom: Reimagining the Open Web

  2. 14 Aug

    In Defense of Defensibility

    Fresh off a vacation in Fire Island, Adam Heimlich joins Gareth Glaser for an episode that balances hilarious observations on human status-seeking—complete with a story about Brooklyn zoo-goers misunderstanding baboon mating habits—with deep strategic analysis on the future of ad tech defensibility. The hosts open by examining what happens to the internet as Google Search's dominance wanes, predicting how publishers will pivot toward "orthogonal" content (highly uncorrelated, novel, or real-world local news) and why Google will artificially feed referral traffic to prevent publishers from blocking AI crawlers.The core of the episode tackles software defensibility and enterprise ad tech architecture. Gareth shares inside details from defensibility meetings at Chalice, challenging engineers to scope out what it would take for a $100M competitor to clone their tech, while highlighting how engineers think in "mind castles." Adam breaks down the fifth and final installment of his "Two Architectures" blog series alongside a new Bayer case study, explaining why modeling on the joint outperforms standard lookalikes, why all ad tech data science problems are edge cases, and why pure SaaS models fail in enterprise ad tech compared to custom per-advertiser infrastructure.The hosts wrap up with a tour through industry news and earnings. They dissect the ongoing backlash against The Trade Desk and CEO Jeff Green, defending Green's long-term conviction around DSP longevity, data privacy, and agentic prompt interfaces. They also critique ADCP and audit chains, evaluate Mark Ritson and Byron Sharp's "share of search" metrics, discuss Martin Sorrell's take on agency holding company defensibility, and celebrate Uber's ad business success story under Mark Grether.00:00 Oatmeal Court Audio & Fire Island Vacation03:16 Baboon Mating Myths & Human Status-Seeking06:48 Accused of AI Writing & The Post-Search Internet14:43 "Orthogonal" Content & Local News Production20:00 Local News Attribution & Google Spotlight Tags22:05 Gareth’s Defensibility Meetings & Engineer "Mind Castles"28:15 Adam’s Bayer Case Study & Data Science Edge Cases32:15 The Myth of Pure SaaS in Enterprise AdTech34:27 Pooled Averages vs. Per-Advertiser Custom Models37:50 AWS Fabric, ARTF Adoption & Measurement Lock-In43:25 Google Buyer Direct, ADCP & The Audit Chain Myth47:11 Share of Search, Byron Sharp & The Dark Lord of Penetration51:20 Martin Sorrell & Holding Company Compute Aggregation55:08 In Defense of The Trade Desk & Jeff Green's Conviction1:01:06 Collapsing Agencies & DSPs into Agentic Systems1:04:00 The Seismic Leak Story & Uber's Ad Success

