The advertising technology sector spent the first days of July 2026 confronting a question it has circled for years: who should own the infrastructure of digital advertising, and at what price? A takeover approach for Criteo from two private investment firms, disclosed on July 6 and dissected across the trade press through July 8, put a number on that question. So did the installation of a new chief executive at Integral Ad Science on July 7, one year after the measurement company itself changed hands in a 1.9 billion dollar buyout. Both moves landed in the same 48 hours as the twentieth edition of IAB Europe's AdEx Benchmark Report, released July 7 in London, which valued Europe's digital advertising market at 131.1 billion euros and documented where the money now flows: video, social, and retail media.

Regulators supplied the counterweight. The General Court of the European Union dismissed Apple's challenge to its Digital Markets Act gatekeeper designation on July 8, closing three joined cases that had been running since late 2023. The European Data Protection Board opened a public consultation the same day on a 33-page framework that replaces the twelve-year-old test for when advertising data counts as anonymous. And in the search channel, Google extended Search Console to Instagram, TikTok, X and YouTube accounts on July 7, a measurement expansion that arrived just as publishers, platforms and analysts traded increasingly blunt assessments of what search traffic is still worth.

This edition covers those five stories in depth, drawing on reporting published between July 6 and July 8, 2026 by PPC Land, AdExchanger, Digiday, Search Engine Roundtable, Adweek and MediaPost.

A 50 percent premium for Criteo, and a leadership reset at IAS

The week's defining transaction story began quietly, with two unnamed sources and a share price that would not sit still. Vista Equity Partners and Quinti Capital submitted an offer to acquire Criteo at a premium of more than 50 percent to its recent share price, two sources familiar with the matter told Reuters, in reporting published July 6 and covered by PPC Land on July 7, 2026. Bloomberg News surfaced the approach first, earlier the same day. Criteo's Nasdaq-listed shares closed 21.4 percent higher at 23.17 dollars on Monday, July 6, giving the Paris-headquartered commerce media company a market value of roughly 1.16 billion dollars. MediaPost's July 7 report captured the intraday picture, noting Criteo's American depositary receipts up 18 percent at 22.49 dollars by mid-afternoon in New York.

The mechanics of the offer remain partly obscured. The sources described the bid as having been submitted during the week of June 29, without specifying a date, and neither Vista, Quinti Capital nor Criteo has confirmed the terms publicly. The premium is measured against the recent share price rather than the post-news close, a distinction that matters because any calculation performed after Monday's 21.4 percent jump would understate the gap between where the stock traded before the approach and what the buyers propose to pay. Whether the offer is all-cash, whether it relies on the debt financing typical of leveraged buyouts, and whether Criteo's board has formally reviewed it all remain unknown. What is established is that Criteo has not yet decided how to respond, and that any completed deal would require shareholder approval and, as AdExchanger's July 8 roundup observed, presumably a delisting from the Nasdaq.

AdExchanger's assessment carried a sting worth registering. The publication noted that the offer, which lands 50 percent above a market capitalization of roughly 1.1 billion dollars, would have read as a lowball as recently as eighteen months ago, when Criteo's stock still performed well. It also placed the approach inside a longer pattern: Criteo has explored a sale multiple times without closing one, including talks with the retail media firm Skai as recently as last year. Digiday's July 8 coverage by Ronan Shields struck a similar note, framing the bid as another twist in a long-running saga over the fate of the France-founded firm, and pointing out that Criteo's market capitalization had fallen 23 percent year over year before the takeover speculation lifted it back above 1 billion dollars.

Why a buyer sees an opening now

The financial arithmetic behind Criteo's discounted valuation traces to a single client decision disclosed fourteen months ago. On May 2, 2025, Criteo revealed that its largest retail media client would discontinue managed services and curtail brand demand sales, a change the company estimated would remove 25 million dollars from 2025 revenue and as much as 75 million dollars across the first ten months of 2026, as documented in PPC Land's coverage of the Q1 2025 disclosure. The damage accumulated across subsequent quarters. Fourth-quarter 2025 revenue fell 2 percent to 541 million dollars. First-quarter 2026 results, announced May 6, 2026, showed GAAP revenue down 6 percent to 424.6 million dollars and net income down 79 percent, even as total media spend crossed 1 billion dollars in a single quarter for the first time. That divergence between the dollar volume flowing across the platform and the revenue Criteo recognizes directly is itself a story about the company's transition from managed services toward a platform model, and it is precisely the kind of multi-year restructuring that private equity owners prefer to execute away from quarterly earnings scrutiny.

