A federal judge in New York ruled on Wednesday that investors can proceed with a lawsuit accusing Zeta Global of misrepresenting how it obtained consumer consent for its advertising data, a decision that landed in the same week two separate European bodies moved to tighten what counts as legitimate consent for data collection in the first place. Taken together, the three developments, one from a US court, two from European regulators and a European court ruling that predates them, describe a narrowing of the legal room advertising technology companies have historically relied on when they describe their data as consensual, opted-in, or freely available.

U.S. District Court Judge Dale Ho rejected Zeta Global's bid to dismiss a class-action complaint at an early stage of the proceedings, writing in a 29-page ruling that the investors, who include the Allegheny County Employees Retirement System, adequately alleged materially misleading statements regarding Zeta's opted-in data set, according to MediaPost. The complaint centers on claims that Zeta publicly represented, between February and October 2024, that it held a data set of more than 240 million opted-in individuals in the United States, only for an outside report published in November 2024 to raise serious concerns about how that consent was actually procured. Zeta's stock price fell 37 percent the day that report came out, the complaint states. By February 2025, the company's own securities filings still described more than 240 million individuals in its data but had dropped the word opted-in entirely. Among the specific allegations, the plaintiffs pointed to Zeta's commenting platform Disqus, arguing it gathered data under what amounts to an opt-out framework rather than a genuine opt-in one, since a user must first sign up and share information before separately accessing a privacy policy to opt out. Zeta's spokesperson said the company denies the claims and expects to prevail once the case is fully determined, adding that the ruling was not a decision on the merits and that the court was required at this stage to accept the complaint's allegations as true. Judge Ho ordered both sides to submit a letter by July 22 outlining next steps.

The Zeta ruling arrived four days after a separate, larger reckoning over consent took shape on the other side of the Atlantic. A ruling issued by the United States Supreme Court on Monday, June 29, has stripped the Federal Trade Commission of the structural independence that underpins the entire legal architecture of the EU-US Data Privacy Framework, according to Austrian privacy group noyb, which PPC Land reported on July 12. The decision, Trump v. Slaughter, found that the FTC's long-standing insulation from presidential removal is unconstitutional, resolving a question that traces back to a 1935 precedent called Humphrey's Executor. The certiorari grant itself was not a total surprise to close observers of American administrative law; a Fifth Circuit concurrence in an unrelated FTC enforcement matter involving Intuit's TurboTax advertising had already flagged, in March 2026, that the Supreme Court was weighing whether FTC commissioners could be shielded from removal, citing the same docket entry months before the ruling landed. Because the European Commission's 2023 adequacy decision cites that same FTC independence more than 250 times as proof that American oversight meets European Union standards, noyb chair Max Schrems argues the foundation for transatlantic data flows no longer exists. "Given that there are no independent authorities in the US anymore, we call on the European Commission to orderly withdraw the adequacy decision on the US," Schrems wrote in a letter sent to the Commission on June 30. Commission Implementing Decision (EU) 2023/1795 remains formally in force for now; it stays that way until the Commission repeals it or the Court of Justice of the European Union annuls it, and noyb has said it will file its own annulment lawsuit within weeks if Brussels does not act first, litigation the group estimates could take two to three years to resolve. Google adopted the framework for its advertising services in September 2023, meaning the ruling touches a company whose ad products sit at the center of the industry PPC Land covers daily, alongside every other advertiser and cloud infrastructure provider that built compliance architecture on the assumption of a durable FTC backstop.

A third and more granular consent question was settled the same week by the body that actually writes the European Union's data protection guidance. The European Data Protection Board adopted Guidelines 03/2026 on web scraping in the context of generative AI on July 7, 2026, a 22-page framework that concludes consent will most probably not serve as a workable legal basis for the kind of large-scale scraping that trains AI models, according to PPC Land. The guidelines state plainly that a person making their data available on an openly accessible web page has not thereby consented to that data being scraped for a purpose such as AI training, and that the mere absence of a robots.txt file does not amount to consent either. That leaves legitimate interest as the legal basis the board expects private companies to lean on most often, subject to a three-part test: a genuine interest pursued by the controller, the necessity of processing personal data to achieve it, and a balancing exercise weighing the data subject's rights against that interest. The balancing test carries real teeth. Where a platform blocks scraping through robots.txt and CAPTCHA and expressly states it will not permit its user data to train AI systems, the guidelines conclude a scraper cannot reasonably expect to use that data regardless of what it does anyway, and the board notes starkly that once a model is trained, personal data cannot easily be deleted from it. Public consultation on the guidelines runs until October 30, 2026, the same closing date the board separately set for its updated anonymisation framework.