  3. 7 Aug

    Revenge of the SSP

    Adam Heimlich (representing the buy side) and Gareth Glaser (representing the sell side) dive into a packed earnings season episode examining major structural shifts across digital advertising. The hosts open with a deep dive into the emerging "infrastructure layer" thesis, comparing cloud plays like AWS Fabric and Amazon Publisher Services (APS/TAM) to Prebid setups. They analyze how Amazon gains auction advantages by running proprietary code on publisher pages outside Prebid’s rules, while also highlighting Amazon Bedrock's leadership in multi-LLM model management.The central theme of the show turns to the "Revenge of the SSP". Gareth and Adam break down why supply-side platforms are reinventing themselves as ad tech infrastructure providers through real-time containerization (ARTF) integrations with companies like Chalice across Index Exchange, OpenX, PubMatic, and Magnite. They explore PubMatic's stellar earnings driven by a 30% reduction in infrastructure costs, Magnite's hybrid cloud/on-prem strategy, and how curation accidentally achieved the true goals of Supply Path Optimization (SPO).The hosts also contrast DSP data-hoarding with SSP log-file transparency, dissect Nielsen's acquisition of DoubleVerify alongside Cybids and Rockerbox, and engage in a spirited debate over biddable media versus packaged direct sales using an "omakase vs. buying the whole cow" butcher metaphor. They wrap up with a spicy "Postmortem" segment roasting Jonah Goodhart’s case study on serving ads to AI agents, with Gareth framing prompt-injecting LLMs as nothing more than black-hat SEO.2. Key TopicsThe Infrastructure Play & Amazon APS Advantage: Gareth and Adam examine how scaled ad tech is shifting toward cloud infrastructure layers, noting that Amazon's APS/TAM gains an edge by running outside Prebid's open-source rules and timeouts."Revenge of the SSP" & Real-Time Containerization (ARTF): Adam and Gareth explain that SSPs are thriving by transforming into open architecture and containerized decisioning hubs (ARTF), enabling tools like Chalice to lower infra costs and drive publisher revenue.DSP Data Hoarding vs. SSP Log Transparency: The hosts contrast SSPs' willingness to share raw log files for custom algorithmic optimization against DSPs' historical tendency to withhold logs under the pretense of privacy.Direct Sales ("The Whole Cow") vs. Biddable Programmatic ("Omakase"): Gareth and Adam debate inventory packaging, arguing that direct-sold packages force buyers to pay for unwanted filler ("the whole cow"), whereas impression-level biddable programmatic delivers targeted value ("omakase").Ads for AI Agents & Black-Hat Prompt Injection: In the Postmortem segment, Gareth critiques Jonah Goodhart's case study on context-injecting ads into LLMs, calling it spammy black-hat SEO that models like Claude will actively reject.00:00 Show Intro & Summer Earnings Energy02:26 The Fabric Architecture & Infrastructure vs Application Layer04:43 Amazon APS/TAM & Prebid Rule Bypassing12:15 Amazon Bedrock & Multi-LLM Modular Architecture14:42 Real-Time Containerization (ARTF) Across Index, OpenX & PubMatic16:58 PubMatic Earnings Blowout & Cost Reductions21:05 Magnite's Hybrid Cloud/On-Prem Whitepaper24:18 "Revenge of the SSP" & Accidental SPO27:04 DSP Shifts, Trade Desk & Log File Access34:02 John D'Souza Comment on Amazon Logs & AMC Clean Rooms36:58 Nielsen Acquiring DoubleVerify, Cybids & Rockerbox43:36 Ali Manning on AdExchanger & The Origins of ARTF45:21 Regional Allocation Optimization (YouTube, Meta & DSPs)48:24 The Great Packaging Debate: Omakase vs. The Whole Cow54:51 Agentic Buying & Strategy Selection58:07 Postmortem: Jonah Goodhart vs. Will Kakowski on Ads for AI Agents

  4. 31 Jul

    I Have a Thesis

    After recovering from a truly bizarre pre-roll video featuring an opossum massage, hosts Adam Heimlich and Gareth Glaser dive deep into defining their core operating philosophies for Chalice and Gamera. Adam lays out his "Two Architectures" thesis, arguing that enterprise advertisers require unique, one-at-a-time AI stacks based on their own data, rather than the "pooled stacks" used by small businesses, to prevent brand equity erosion and slows growth. Gareth counters with a critique of the shaken faith in programmatic, specifically among publishers moving back toward direct sales, necessitating a new thesis on "Trust in Programmatic." The hosts litigate the complete lack of standards and enforcement in "seller-defined" programmatic fields, which publishers have historically gamed to satisfied generic, pooled DSP algorithms that fail to spread value across opportunities. The conversation shifts to the distinction between data, measurement, and optimization. Adam asserts that Chalice is uniquely optimizing empirical page metrics—like network calls and refresh rates—against robust per-advertiser valuations. Gareth advocates for a zero-trust approach, positioning Gamera as the necessary neutral third party to objectively label inventory and create feedback loops for publishers. In the news segment, the hosts analyze Omnicom’s recent earnings call, slamming CEO John Wren for punting on analyst questions regarding whether AI efficiency savings benefit the agency or the advertiser. Adam and Gareth express deep skepticism over agency claims of multi-year internal AI builds, suggesting they are likely relying on generic tools from Google and Microsoft. They close with Ian Whitaker’s analysis of Microsoft versus Meta capex spending, analyzing why the market rationalizes Microsoft’s compute investments as widgets to be sold, but remains skeptical of Zuckerberg’s ability to translate massive hardware bets into an impregnable advertising moat. 00:00 Apossum Massage Madness & Intro02:29 The Adam Loves AdTech Blog Engagment04:31 Adam's Thesis: The Two Architectures of AI Advertising08:23 Gareth's Thesis: Shaken Faith in Programmatic09:37 The Lack of Trust & Standards in Bidstream Fields11:44 Publisher Shenanigans & The Failure to Spread Value13:20 The Need for Neutral Third Parties & Zero Trust16:40 Objective Site Assessment vs. Gamed GPIDs22:00 The App Nexus Origins of Curation24:30 Why Curation Works: Beating the DSP Algo29:23 Does AdTech Actually Care About Advertiser Performance?31:47 Dissecting Data, Optimization, & Measurement34:54 Viewability Table Stakes vs. Gamed Inventory39:50 JS Logs, iframes, and Closing the Pub Loop43:30 Gamera Business Model & Free Publisher Monitoring45:41 News Pubs & Oracle's Brand Safety Vacuum50:50 News: Agency Earnings, AI Savings & The John Wren Punt52:55 Moving too Fast: The Pony Scheme Analogy55:51 News: Meta vs. Microsoft Capex & Ian Whitaker Analysis59:12 Social Media is Just a Website1:00:32 Singularity & Next Week: Kaiju Open Weight Model Fight