Criteo also spent the past nine months dismantling the legal barrier that once made a transaction of this kind difficult. On October 29, 2025, the company announced it would abandon its French corporate domicile for Luxembourg, on the specific ground that French law provides no framework for a direct merger into a United States corporation, a rationale spelled out in Criteo's own investor documentation. Shareholders endorsed the move at an extraordinary general meeting on February 27, 2026, casting 50,511,371 votes in favor against 114,993 opposed, with 37,908 abstentions, a landslide PPC Land reported at the time. The conversion is expected to complete in the third quarter of 2026. The Vista and Quinti Capital offer arrived, then, inside the exact window Criteo itself identified as the point at which a transaction with an American acquirer would become structurally feasible. Coincidence is possible. It is not the most parsimonious explanation.

What the buyers would actually acquire

Any acquirer takes possession of a company in mid-transition rather than a settled asset. Criteo, founded in 2005 and trading under the ticker CRTO, operates two segments: Retail Media, which sells sponsored product placements and display formats on retailer-owned digital properties for approximately 225 retailers, and Performance Media, the older retargeting business built on the open web. More than 4,100 brands transact across the platform. Michael Komasinski, who joined from dentsu, has run the company as chief executive since February 15, 2025, with Sarah Glickman as chief financial officer. Earlier acquisitions, including HookLogic, Storetail, Mabaya, Iponweb and Brandcrush, extended the business from retargeting into marketplace monetization and programmatic infrastructure, a build-up MediaPost traced in its July 7 account.

The newer, more speculative asset is Criteo's position in what the industry has taken to calling agentic commerce. The company introduced its Agentic Commerce Recommendation Service on February 5, 2026, reporting up to 60 percent improvement in recommendation relevancy over approaches that rely on product descriptions alone, based on internal testing conducted in January 2026, as PPC Land detailed at launch. On March 2, 2026, Criteo became the first advertising technology partner in OpenAI's ChatGPT advertising pilot, connecting its roughly 17,000 advertiser clients to ChatGPT's Free and Go subscription tiers in the United States. By early May, more than 1,000 brands were running active campaigns through that integration, with AI-referred conversion rates approaching twice the rate of traditional search referrals in some retail categories, per PPC Land's May 2026 reporting. MediaPost's July 7 piece put the current figure at about 2,000 brands advertising on ChatGPT through Criteo. A financial buyer weighing the offer is, in effect, pricing the option value of that AI positioning against the near-term revenue erosion from the retail media client losses, with a 40 million euro GDPR fine, upheld by France's Conseil d'Etat on March 4, 2026, sitting in the liability column as a reminder of the regulatory exposure attached to cross-site tracking businesses in Europe.

Vista's own track record in advertising technology frames the approach. The Texas-based firm manages more than 90 portfolio companies serving over 450 million users, and it exited Integral Ad Science in September 2025 through a 1.9 billion dollar all-cash sale to Novacap. Vista executives Rod Aliabadi and Eric Roza have held board and operating roles at TripleLift, another ad tech holding. Quinti Capital remains the lower-profile half of the pairing, with little public detail about its participation. The broader consolidation wave gives the bid its context: Publicis Groupe agreed on May 17, 2026 to acquire LiveRamp at a 2.2 billion dollar enterprise value and a 29.8 percent premium, Novacap absorbed IAS, and Broadsign bought Place Exchange in out-of-home, a sequence PPC Land's July 8 industry analysis reads as private capital and strategic buyers systematically absorbing mid-sized, publicly listed ad tech platforms as the category matures. Industry estimates cited in Criteo's own disclosures project retail media capturing 20 percent of global advertising revenue by 2030, roughly 300 billion dollars. Who owns the pipes when that happens is the question the bid forces.

