AI Featured

This week: Platform battles and AI shifts dominate advertising

Google, Meta, and AI platforms clash over privacy regulations and advertising control as publishers struggle with AI crawler economics and platforms race to deploy agentic features for holiday shopping campaigns this week.

Luis Rijo

Nov 22, 2025 • 17 min read

Meta and Google compete for holiday retail ad dollars during Black Friday and Cyber Monday shopping season

The advertising industry faced a week of stark contradictions. While major platforms rolled out sophisticated AI tools promising autonomous shopping and campaign optimization, publishers discovered the harsh economics of artificial intelligence—one media company earned just $174 from AI crawlers scraping millions of pages. The gap between platform ambitions and publisher realities has rarely appeared wider.

Subscribe PPC Land newsletter ✉️ for similar stories like this one

Holiday automation drives platform updates

Google announced four new Demand Gen capabilities on November 17, targeting advertisers preparing for the compressed 2025 holiday season. The platform introduced brand suitability controls expanding to YouTube Home feed, watch next feed, and Discover placements on November 7. Three inventory modes now determine content adjacency thresholds: Standard inventory reduces exposure to strong profanity and dramatized violence; Expanded inventory maximizes reach by including potentially sensitive themes; Limited inventory applies the strictest filters for brands with rigorous guidelines around language and sexual references.

The update included asset-level experimentation for Demand Gen campaigns, allowing advertisers to test creative variations systematically. Google emphasized that organizations conducting monthly optimization tests achieve 3.1x return improvements over two-year periods when implementing 5% effectiveness gains per experiment. Video creative enhancements and AI-powered image automation rounded out the November "Drop" update, which builds on internal measurement showing Demand Gen advertisers achieved an average 20% increase in conversions during the first half of 2025.

Meta published comprehensive festive season guidance on November 16, arriving with just 12 days before Thanksgiving and 13 days before Black Friday. The platform emphasized Advantage+ automation, citing 22% average return on ad spend improvements throughout 2025. The two-phase framework divides festive preparation into third-quarter planning focused on goal definition and testing, followed by fourth and fifth quarter execution applying pre-season learnings.

Shopping activity patterns revealed in the guidance showed 15% of shoppers starting in late November and 18% beginning in early December. Meta deprecated legacy campaign APIs for Advantage+ structure consolidation in October 2025, with full migration required by the first quarter of 2026. This technical consolidation reflects Meta's strategic direction toward AI-powered campaign management as the default advertising experience. The guidance competes with similar resources from other major platforms, as Google launched agentic checkout and AI shopping tools on November 13.

Autonomous shopping faces skepticism

Google deployed autonomous capabilities across Search and Gemini platforms on November 13, introducing features enabling AI systems to complete purchases, contact stores, and process complex product queries. The company announced three primary capabilities: agentic checkout that autonomously purchases tracked items when prices drop, Duplex-powered phone calls verifying local inventory, and conversational shopping through AI Mode in Search and the Gemini app. These features mark a substantial expansion of Google's Shopping Graph infrastructure, now containing more than 50 billion product listings with 2 billion receiving hourly updates.

The integration represents a shift toward unified shopping capabilities across Google's AI products, following comprehensive advertising updates announced at Think Week 2025 in September. Shopping features previously limited to Search expanded to the Gemini app on November 13 for all U.S. users. The app processes shopping-related questions and returns product listings, comparison tables, pricing data, and purchase locations directly within conversations. Users can transition from brainstorming gift ideas to browsing specific products without leaving the Gemini interface.

Merchant eligibility for agentic checkout depends on integration with Google's payments infrastructure. The company confirmed Wayfair, Chewy, Quince, and select Shopify merchants as initial participants, with additional retailers scheduled for future inclusion. Google did not specify technical requirements for merchant participation or commission structures. The autonomous shopping deployment occurs amid broader industry movement toward agentic AI systems. However, analysis published on PPC Land in October 2025 identified eight structural challenges to agentic commerce adoption, including retailer incentives against AI intermediation, high ecommerce return rates, and consumer preferences for evaluating options before purchasing.

Measurement infrastructure expands access

Google announced substantial changes to incrementality testing capabilities on November 11, reducing the minimum experiment budget from previous thresholds approaching $100,000 to just $5,000. The platform implemented improved statistical models delivering up to 50% more conclusive results and faster reporting capabilities enabling custom test sizes and preferred statistical confidence levels within the Google Ads interface. The features became available across Video, Discovery, and Demand Gen campaigns immediately, while Display, Search, Shopping, and Performance Max campaign access requires coordination with Google account representatives.