  5. 24 Jul

    Is It Recipe Sites? Or Is It Cake?

    Adam and Gareth return after a week off, diving straight into high-level ad tech analysis. The hosts tackle how user IDs and audience targeting are evolving in modern advertising. Rather than viewing IDs as rigid segment buckets, they explore how advanced buyers like Chalice treat IDs as correlation fields to drive flexible propensity scoring and true incrementality modeling.The conversation moves into market economics, the shifting balance of power between CMOs and CFOs, and the bull vs. bear cases for programmatic technology. Gareth and Adam break down why ghost bidding often produces messy data compared to clean public service announcement (PSA) control groups, why retail media networks are primarily useful for conquesting rather than loyalty targeting, and how programmatic tech fees could theoretically compress down to a lean 15% to 18%.Finally, the hosts unpack technical mechanics on the supply side, explaining how open web native ads operate at a component level and why supply-side platforms (SSPs) like Media.net, TripleLift, and Magnite lead format innovation. They close out with a candid discussion on publisher revenues dropping 40% in Q2, an infuriating case study of a luxury brand irrationally blocking news publishers, and a witty critique of Sam Altman's rumored $100 billion AI ad network ambitions.Key TopicsUser ID Propensity Scoring & Incrementality: Adam and Gareth explained that user IDs should be treated as flexible correlation fields for propensity scoring and incrementality rather than rigid target audience buckets.The Flaw of Retail Media Networks & Ghost Bidding: Adam noted that retail media networks mostly hold repeat-buyer data best suited for brand conquesting, while ghost bidding creates unreliable data compared to clear PSA control testing.Bull vs. Bear Cases for Programmatic & AI: Gareth presented a bear case where middleware vendors re-complicate ad tech and Google Cloud locks in buyers, while Adam shared a bull case where tech fees compress to 15-18% and programmatic expands into mid-market performance budgets.SSP Formats & Mechanics of Open Web Native Ads: Gareth detailed how true native advertising delivers structured component arrays rather than pre-assembled banner tags, explaining why SSPs like Media.net, TripleLift, and Magnite drive format creation.Publisher Squeeze & Irrational News Blocking: Adam and Gareth discussed how search retention hurts publisher traffic, sharing a real-world story of a luxury brand cutting news inventory despite high campaign performance.00:00 Toilet Privacy Sensor & Show Intro01:19 User IDs, Audience Targeting & Chalice04:02 Propensity Scoring vs. Rigid Bucketing06:10 The Retail Media Network Conquesting Dilemma08:27 Incrementality: PSAs vs. Ghost Bidding12:02 CFO Power & Marketing Mix Modeling (MMM)15:15 The Branded Search Tax & ADT Security Story18:15 Low Viewability Waste & Zero-Trust Bid Fields24:24 Bull vs. Bear Cases for AI & Ad Tech Efficiency36:30 Open Web Video Distribution & Maldo's Birthday39:22 Media.net & Why SSPs Own Format Innovation45:19 How Open Web Native Ads Actually Work49:20 Publisher Q2 Traffic Squeeze & Google Retention53:31 Luxury Brands & The Irrational News Block55:40 Blog Plugs, Twitter Feuds & Sam Altman's Ad Dreams