IAS installs an AI-era chief executive

The Vista thread connects directly to the week's most consequential personnel move. Integral Ad Science, the media quality and verification company Novacap acquired from Vista last year, announced on July 7, 2026 that Lidiane Jones would become chief executive officer, effective immediately, succeeding Lisa Utzschneider after a run of more than seven years, as reported by PPC Land on July 8. Utzschneider remains as Special Advisor to the Board through the end of 2026 and will separately advise Novacap and its portfolio companies, details laid out in the announcement carried on Adweek's wire. Jones arrives with an unusual resume for an ad verification firm: chief executive of Bumble, chief executive of Slack following its acquisition by Salesforce, where she served as executive vice president and general manager of digital experiences, earlier leadership roles at Sonos, and more than a decade at Microsoft.

The scale of the handover is easy to understate. Utzschneider joined IAS in 2019, steered the company through its 2021 initial public offering, and remained in the chair through the September 2025 sale that returned the business to private ownership. Her seven-plus years span the entire arc of ad verification's maturation, from viewability audits and brand safety block lists through pre-bid optimization to the AI-driven classification products that now generate much of the category's growth. Handing the company to an executive whose formative experience lies in consumer software, workplace collaboration and platform product rather than in advertising suggests the board views the next chapter as a product-engineering problem more than a sales one. It also mirrors a wider pattern in ad tech leadership, where boards increasingly recruit from outside the industry when the strategic question shifts from selling into the existing ecosystem to anticipating how AI rearranges it.

Two days into the role, Jones sat down with Adweek and signalled where the company is headed. The July 8 interviewdescribes her framing IAS as the "trust layer" for an industry being reshaped by AI assistants, agentic bidding and chatbot-based discovery, with an openness to mergers and acquisitions as part of the plan. Her stated priority for the coming months centers on meeting customers and accelerating product work around AI-mediated media buying and selling. The appointment reads as a bet by Novacap that measurement and verification, categories built for a world of human-viewed web pages and video players, must be rebuilt for one in which software agents increasingly select, transact and even consume advertising. Whether a verification vendor can become infrastructure for machine-to-machine media trading is an open question. That the question now sits at the top of the IAS agenda, one year into private ownership, says something about how quickly the ground has moved.

The open questions a board must now answer

Several elements of the Criteo situation remain outside the public record, and their resolution will determine whether the approach becomes a transaction or another entry in the company's long history of near-sales. It is not known whether Criteo has retained financial or legal advisors to evaluate the offer, whether an exclusivity or go-shop period has been discussed, or whether other bidders have surfaced. The structure of the bid, cash versus debt-financed, whole company versus a carve-out of one segment, has not been disclosed. Nor is it clear whether the Luxembourg conversion, expected to complete in the third quarter, had been finalized as a matter of corporate law at the moment the offer arrived or remains in process, a technicality with real consequences for how quickly any merger could execute. The history counsels patience. Talks with Skai went nowhere last year, and earlier exploratory processes under previous management produced headlines rather than signatures. A 21.4 percent single-day repricing nonetheless tells its own story about how credible the market judges this particular approach.

For the retailers, brands and agencies that run on Criteo's infrastructure, ownership matters in practical rather than abstract terms. The platform functions as commerce plumbing for roughly 225 retail media networks, which means a change in strategic direction, product investment or pricing after an acquisition would propagate through the retail media ecosystem rather than staying contained within one vendor relationship. Private equity ownership historically brings sharper focus on cash flow and margin, which could accelerate the managed-service-to-platform transition Criteo has been navigating on its own terms since 2025, and it typically removes the quarterly disclosure rhythm through which clients and competitors currently track the company's health. The last comparable case offers a template: IAS spent a decade under Vista, went public in 2021, and returned to private hands in 2025 when the public market's patience ran out. Criteo, public since 2013, may be tracing the same arc in reverse order.

Taken together, the Criteo approach and the IAS transition describe an industry whose ownership structures are being renegotiated faster than its operating models. A French retargeting pioneer engineered its own legal availability for an American buyout. A verification firm passed from one private equity owner to another and then reached outside advertising entirely for a product-and-AI executive. The public markets, which spent a decade pricing ad tech companies with visible skepticism, are watching private capital arrive at rather different valuations, one bid at a time.