Google released its Meridian Marketing Mix Model platform in January 2025, employing Bayesian causal inference to combine prior knowledge with observed data. The open-source framework incorporates incrementality experiment results regardless of channel or methodology, calibrating models with real-world experimental outcomes. Conversion Lift measurement data becomes available at product or brand levels within the Lift Measurement table in Google Ads. Advertisers can access granular results by selecting specific studies. The interface includes column customization options enabling Conversion Lift metric visibility alongside standard campaign performance indicators.

Follow on Google, Google News, X, LinkedIn, Mastodon, Bluesky, or via RSS

The measurement improvements complement broader platform developments including enhanced customer lifecycle targeting options announced in April 2025. Those features enable differentiated bidding strategies for new customer acquisition versus retention objectives, creating additional use cases for incrementality measurement to validate lifecycle targeting effectiveness. The announcement follows comprehensive measurement improvements unveiled at Think Week 2025 in September, which introduced the initial $5,000 threshold reduction alongside AI-powered advertising innovations.

Technical infrastructure receives attention

Google published updated documentation for its crawling infrastructure on November 20, expanding technical specifications for webmasters managing crawler interactions with their websites. The documentation updates provide detailed information about HTTP caching implementation, supported transfer protocols, and content encoding standards used by Google's automated systems. Google's crawling infrastructure now supports heuristic HTTP caching as defined by the HTTP caching standard. The implementation specifically utilizes ETag response and If-None-Match request headers, alongside Last-Modified response and If-Modified-Since request headers.

The documentation specifies support for Brotli, Gzip, and Deflate content encoding methods. Transfer protocols detailed include HTTP/1.1, HTTP/2, and HTTP/3 with QUIC support. The updates provide website administrators with more precise technical specifications for managing crawler interactions, addressing server performance concerns, implementing effective caching strategies, and optimizing crawl budgets. The documentation helps sites configure systems appropriately for Google's crawlers while maintaining server stability and efficient resource utilization across the search engine's distributed crawling infrastructure.

Display & Video 360 implemented exchange name changes during the week of November 17, reflecting updated branding and partnership arrangements. Three exchanges received updated names in reporting systems: Magnite Streaming became Magnite SpringServe, Xandr changed to Microsoft Monetize, and Criteo—Commerce Grid updated to Criteo Commerce Grid. The changes require advertisers to adjust saved reports, custom dashboards, and automated workflows relying on exchange-level reporting data for optimization decisions and supply path analysis.

The Magnite Streaming to Magnite SpringServe change acknowledges the company's April 23, 2025 announcement combining its streaming ad server with supply-side platform capabilities. The unified platform streamlines buyers' connection to 99% of U.S. streaming supply, with initial clients including Disney Advertising, LG Ad Solutions, Paramount, Roku, Samsung, and Warner Bros. Discovery. SpringServe added machine learning capabilities on October 24, 2025, to optimize ad pod construction for connected television publishers, reducing redundant bid requests while maintaining yield and competitive separation between brands.

Search ranking volatility persists

Google Search ranking volatility heated up around November 7, prompting widespread discussion across SEO communities. The so-called "Movember" update followed previous fluctuations identified as the Halloween update and movements between October 15-17. SEO professionals reported mixed signals, with some restaurant chains showing steady seasonal growth while others experienced 20% search traffic losses on thin location pages lacking physical addresses. Third-party tracking tools showed measurable spikes, though SEO community chatter remained more subdued than previous updates.

The pattern continued with additional ranking volatility detected around November 20, potentially related to either the Cloudflare outage that occurred during this period or the rollout of Gemini 3 models. Some SEOs speculated Google Search might have quietly incorporated Gemini 3 models into core ranking systems, though this remained unconfirmed. The reduced visibility stems partly from Google's October decision to disable the num=100 parameter, which had allowed viewing 100 results per page—a tool many SEO professionals relied upon for validation. Since this change, third-party tracking tools have been recalibrating their methodologies while SEO professionals increasingly depend on direct traffic observations rather than tool confirmations.

Gemini 3 powers AI Mode expansion

Google's AI Mode received Gemini 3 integration on November 19, bringing what the company described as "incredible reasoning capabilities" to Search. The new model can understand complex questions and provide detailed responses through generative layouts that go beyond simple text answers. Examples from early users demonstrated Gemini 3 creating kaleidoscope simulations, visualizing constellation positions over specific horizons, explaining DDoS attacks, and illustrating how neural networks function through interactive diagrams.