The guidelines also settle a question that has sat unresolved across the marketing technology stack for years: who counts as the legally responsible party when scraping itself is outsourced to a third party rather than performed in-house. Where a specialist scraper builds a training dataset on behalf of an AI developer, following documented instructions about which sources and categories to collect, the scraper functions as a processor and the developer as controller, since the developer determines the purposes and essential means of the processing. Where a developer instead buys or reuses a dataset someone else already scraped, the two parties become separately responsible for their own distinct processing activities, and the guidelines state explicitly that the original scraper is not, in principle, responsible for how a downstream buyer later reuses the data. A third configuration, joint controllership, arises only when two organisations jointly decide to build a model and jointly determine the scraping criteria, even where one company performs the technical extraction and the other trains the resulting model.

Read side by side, the Zeta lawsuit, the Supreme Court ruling, and the EDPB's scraping guidance describe the same underlying pressure from three different directions. A US court is now testing whether a company's own description of its consent practices holds up under scrutiny once litigated. A constitutional ruling has knocked out the enforcement mechanism the European Union relied on to trust that American oversight of consent claims was independent in the first place. And Europe's own regulators have concluded that consent, as a legal basis, does not scale to the volume at which modern AI training actually operates, pushing the industry toward a legitimate-interest test that explicitly weighs whether a person could have reasonably expected their data to be used this way. None of the three developments alone forces an immediate change in how advertising technology companies operate. Together, they narrow the set of arguments available to any company that has historically pointed to consent, broadly and loosely defined, as sufficient legal cover for how it collects or uses personal data.

Google makes AI ad labels mandatory and unremovable

A second thread running through the same window concerns disclosure rather than collection, and it produced a policy change with an unusual feature: once applied in certain circumstances, the label cannot be taken off. Google updated its advertising policies on July 9, 2026, introducing what it calls the AI label setting, a control rolling out gradually through July across five products, Google Ads, Display and Video 360, Campaign Manager 360, Merchant Center, and Ads Editor, according to PPC Land. The mechanism works on two layers. Any ad containing an asset an advertiser designates as AI-created or AI-edited will carry a disclosure inside the "How this ad was made" section of My Ad Center, a layer that exists globally regardless of where the ad runs. For campaigns targeting the European Union, India, or New York specifically, flagged assets also trigger a visible overlay directly on the ad itself, meaning the same creative can look different depending on where it serves.

The unusual part concerns when Google applies a label without the advertiser touching the setting at all. Google's own documentation lists three circumstances: when the company is legally required to label an ad in a given region, when it receives signals from other platforms indicating AI involvement, or when an advertiser uses one of Google's own fully automated tools to generate and serve creative. In each of those cases, according to the documentation, labels cannot be overwritten, and there is no interface control to remove one applied this way. An advertiser who never opens the AI label setting can still end up with a labeled ad. Beneath the visible layer, Google also embeds SynthID, a non-visible digital watermark, alongside C2PA metadata describing an asset's creation history, into any image or video generated through its own tools, independent of whatever an advertiser selects in the interface.

The timing places the rollout close to a hard regulatory deadline. Rules in the European Union, India, and New York already require disclosure when ads contain certain AI-generated or AI-edited assets, and the EU AI Act's Article 50transparency obligations become applicable on August 2, 2026, less than a month after Google's changelog entry. MediaPost separately confirmed the same feature the previous day in a July 10 item titled "Google Adds Transparency To AI-Generated Ads," listed among the outlet's most recent MediaDailyNews coverage, giving the rollout independent confirmation across two of the six sources tracked in this edition. Google's documentation is careful to note that neither disclosure route, the automatic setting or an advertiser's own custom label, guarantees compliance with any specific law on its own, and it urges advertisers to seek legal guidance. For teams managing creative across more than one of the five affected products, the practical complication is less about any single button and more about consistency: an asset labeled inside Google Ads does not inherit that status inside Display and Video 360 or Campaign Manager 360 if it was uploaded separately into each, since every product keeps its own independent AI label field.

The mechanics differ noticeably across the five products because each interface was built around its own existing creative-management logic rather than a single shared review screen. Inside Google Ads, the AI label status lives as a filterable column visible from the asset library, with an in-product prompt to review assets appearing whenever a campaign contains creative that may require a label. Merchant Center folds the same choice into its existing image-management dialog, tucked inside a three-dot menu on each product asset. Display and Video 360 offers the widest range of entry points, including a bulksheet with a dedicated AI label column for large-scale uploads and a filter bar that separates assets into three states: on, off, and not set. Advertisers who prefer not to rely on the platform-generated disclosure at all can build a custom label using their own design tools, though Google's guidance specifies that any such label should sit outside a 5.5 percent margin around an asset's full perimeter, since automatic image cropping and enhancement features can otherwise cut a custom disclosure out of frame without warning.