Europe's 131.1 billion euro ad market and the formats pulling it forward

While bankers priced Criteo, economists priced the continent. IAB Europe released the AdEx Benchmark 2025 Report on July 7, 2026, at a dedicated event in London, and the headline figure landed at 131.1 billion euros, a 10.5 percent increase that added 12.5 billion euros to Europe's digital advertising market in a single year, as PPC Land reported on July 8. The twentieth edition of the study, compiled with national advertising associations across 30 markets, found that every single market tracked posted growth during 2025. That breadth is unusual. So is the deceleration hiding inside it: the market grew 16.0 percent in 2024, from 102.2 billion euros in 2023 to 118.6 billion euros, before slowing to the current pace, which the report situates close to the 10 to 13 percent long-run average recorded across most years between 2012 and 2019.

The report's most quoted line came from its author. Dr. Daniel Knapp, IAB Europe's chief economist, argued that the year's performance, achieved against sluggish GDP growth, trade policy uncertainty and cautious consumers, shows digital advertising "decoupling from the macro cycle." Businesses, in his assessment, now fund digital advertising the way they fund distribution rather than the way they historically funded media, treating it as sales infrastructure, shelf space and shopfront simultaneously. Digiday's July 8 analysis by Seb Joseph added a memorable scale comparison from Knapp's July 7 panel appearance: continental ad spend now roughly matches the size of Morocco's entire economy, with digital soaking up around 70 percent of the total. Europe remains barely half the size of the American market in absolute terms, yet the direction of travel is unambiguous, and the money keeps moving toward whichever channel promises the most measurable attention.

Video crosses the halfway line

The structural milestone of the 2025 data belongs to video. Video advertising grew 19.6 percent to reach 34.0 billion euros, and for the first time in the study's twenty-year history it accounts for more than half of all display investment across Europe. The threshold had been approaching for years, and the composition of the growth explains why it finally arrived. Subscription video-on-demand advertising, the category spanning Netflix, Disney Plus and Amazon Prime Video, expanded 59.6 percent in 2025. That figure only looks restrained beside the 222.4 percent the same category posted in 2024, when it was growing from a much smaller base. Broadcast video-on-demand grew 14.5 percent, down from 29.5 percent a year earlier, while the free ad-supported category that includes YouTube grew 9.6 percent, decelerating from 16.9 percent.

Ten European markets now see video account for more than half of their total non-social display advertising. Slovenia leads at 89.4 percent, followed by Turkey at 72.7 percent and Ukraine at 68.3 percent. Across the continent, video's share of the non-social display market reached 48.2 percent, just below the crossover point, though the combined view of display and social video splits the market evenly at 50 percent each. Programmatic buying absorbed a growing share of the shift: programmatic display and video spending outside social rose 10.9 percent to 15.771 billion euros, against non-programmatic growth of just 1.5 percent. Inside that programmatic total, video spending grew 19.8 percent to 9.832 billion euros while programmatic display outside video fell 1.2 percent to 5.939 billion euros. Budget is not merely migrating into programmatic pipes; within those pipes, it is abandoning static formats for moving ones.

Social advertising, measured separately in the report's taxonomy, grew 19.2 percent to 35.5 billion euros, and social video was the fastest-growing individual format anywhere in the study, up 25.7 percent to 18.602 billion euros. Social video now represents 52 percent of total social ad revenue, with standard non-video social formats growing a more modest 12.9 percent to 16.908 billion euros. The digital outperformance gap against traditional media widened as a result: digital media grew 17.0 percentage points faster than non-digital media in 2025, up from a 13.7 point gap in 2024.

Retail media passes a tenth of the market, but where does the money come from?