The generative UI capabilities represent a significant advancement in human-computer interaction, allowing AI Mode to produce custom visualizations, infographics, and interactive content responding to specific queries. However, not all responses received positive reception. Recipe bloggers expressed frustration over what they characterized as plagiarism, with one noting the level of content reproduction "is now completely out of control." The concerns highlight ongoing tensions between AI systems that can generate comprehensive responses and content creators who view such capabilities as appropriating their intellectual property without compensation.

Google simultaneously released Nano Banana Pro on November 20, built on Gemini 3 Pro and representing the company's most advanced image generation model. Nano Banana Pro became available in Search starting with AI Mode for Google AI Pro or Ultra subscribers in the U.S. The tool can create detailed infographics for educational purposes, produce marketing materials with better brand alignment through advanced reasoning that understands prompts and brand guidelines, and generate reference images ensuring critical brand consistency across campaigns.

Google Ads received access to Nano Banana Pro, enabling advertisers to create, edit, and test ad creatives with greater velocity. The platform emphasized better brand alignment through advanced reasoning that understands prompts, brand guidelines, and reference images to ensure consistency and authenticity across campaigns. The release builds on Google's broader push into generative AI for advertising, following earlier introductions of Asset Studio and other AI-powered creative tools throughout 2025.

Google updated its list of countries prohibiting offline gambling advertising on November 19, adding 23 countries to match locations where offline gambling is not allowed. The platform also began testing a much larger disclaimer on AI Overviews reading "Double-check important information" and "It's a good idea to check info in multiple places" with a "Learn more" button. This expansion from April's smaller "Check important info" notice reflects ongoing concerns about AI-generated response accuracy.

Buy ads on PPC Land. PPC Land has standard and native ad formats via major DSPs and ad platforms like Google Ads. Via an auction CPM, you can reach industry professionals.

Learn more

Legal battles escalate over content controls

Google, YouTube and Meta filed separate federal lawsuits on November 13, challenging California Senate Bill 976's restrictions on personalized content feeds for users under 18. The complaints, filed in the United States District Court for the Northern District of California, target specific provisions requiring parental consent before minors can access personalized feeds and mandating default settings that disable such features. Both companies argue these restrictions violate their editorial rights to curate third-party content and burden minors' ability to access speech.

Google and YouTube filed their complaint as Case No. 5:24-cv-07886, while Meta filed separately as Case No. 3:25-cv-09792. Both cases name California Attorney General Rob Bonta as defendant in his official capacity. The lawsuits follow a complex procedural history involving NetChoice, an internet trade association representing both companies. NetChoice filed its lawsuit challenging SB 976 on November 12, 2024. Before the Act's January 25, 2025 effective date, a federal court granted a preliminary injunction in part but declined to enjoin most provisions.

The court rejected NetChoice's facial challenge to the personalized-feed restriction because NetChoice had "not made a record ... to show facial unconstitutionality." The court also concluded that NetChoice lacked associational standing to raise an as-applied challenge for its members, reasoning that "NetChoice's individual members" would need "to participate in this lawsuit" for the court to decide the as-applied First Amendment claims. The Ninth Circuit affirmed the lower court's ruling on November 6, 2025, holding that NetChoice lacked associational standing to mount an as-applied challenge and that NetChoice failed to show those provisions were facially unconstitutional. The Ninth Circuit denied NetChoice's petition for rehearing and issued its mandate on November 13, 2025—the same day Google, YouTube and Meta filed their individual lawsuits.

Privacy settlement creates programmatic uncertainty

Google recently settled a class-action lawsuit in California by agreeing to create an off-switch for data sharing in bid requests, marking a significant development in programmatic advertising privacy. The RTB Control, as the mechanism is being called, would apply to Google account holders and remove all tracking data from bid requests, including pseudonymous identifiers. That means no cookies, IP addresses, device IDs, encrypted Google user IDs, user agent details or any other identifying or potentially identifying information.

The settlement could set a precedent by providing a template that regulators, lawmakers and privacy advocates point to when pushing for similar controls elsewhere, both in the United States and abroad. Privacy attorney Alan Chapell noted that if the mechanism gets codified, regulators in U.S. states or the European Union will likely insist on similar capabilities. The settlement hasn't been finalized, with the earliest possible date for a court hearing to approve it set for January 13, 2026. The mechanism doesn't exist yet, and it remains unclear exactly how it would work. It's an opt-out system, meaning users with Google accounts would have to manually turn it on to opt out of data sharing.

Despite Google having agreed to the settlement, a Google spokesperson emphasized that the company does not sell or share personal data and already has the strictest real-time bidding restrictions in the industry. Google emphasized it doesn't include any personally identifiable information in bid requests or share sensitive details like health, race, political views or precise location. What does get shared are things like page URL, app ID and general location based on a broad group of users. The settlement addresses a persistent concern: that Google rarely gives ground unless it's confident the ground isn't all that valuable. If this proposed mechanism makes Google's RTB ecosystem less efficient, it could open the door for rivals and spark new waves of innovation.