  6. 10 Jul

    Fruit of the Poisonous Tree

    Adam and Gareth return to tackle the pollution of the media ecosystem, from AI-generated "slop" on LinkedIn to the strategy of flooding the zone with fake news. The guys dive deep into the philosophical roots of Palantir CEO Alex Karp, exploring how his studies in ontology under Jurgen Habermas in Germany led to the realization that AI training models desperately need human knowledge workers for proper data labeling. This same need for consistent taxonomy paved the way for companies like Jounce Media and Sincera to redefine quality in ad tech. The main event focuses on Google's new "Buyer Direct" product within Google Ad Manager (GAM). Gareth argues that Google's unified architecture could easily kill complex Agentic workflows in the crib, proving that "automation doesn't beat architecture". Adam points out that while publishers may love Google's unfathomably low 10% take rate, it's ultimately "fruit of the poisonous tree"—a massive competitive advantage born directly from Google's monopolistic tying of AdX and GAM. Plus, the guys discuss Warner Bros. Discovery building an ad stack on AWS, why CTV is still wildly mispriced due to information asymmetry, and why publishers are finally opting out of Google Search. Key Topics: - The Flood of AI Slop: How AI is polluting the media landscape and making it impossible to discern genuine expertise on LinkedIn. - Alex Karp's Ontology: How the Palantir CEO's background in German social theory applies directly to the manual labeling required for AI training data. - Google Buyer Direct: Why Google's low-fee, direct-to-publisher pipeline could completely bypass DSPs and destroy fragmented Agentic workflows. - Fruit of the Poisonous Tree: Why Google's 10% take rate is an anti-competitive advantage stemming from their illegal tying of the publisher ad server. - Opting Out of Google Search: Why publishers are finally blocking Google's AI crawlers as AI overviews cannibalize their organic traffic. Chapter Timestamps: 00:00 Intro: Restraining Ad Tech Opinions 01:20 The Rise of AI-Generated LinkedIn Takes 04:06 Flooding the Zone & Media Pollution 06:21 Alex Karp, Ontology, and the Need for Data Labeling 10:12 Sincera, Jounce Media, and Ad Placement Taxonomy 15:09 Google's Buyer Direct vs. Agentic Workflows 21:26 Why Automation Doesn't Beat Architecture 23:21 Fruit of the Poisonous Tree: Google's 10% Margin 25:33 The Overlord Model vs. Free Financial Markets 33:43 SMB Advertising, Taboola, and Outbrain 38:03 Ad Tech for Car Dealerships & Healthcare 40:04 Warner Bros. Discovery Builds on AWS 44:24 Information Asymmetry and CTV Mispricing 47:39 Publishers Opting Out of Google Search 56:40 The Token Economy: Anthropic vs. Fable AI

  7. 2 Jul

    Barrage of Buzzwords

    Adam and Gareth return from their post-Cannes Lions haze to break down a wild week of acquisitions, executive manifestos, and deleted LinkedIn posts. The guys kick off with a recap of the Cannes "boondoggle," where ad exchanges and platforms throw massive yacht parties and take over entire hotels to woo CMOs. Shifting to the tech stack, they discuss how AI is supercharging CRMs and customer service, including the shock for Pylon's CEO when their Claude enterprise pricing jumped from $600k to $3M. Adam and Gareth then dissect Sir Martin Sorrell's critique of holding company bureaucracy, as well as Keith Weisberg's quickly deleted post calling out agencies for pitching a "barrage of buzzwords" rather than proving their AI capabilities. They also unpack the juicy details of a controversial Meta tell-all book, Walmart's rumored $1.4 billion acquisition of Vibe, Satya Nadella's AI manifesto, and Alex Karp's buzzword-heavy Palantir-Zeta Global partnership. Key Topics: - The Cannes Lions Boondoggle: Why the festival is less about shared events and more about massive corporate flexing and yacht parties. - The Cost of AI in SaaS: How tools like Pylon and Granola are revolutionizing asynchronous communication, but facing massive enterprise API costs. - Holding Company Bureaucracy: Sir Martin Sorrell's debate on whether agencies should organize by market or capability. - Fake Agentic AI: Calling out the industry's habit of selling "a barrage of buzzwords" instead of showing real ROI and case studies. - The Vibe Acquisition: Breaking down the massive 17x multiple and $180M founder earnouts in Walmart's reported acquisition of Vibe. - The White-Collar Industrial Revolution: Satya Nadella's pushback on AI doomerism and why LLMs are uniquely targeting the bourgeoisie. Chapter Timestamps: 00:00 Intro & Cannes Lions Recap 11:27 AI in Sales, Pylon, & The Cost of Claude 19:03 Sir Martin Sorrell on HoldCo Taxonomy 28:04 Keith Weisberg Calls Out Fake Agentic AI 33:54 The Meta Tell-All Book "Careless People" 40:08 Walmart's $1.4B Acquisition of Vibe 54:18 Satya Nadella's AI Manifesto 58:48 Alex Karp, Palantir, and Zeta Global

