Every year in advertising technology produces a few defining narratives. 2024 produced four — and they’re interconnected in ways that set up a substantially different industry landscape for 2025.
The cookie didn’t die. Google lost two antitrust cases. Retail media hit $60 billion with measurement standards still unresolved. And AI moved from road map to product reality in ways that changed what every major platform actually does.
This is where we are.
The Cookie Story: Preparation Won Even When the Deadline Disappeared
The most dramatic single news event in adtech in 2024 was Google’s July announcement that it would not deprecate third-party cookies in Chrome. After four years of preparation, billions of dollars of investment in alternative identity infrastructure, and an entire consulting industry built on cookieless futures, the deprecation was canceled.
The reflexive response from some corners was “all that preparation was wasted.” That reaction was wrong then and the data through year-end confirms it.
The organizations that built first-party data programs, deployed CAPI on their media platforms, implemented authenticated identity in their publishing infrastructure, and developed clean room capabilities are measurably better positioned than they were at the start of the preparation period — regardless of Chrome’s cookie status. Publishers with logged-in user bases are commanding 40%+ CPM premiums on identity-resolved inventory. Advertisers with complete CAPI implementations are running Meta campaigns with higher signal quality than their competitors. Clean room users are executing audience activations that weren’t possible before the infrastructure existed.
The cookie survived. The preparation infrastructure compounded value anyway. This is the story that the year-end data tells clearly: the companies that built for cookie indifference rather than cookie replacement won.
What 2025 inherits: the user choice prompt architecture Google is implementing will bifurcate Chrome users into cookie-accepting and cookie-refusing segments. Managing that bifurcation is the new operational question, and organizations with identity infrastructure are better equipped to do it than those without.
The Antitrust Story: Two Cases, Two Findings, One Open Remedies Chapter
2024 produced the most consequential antitrust findings against a major tech platform since the Microsoft case of the late 1990s.
In August, Judge Mehta found Google to be an illegal monopolist in general search — the first time a US court has ruled that Google’s core search business involved anticompetitive conduct. The remedy proceedings are active and the options on the table (Chrome divestiture, Android restrictions, search data sharing) are structural rather than merely behavioral.
The DOJ adtech trial concluded testimony in October. Judge Brinkema’s ruling is expected in early 2025, and the evidence presented — Project Jedi last look, Project Poirot header bidding throttling, DFP/AdX tying — is detailed and well-documented. The probability of a liability finding seems high based on the trial record; the question is remedy scope.
What 2025 inherits: two active remedy proceedings, certain appeals, and a multi-year timeline to structural change. But the competitive narrative has shifted. Publishers evaluating ad server alternatives, independent SSPs investing in direct demand relationships, and advertisers diversifying programmatic infrastructure away from Google-owned products are all making rational forward-looking decisions. The verdict doesn’t need to be final to change behavior.
The Retail Media Story: The Category Is Real, the Measurement Isn’t
US retail media spending crossed $60 billion in 2024, according to eMarketer’s full-year estimates. The category is real, the money is significant, and the major networks — Amazon, Walmart Connect, Kroger Precision Marketing, Instacart, Target Roundel — have all demonstrated that advertisers will pay premium CPMs for access to shopper first-party data.
What didn’t happen in 2024: meaningful standardization of measurement methodologies across networks. Advertisers who ran campaigns across multiple retail media networks ended 2024 in the same measurement predicament they started it — with incompatible attribution models, inconsistent closed-loop definitions, and no reliable way to compare performance across networks.
The IAB’s retail media measurement guidelines exist. The major networks participated in developing them. Adoption timelines remain undefined. The networks with the most advertiser leverage (Amazon) have no meaningful competitive incentive to standardize measurement in ways that would make comparison with other networks straightforward.
What 2025 inherits: retailer pressure from brands that are now sophisticated enough to demand incrementality measurement, or at minimum, methodological disclosure. The brands that entered retail media in 2021-2022 with minimal measurement sophistication have, in most cases, built internal capability that makes them harder to mislead with attributed sales that include baseline purchase behavior. The measurement pressure will increase.