Retail media supplied the second milestone. The category grew 16.7 percent in 2025 to reach 13.314 billion euros, crossing 10 percent of total European digital ad spend for the first time since IAB Europe began tracking it. The composition tilts heavily toward search: retail search reached 10.002 billion euros after growing 19.1 percent, close to three times the pace of standard search advertising, while retail display grew 9.8 percent to 3.312 billion euros. Combined search advertising of both kinds rose 8.8 percent to 56.155 billion euros, with standard search growing 6.8 percent to 46.153 billion euros. Retail search's share of the total search market climbed to 17.8 percent from 16.3 percent a year earlier. The measurement scope matters here, because the AdEx figure excludes in-store retail media, off-site retail media and non-retail revenue streams such as streaming ad revenue earned by retailers, meaning the true scale of commerce advertising in Europe runs higher than the benchmark alone suggests. An earlier IAB Europe study using a broader scope put 2024 retail media at 13.7 billion euros with 21.1 percent growth.

Digiday's reading of the retail media numbers added an important caveat that the celebratory framing tends to skip. Most of the category's money, the publication noted, is not new advertising investment at all but old trade and shopper marketing budgets being reclassified as media, a shift in accounting as much as a shift in demand. The distinction carries planning consequences. A category growing by conquest of adjacent budgets behaves differently from one growing on fresh demand, particularly when retailers set the terms of measurement on their own properties.

Not every format shared in the growth. Non-video display spending, the banners, native placements, newsletter ads and affiliate formats that once defined digital advertising, actually contracted 0.8 percent to 15.461 billion euros, standing out against nearly every other line in the report. Digital audio grew 13.9 percent to 1.230 billion euros, split between podcast advertising at 539 million euros, up 16.7 percent, and streaming and internet radio at 691 million euros, up 11.8 percent. Classifieds and directories grew 6.5 percent, the slowest of the five headline formats, with a spread of national outcomes wide enough to signal structural change: Turkey posted 67.1 percent growth in the category while Finland contracted 10.0 percent and Denmark 5.9 percent.

Geography, inflation and the limits of the data

The growth map skews decisively east. All ten of the fastest-growing markets sit in Central, Eastern or South-Eastern Europe, a bloc that expanded by more than 20 percent in aggregate, roughly twice the pace of the mature Western markets. Turkey led nominal growth at 37.0 percent, Ukraine followed at 22.0 percent, and Serbia, Croatia, Poland, Romania, Bulgaria, Slovenia, Hungary and the Czech Republic completed the list. Sixteen of the thirty markets grew at double-digit nominal rates. Value, however, remains concentrated where it has always been. The United Kingdom, Germany and France together account for 62 percent of total European digital ad spend, with the UK alone at 46.933 billion euros, more than double Germany's 21.583 billion euros and well clear of France's 12.701 billion euros. The top five markets hold 70 percent of spend, the top ten hold 88 percent, and the top fifteen hold 95 percent, leaving fifteen countries to divide the remaining twentieth among themselves.

For the first time in the report's history, IAB Europe published an inflation-adjusted view of growth, built on World Bank consumer price data cross-checked against the International Monetary Fund's World Economic Outlook database. The methodology classifies Turkey as a hyperinflationary Tier 1 market, with cumulative three-year inflation of 229.1 percent, and Ukraine as a high-inflation Tier 2 market at 35.4 percent cumulative. Applying the adjustment compresses Turkey's headline 37.0 percent growth to 1.6 percent and Ukraine's 22.0 percent to 8.2 percent. It also shaves the continental figure: excluding Turkey, European growth stood at 9.7 percent, and the fully inflation-adjusted rate comes to 9.4 percent, meaning roughly 1.1 percentage points of the headline 10.5 percent reflects price increases in high-inflation markets rather than expanding advertiser demand. The number of genuinely double-digit markets narrows from sixteen to fourteen under the adjusted view.

The report is candid about its own blind spots, and the candor deserves notice. Five of the thirty markets, Belgium, Bulgaria, Greece, Portugal and Romania, have no national association submission and are fully modelled from regional growth rates, company filings and historical trajectories. A separate disclosure flags six markets, Belgium, Romania, Turkey, Slovakia, Serbia and Croatia, where national measurement likely misses the self-serve long tail of search and social spending because it captures only association members and agency-transacted budgets. Slovakia's case is explicit: its national source covers only the country's largest advertisers, with small-business spending excluded at IAB Slovakia's own request. Where a market's search share falls below half the European average of 42.8 percent without a structural explanation, the report flags it for coverage review rather than inflating the figure. Fourteen markets recorded double-digit search growth, led by Croatia at 64.2 percent and Turkey at 40.0 percent, both figures shaped more by currency and inflation dynamics than by comparable underlying demand.