Retail media platforms keep adding inventory faster than measurement can absorb it

A third pattern across the same window concerns retail and commerce media expanding into new formats and partnerships, spanning both physical and digital inventory, even as separate measurement data raised its own doubts about whether ad spend on the digital side is converting as claimed. Perion won the contract to power Best Buy Canada's in-store retail media network across all 308 Canadian stores, replacing legacy loop-based signage with an ad server, supply-side platform, and header bidding tools, according to PPC Land. The deal tests Perion's claim to run a full technology stack for physical retail advertising rather than a single point solution, and it lands in a category, digital out-of-home retail signage, where Best Buy's specific combination of scale and store count gives the contract outsized visibility for other retailers weighing similar upgrades.

Taboola made a comparable expansion move on the AI side of retail media. The company opened its DeeperDivenetwork, which reaches nearly 7 million monthly users and spans roughly 9,000 publishers, to outside AI answer engines and chatbots, letting those third-party assistants plug directly into Taboola's ad demand network rather than requiring publishers to negotiate separate monetization arrangements with each AI platform individually, PPC Land reported. The move effectively repositions an existing content-recommendation network as inventory for a category of traffic, AI-driven answer engines, that most publishers currently have no monetization pathway for at all.

Amazon, meanwhile, lowered the barrier to entry on the seller side rather than the ad-buying side. The company removed the login requirement from its Seller University training library, opening more than 125 free topics, ranging from listing products to fulfillment logistics, to anyone browsing rather than only to registered sellers, according to PPC Land. The change signals Amazon treating seller education as a top-of-funnel acquisition tool in its own right, rather than a resource reserved for people who have already committed to opening a store.

A fourth expansion arrived through the agency rather than the platform side. Conagra Brands named Barrows, a commerce and shopper marketing agency owned outright by WPP since earlier this year, as its United States commerce marketing agency of record following a competitive review, according to MediaPost. The assignment covers shopper and customer marketing, including digital shelf presentation, across a brand portfolio spanning Birds Eye, Duncan Hines, Healthy Choice, Marie Callender's, Reddi-wip, and Slim Jim. Lisa Matos, vice president of sales capabilities and commerce at Conagra, said Barrows stood out for its ability to connect shopper marketing, digital shelf, and retail media into a single system, language that echoes exactly the kind of unified stack Perion is now building for Best Buy Canada and that Taboola is building for AI-driven answer engines. Barrows' first Conagra programs are expected to launch in late July 2026.

Those four expansion moves ran alongside data suggesting the underlying measurement problem in digital and performance advertising has not eased in step with the growth. Channable data spanning 1.38 billion euros in ad spend found advertisers running Performance Max campaigns lost 46 percent of their return on ad spend within a single year, according to PPC Land, as the cost of clicks on Google's automated bidding system continued climbing faster than the incremental value those clicks delivered. Separately, Integral Ad Science found that invalid traffic remained 30 percent above its own forecast during the surge in advertising activity accompanying the FIFA World Cup, settling at a rate of 1.134 percent against a projected 0.876 percent, PPC Land reported, keeping automated bidding systems exposed to low-quality inventory through the remainder of July. Neither figure names a single culprit; both describe a gap between spend and verified outcome that keeps widening even as the platforms carrying that spend keep adding features. Read against the Perion, Taboola, Amazon, and Barrows moves, the pattern suggests commerce media inventory, and the agency and technology relationships built around selling it, is growing on the supply side considerably faster than the measurement tools tracking whether that spend actually converts are keeping pace. Each of the four companies is, in its own way, betting that unifying previously separate systems, in-store signage and programmatic bidding, publisher inventory and chatbot traffic, seller education and acquisition funnels, shopper marketing and digital shelf, will outrun the measurement gap rather than close it directly.

Also noted

  • July 12, 2026: Adobe found that 81 percent of brands using a unified AI-SEO strategy gained traffic, with Semrush's 126-million-prompt index tying the gain to travel-site AI traffic up 2,215 percent since 2024, though 45 percent of brands still cannot fully measure their AI visibility, according to PPC Land.
  • July 12, 2026: Google began letting Gemini agents run background tasks without breaking their connections, adding remote MCP links, custom function calls, and credential refresh to end a longstanding fragility in long-running automated tasks, per PPC Land.
  • July 12, 2026: Meta dropped an Instagram AI tagging tool after user objections while preparing Muse Image, a new generative tool that will give advertisers access to Advantage+ creative within weeks, according to PPC Land.
  • July 12, 2026: Ten European media groups launched an advertising marketplace explicitly pitched around data sovereignty and local processing, positioning itself against what the coalition describes as an 80 percent Google grip on the region's ad growth, with a pilot planned for September 2026, per PPC Land.
  • July 12, 2026: iHeartMedia opened its podcast and digital audio inventory to Amazon DSP for the roughly 1,000 sellers already buying through the platform, adding shopping, browsing, and streaming signals alongside reach into Twitch, Amazon Music, Fire TV, and Alexa, according to PPC Land.
  • July 13, 2026: Colle McVoy was named brand development partner for Allegiant, taking on identity work across Allegiant Air and Sun Country Airlines as the combined carriers serve 22 million customers annually across more than 650 routes, according to MediaPost.