The AI Story: From Roadmap to Reality
AI moved from advertising platform roadmap language to actual product architecture in 2024 in ways that are changing what campaign management means operationally.
Meta’s Advantage+ Shopping campaigns reached a maturity point where head-to-head test data consistently shows AI-automated campaigns outperforming manually structured equivalents — at least for advertisers with sufficient conversion volume and complete CAPI implementation. The “give up control to get better results” trade-off that Advantage+ represents is now backed by enough performance data that most direct response advertisers have made the transition or are actively running tests.
Google’s Performance Max became more prevalent by default — new campaign creation funnels advertisers toward PMax, and existing campaigns have migrated. The transparency concerns haven’t been resolved. The performance case in most accounts has been net positive even without transparency.
Google Marketing Live’s AI Overviews announcement changed the reference architecture for search advertising. The commercial query click volume question — does AI-summarized search reduce or maintain paid click opportunity — remains partially unresolved but early data through Q4 suggests commercial intent queries continue to produce click volume comparable to pre-AI-Overviews baselines.
Generative AI for creative is the story that 2024 set up for 2025. Every major platform — Meta, Google, Amazon, TikTok — has shipped AI creative generation tools in 2024. The tools are useful but not yet at the quality threshold where they replace human creative for premium brand advertising. For direct response assets at scale, they’re already operational. For brand creative where visual quality and brand voice matter, they’re not yet there. The gap is closing faster than most brand teams are prepared for.
What 2025 Is Set Up to Deliver
Four converging dynamics set up 2025 specifically:
First, Google’s adtech verdict will land and the remedies debate will intensify. Independent adtech will have a clearer understanding of its competitive opportunity within twelve months.
Second, retail media measurement pressure will produce at least partial standardization — either voluntarily from networks seeking to attract larger budgets, or through the IAB frameworks gaining enforcement teeth from major advertisers.
Third, the cookie bifurcation in Chrome — cookie-accepting vs. cookie-refusing users — will produce the first clean data on user behavior under choice architecture, which will tell the industry more about the actual post-cookie landscape than four years of theoretical preparation did.
Fourth, AI creative and AI campaign automation will move from test to standard for most digital advertisers. The agencies and in-house teams that haven’t built workflows for AI-assisted creative production at scale will face accelerating competitive pressure from those that have.
The industry is better equipped for each of these dynamics than it was twelve months ago. 2024 was a year that produced more consequential change than its surface-level narrative suggested.
FAQ
Q: What is the most important single action adtech companies should prioritize heading into 2025? For publishers: accelerating authenticated identity programs that generate CPM premiums on identity-resolved inventory while hedging against cookie bifurcation. For buyers: completing CAPI implementation on all major platforms and building incrementality measurement into retail media spending. For SSPs and DSPs: contingency planning for programmatic market structure changes resulting from adtech antitrust remedies.
Q: Did any meaningful cookie alternative technologies emerge in 2024 that have traction? UID2 reached meaningful publisher enrollment milestones and is the most scaled open-source identity alternative to cookies. ID5 has grown adoption particularly in Europe. Google’s Topics API is technically available but commercially underutilized. None of these have achieved the signal density that cookies provided at scale in Chrome — but all have genuine utility for cookieless traffic segments.
Q: How should brands think about allocating between retail media and traditional programmatic going into 2025? Retail media is most efficient for brands with strong existing product presence in a retailer’s physical or digital shelves and sufficient transaction volume to measure closed-loop outcomes. Traditional programmatic is more flexible for brand building, reaching audiences outside shopping contexts, and campaigns where retail purchase isn’t the conversion goal. The allocation decision should follow measured incrementality, not category trend.
Q: What is the biggest risk to adtech stability in 2025? The convergence of the Google antitrust remedies process with whatever Chrome cookie choice architecture actually gets implemented creates genuine structural uncertainty. If either remedies proceeding produces unexpected outcomes — aggressive Chrome divestiture, immediate DFP functional separation — the market could adjust faster than companies are prepared for. Scenario planning for a Google-less or reduced-Google infrastructure environment is genuinely prudent.