The transatlantic comparison, presented at this level of detail for the first time, cuts against European self-congratulation. The United States digital advertising market reached 260.5 billion euros in 2025, converted at a constant rate of 1.131 dollars per euro, and grew 13.9 percent. America grew faster than Europe in percentage terms and from a base twice the size, widening rather than narrowing the absolute gap. Artificial intelligence, meanwhile, received deliberately measured treatment. Knapp confined AI's 2025 effect to the supply side, cheaper creative production, better optimization, more automation, with productivity gains flowing disproportionately to smaller advertisers, and deferred the second-order effects on search behaviour, agentic buying and budget allocation to next year's edition. Townsend Feehan, IAB Europe's chief executive, paired the growth story with a call for the innovation to be matched by trust, transparency and accountability.

Why does a single benchmark study command this much attention? Since 2006, the AdEx series has functioned as the most cited harmonised source for aggregate European digital ad spend, assembled from national association submissions supplemented by public filings and third-party data where local figures are unavailable. Budget planners benchmark national performance against the continental trend line, agencies lean on it to support channel-shift recommendations, and platform vendors contextualise their own growth claims against an independent baseline. The figures also feed the national reporting cycle in both directions: IAB member associations publish country studies throughout the year, several of which diverge from the harmonised European numbers because of differing scope, category boundaries and net-versus-gross conventions, gaps the AdEx report itself acknowledges. Poland's national study, showing an 11 billion zloty domestic market, and Ukraine's projection of 18 percent growth for 2026 led by video and connected TV, both published in June 2026, illustrate the granularity sitting beneath the pan-European totals. Last year's edition documented the market's first crossing of the 100 billion euro threshold, a milestone PPC Land covered in mid-2025; the 2024 baseline has since been restated from 118.9 billion to 118.6 billion euros under the current methodology, a reminder that even the anchor figures move as the measurement improves. No date has been announced for the next edition, though the publishing cadence points to a mid-2027 release covering full-year 2026, the edition Knapp has already flagged as the one where AI's effects on search behaviour, agentic buying and budget allocation will finally register in the numbers.

One dissenting voice from the same London stage merits the final word on the report. Ian Whittaker, the former City equity analyst who now runs Liberty Sky Advisors, told the July 7 forecast panel that "advertising has lost its way," a blunt assessment Digiday relayed on July 8. His argument holds that the straight line that once ran from a business's commercial needs to what its agency did on its behalf has dissolved, leaving an industry that talks to itself, debating agentic transaction protocols that brand-side finance teams never asked for, while telling investors a story about AI-driven margin expansion. Against a 131.1 billion euro market growing through macroeconomic gloom, the critique may sound churlish. It may also be the most useful sentence spoken at the event.

Apple loses in Luxembourg, and the DMA's architecture hardens

Apple's three-year effort to loosen its Digital Markets Act obligations ended, at least at first instance, on the morning of July 8, 2026. The General Court of the European Union, sitting in its Eighth Chamber with five judges in Luxembourg, dismissed three joined cases, T-1079/23, T-1080/23 and T-214/24, in which the company had challenged its designation as a gatekeeper for the App Store and iOS along with the Commission decisions opening and closing a market investigation into its iMessage service, as PPC Land reported the same day. The accompanying press release from the Court of Justice stated that the court "dismisses all the actions brought by Apple," confirming the gatekeeper designation and finding the iMessage challenges inadmissible. Apple Inc. and Apple Distribution International were ordered to pay the Commission's costs across all three cases, along with those of the Coalition for App Fairness, which intervened in support of the regulator. One narrow route remains: an appeal to the Court of Justice, limited to points of law, within two months and ten days of formal notification.