  8. 12 Jun

    Hands on the Contract, Eyes on the Bag

    Adam and Gareth return for an absolute blockbuster episode of Adtech Ad Talk, kicking off with the biggest breaking news in the industry: Publicis and The Trade Desk have officially buried the hatchet and are back in the room together. To celebrate the end of this fierce corporate feud, the guys debut an incredible, AI-generated southern rap track summarizing the high-stakes reconciliation.The duo tackles the shifting dynamics of the AdTech landscape, breaking down why the hard binary split between the buy-side and sell-side is fundamentally dissolving as platforms move toward unified execution environments. Gareth unpacks how Meta's post-ATT pivot allowed them to perfect their "Andromeda" value function, driving an insane $50 billion optimization lift completely independent of traditional user IDs. Finally, they pull back the curtain on the "nasty" reality of dynamic take rates and hidden arbitrage in the middle, explore why the smartest marketing teams are born out of hyper-competitive industries like CPG, and feature a stunning, jarring viewer submission in Creative Corner.Key Topics:- Publicis & The Trade Desk Reunite: Breaking down the end of the feud and the power dynamics of their new handshake.- The Dissolving Buy/Sell Divide: Why legacy, siloed DSP/SSP integrations are giving way to single-environment optimization.- Meta's Post-ID Blueprint: How Apple's privacy changes forced Meta to build a superior, multi-feature recommender system.- The Dynamic Take Rate Trap: Unpacking how middleman platforms exploit averaging bounds to harvest massive margins from publishers.- Creative Corner: A powerful, unforgettable print ad highlighting period poverty and institutional shame.00:00 Intro: An Amped-Up Friday Blockbuster01:36 World Premiere: The Publicis & Trade Desk Reconciliation Rap05:23 Pre-Cannes Musings & The Certainty of Change06:18 The History of Failed DSP/SSP Mergers (Rubicon & Chango)08:08 The Death of the Legacy Buy-Side vs. Sell-Side Mindset11:32 Harvesting the Spread: The Advertiser's Right to Alpha14:18 The User ID Imbalance and Pacing Complexities16:02 How Apple's ATT Birthed Meta's $50B Recommender Engine18:41 The Looming AI Carnage for Unprepared Agencies22:55 The Rise of MCP and Direct Data Set Access25:31 Smartitude: Ranking Clients by Industry Competition26:30 The "Mafia State" of Publisher Arbitrage and Last Look28:29 Why Competition is the Only Way to Raise Publisher CPMs30:20 The Secret Rules of Transparency Pricing in CTV32:55 Beyond Taxonomies: Scoring Attributes on a Spectrum34:36 The Legacy of Cognitive (The OGs of Custom Modeling)35:46 Obstacles in the Trade Desk API (An Accidental Origin Story)37:02 Creative Corner: A Uniquely Jarring Print Ad from South Africa37:37 Outro: Bringing Back the Beat

See All (62)

About

Buy Side Meets Sell Side. Live Q&A with CEOs telling the truth about where Ad Tech is headed.

Information

  • Creator
    Adam and Gareth
  • Years Active
    2025 - 2026
  • Episodes
    62
  • Rating
    Explicit
  • Copyright
    © Adam and Gareth
  • Show Website
    AdTech AdTalk

You Might Also Like