Publisher economics face AI reality

A publishing company earned just $174 from AI crawlers over an extended period, exposing the harsh economics facing media companies as artificial intelligence reshapes content distribution. The data point, revealed on November 20, stands in stark contrast to the existential threat AI has posed to the media industry. In June, The Wall Street Journal called AI an "armageddon" whose downstream effects on search traffic and programmatic ad revenue have upended the economics of digital publishing. In May, Business Insider cut nearly a quarter of its staff, citing AI as the primary reason.

The meager revenue from AI crawlers highlights a fundamental imbalance: while AI companies scrape millions of pages to train models and power answer engines, publishers receive minimal compensation despite bearing the costs of content creation, hosting, and bandwidth. The situation has prompted various responses from publishers, ranging from blocking AI crawlers entirely to filing lawsuits demanding fair compensation. Some large outlets, including Vox Media, Axel Springer, and People Inc., have signed bespoke licensing deals for use of their data. But the vast majority of websites have received no payment in exchange for AI firms using their content to train their models or fuel answer engines.

The economics become even more challenging when considering the infrastructure costs publishers bear. AI crawlers create significant server load, bandwidth consumption, and resource strain on publishing systems. When these costs are weighed against the $174 in revenue generated, the business case for allowing AI access to publisher content appears fundamentally broken. The data underscores the urgency behind industry efforts to establish sustainable compensation frameworks, whether through usage-based fees, revenue-sharing arrangements, or other mechanisms that acknowledge the value publishers provide to AI systems.

Measurement shifts at streaming platforms

Netflix ads reached more than 190 million global Monthly Active Viewers in October, the company announced at a press roundtable on November 6. According to VP Mitzi Reaugh, the new MAV metric represents members who have watched at least one minute of ads on Netflix per month, multiplied by the estimated average number of people per household. MAVs take the place of MAUs – monthly active users. Reaugh noted one advantage is that the MAV metric better captures co-viewing behaviors, which explains Netflix's preference for the new measurement approach.

MAVs also include greater viewership outside of ad-supported plans, like during livestreamed events. Switching from MAUs to MAVs won't affect how Netflix sells ads, said President of Advertising Amy Reinhard. Instead, the goal is to be "more transparent and clear about how our audiences are interacting with the service." The metric change reflects streaming platforms' ongoing efforts to provide advertisers with more accurate audience measurements that account for multiple viewers per household and viewing behaviors that don't fit traditional one-user, one-account models.

The measurement update arrives as streaming platforms continue expanding their advertising businesses and competing for brand budgets. By multiplying household accounts by estimated co-viewing numbers, MAVs provide a larger addressable audience figure than MAUs while claiming to more accurately reflect actual viewing patterns. Whether advertisers embrace this new metric or view it skeptically remains to be seen, but the change demonstrates how streaming platforms are working to distinguish their measurement approaches from traditional digital advertising metrics.

Alternative browsers show growth signals

Brave Browser generated $25 million in revenue in the first quarter of 2025, most of which came from the search ads product built into its browser. The Information reported this data on November 12, highlighting Brave as perhaps the most successful browser that general web users have never heard of. Brave distinguishes itself by building its own search index, a pricey commitment that most other search engines avoid by licensing Bing or Google's search index as the core of their offerings.

The revenue figure demonstrates that privacy-focused browsers can establish viable business models through search advertising, even while operating at scales far smaller than dominant players. Brave's approach of building its own index positions the browser to benefit from potential regulatory changes requiring Google to share search index data with rivals. The federal judge in Google's search monopoly case decided that Google "must share some of its valuable search index data with rivals," creating a glimmer of hope for search engines wanting to quickly build up their own indexes without bearing the full infrastructure costs Brave has shouldered.

The business model success of a privacy-focused browser generating meaningful revenue through search ads presents an interesting counterpoint to industry assumptions about the necessity of extensive user tracking for advertising viability. Brave's $25 million quarterly revenue, while modest compared to industry giants, suggests that alternative approaches combining privacy protection with contextual advertising can sustain operations and potentially scale. Whether this model can grow to compete effectively with established platforms remains an open question, but the data point provides evidence that the advertising ecosystem contains more diversity than the dominance of major platforms might suggest.

European regulatory landscape shifts

The European Union published its Digital Omnibus proposal on November 19, introducing a bundle of updates meant to simplify Europe's digital rulebook and remove complexity that has built up since the General Data Protection Regulations and other tech laws came into force. The "simplification package" includes fundamental changes affecting how data can be used, how AI is trained, and how companies across advertising, media, and tech operate. The European Commission set a clear target to deliver at least 25% reduction in administrative burdens, and at least 35% for small and medium-sized enterprises, by the end of 2029.

The Digital Omnibus arrives as the advertising industry continues navigating the complex regulatory landscape created by GDPR, the Digital Services Act, the Digital Markets Act, and various other European technology regulations. While positioned as simplification, the changes may have unintended consequences for different market participants. Some provisions could provide relief for ad tech companies struggling with compliance burdens, while others might hand more power to Big Tech and AI agents by clarifying their operational parameters.

The package reflects European regulators' ongoing efforts to balance innovation, competition, and consumer protection in digital markets. Whether the simplification achieves its stated goals of reducing administrative burdens without compromising privacy protections and competitive dynamics remains to be seen. The 25% and 35% reduction targets represent ambitious commitments that will require substantial changes to existing regulatory frameworks and enforcement mechanisms.

Display advertising faces pressure

Brands are set to cut open web display spend 30% in response to AI search, according to Forrester analysis published on November 6. Zero-click search adoption has critically affected web traffic from search engines to publishers in the year since Google first introduced its Overviews and AI Mode features. In 2026, advertisers may very well respond by cutting their display investments with publishers on the open web by 30% in favor of connected television and paid social. The research outfit predicts that, as the addressable audiences reachable via publishers shrink, brands will look to find them elsewhere.

Tim Lathrop, vice president of platform digital at Mediassociates, confirmed his clients had reduced their open web display spending by 20-30% during 2025, without providing exact figures. Clients are shifting budget to reallocate toward channels offering better measurement and performance visibility. Meanwhile, connected television and ad-supported streaming video offered "storytelling" capabilities, increasingly alongside shoppable and interactive formats supporting brands' e-commerce ambitions. Jonathan Geller, co-founder of indie full-service agency Lower Cross, questioned the value proposition: "Why would we allocate a budget to something that's almost unmeasurable, that we really can't see the performance lift of without doing tons of modeling… Why not throw that more into something where you know what the input and the output is?"

The projected 30% decline in display spending represents a fundamental shift in advertiser confidence about the open web's ability to deliver measurable results. This isn't about ad spend leaving the market—dollars are reallocating to channels perceived as more accountable and effective. However, the IAB estimated in January that digital display's overall share of global ad spend would fall 1% this year, to 12.8% of overall media investment. Its annual Outlook study also projected annual growth in display investment falling from 7.4% in 2024 to 4.4% in 2025.

Publishers recognize the shift but face uncertainty about paths forward. "You can't rely on Google anymore," Donna Ogier, Reach plc's director of U.S. audience, said at the Digiday Publishing Summit Europe in Lisbon. Another publishing executive speaking under Chatham House rules at the Summit's Town Hall session described the impact: "The last upgrade… the change Google made, has just decimated Google traffic. To my mind, permanently. It's not coming back… You look at the screen when you search for Google, you don't see the listings anymore. You see the snippet or the overview, or a couple of Wikipedia bits, even a couple of pictures."

Publishing diversification accelerates

Forbes launched an AI-powered dynamic paywall on November 14, rolling out the system to 20% of its audience on October 17 before expanding to the entire Forbes audience. The move is part of Forbes' broader push to diversify its commercial revenue for a post-search media era where open-web traffic, and the ads that ride on it, can no longer be taken for granted. Forbes CEO Sherry Phillips revealed at Digiday's Publishing Summit Europe that the publisher saw a 40% decline this year in search referral traffic.

The AI-powered dynamic paywall learns from reader behavior, adjusting in real time to show the right subscription offer to the right person at the right moment, instead of everyone getting the same "subscribe now" message. For instance, if a person clicked on a Forbes story about small-business tax strategy, the AI will notice if they're a repeat visitor on that topic, that they've read business-finance stories all the way to the end, and linger on expert-driven content. The system then tailors subscription offers based on these behavioral signals, potentially improving conversion rates beyond static paywall approaches.

Subscriptions are moving to the center of more traditionally free-to-access publisher revenue strategies as they look to reduce future reliance on open-web digital advertising. Zero-click search eats into programmatic revenue, forcing publishers to find alternative revenue streams. But converting readers isn't easy when willingness to pay for digital subscriptions has stagnated, according to the Reuters Institute's 2025 Digital News Report. For Forbes, the AI-powered dynamic paywall may be what helps it unlock new paying subscribers who have previously been difficult to reach through one-size-fits-all conversion strategies.

Subscribe PPC Land newsletter ✉️ for similar stories like this one