The procedural history explains what was actually at stake. Apple notified the Commission in July 2023 that four of its services met the DMA's quantitative thresholds: the App Store, iOS, Safari and iMessage. For iMessage, the company argued the service should not count as a number-independent interpersonal communications service at all, and that exceptional circumstances should exempt it regardless. The Commission split the difference on September 5, 2023, designating Apple as a gatekeeper for the App Store, iOS and Safari while opening a five-month investigation into iMessage. That investigation closed on February 12, 2024 with a decision not to designate the messaging service, though the closing decision, like the original one, continued to classify iMessage as an NIICS in its reasoning. Apple's obligations took legal effect on March 7, 2024, and the company filed three separate actions attacking different pieces of the regulatory architecture. A hearing followed on October 21, 2025. The judgment delivered this week resolves all three.

Three defeats, each with different consequences

The first defeat concerns interoperability. Apple sought to challenge Article 6(7) of the DMA, the provision requiring operating system gatekeepers to open interoperability to hardware and software providers free of charge on terms equivalent to those Apple's own products enjoy, arguing the rule disproportionately infringes the right to property under the EU Charter of Fundamental Rights. To attack a general rule through an individual decision, EU procedure requires what is called a plea of illegality under Article 277 of the Treaty on the Functioning of the European Union, and the rule challenged must form the legal basis of the decision under attack. The court found it does not. Article 3 of the DMA, which governs designation, is the legal basis of the designation decision; Article 6(7) defines obligations that attach only after designation has occurred. Designation and obligation are two structurally separate stages, and separating them for legal purposes reflects how the regulation is built rather than an artificial construction. The court also rejected Apple's argument that this sequencing denies effective judicial protection, noting that if the Commission later adopts an implementing decision under Article 8 specifying how Apple must comply, the company can challenge that decision and raise the illegality plea then. The path exists. It simply has not opened.

The second defeat carries the most immediate commercial weight. Apple argued that its five App Store storefronts, spanning iPhone, iPad, Apple Watch, Mac and Apple TV, should be assessed as five distinct services, in which case only the iOS storefront would clear the DMA's user and revenue thresholds, leaving the other four outside gatekeeper scope. The court sided with the Commission's view that all five stores perform the same function regardless of device: connecting business users offering applications with end users who download them. Under Article 2(14) of the DMA, a software application store is defined by the transaction it enables, not the hardware it runs on, and the court drew on recital 14 of the regulation to hold that the definition of core platform services must remain technology neutral precisely so that a gatekeeper's own technical architecture cannot determine what counts as one regulated service or several. Usage differences between the stores, how often each opens, what gets downloaded where, go to the character of the device experience, not the purpose of the store. For developers building cross-device distribution strategies and for measurement vendors whose tooling depends on how the ecosystem is regulated device by device, the ruling forecloses any argument that the smaller storefronts operate under a lighter regime.

The third strand, on iMessage, produced the ruling's most structurally significant principle while changing nothing about Apple's present obligations. All three iMessage challenges were declared inadmissible, not because the NIICS classification was right or wrong on the merits, but because a classification confined to a decision's recitals, without a listing in the operative part as an important gateway, produces no binding legal effect. No DMA obligation applies to iMessage today. None can, unless a future designation places the service in an operative part, and Article 4(2) requires the Commission to conduct a fresh assessment whenever it revisits the question rather than inheriting its earlier reasoning. National telecoms regulators enforcing the separate European Electronic Communications Code remain free to reach their own conclusions, since Article 1(4) preserves their powers without binding them to the Commission's classification. The practical lesson generalizes beyond Apple: contesting a Commission classification before it bites is, per this judgment, not an available legal avenue. Companies must wait for an actual obligation before seeking review of it. A month earlier the same court applied a consistent logic to Meta, annulling Facebook Marketplace's gatekeeper designation while upholding Messenger's classification, and the two rulings together sketch a stable line around what counts as a challengeable designation versus a preliminary classification without teeth.

The cast list around the litigation illustrates how much of the industry treated the outcome as its own business. The French Republic, the Federal Republic of Germany, the European Parliament, the Council of the European Union and the Free Software Foundation Europe all intervened at various points, each ordered to bear its own costs, while the Coalition for App Fairness, the developer group whose members include some of Apple's most vocal commercial adversaries, intervened in support of the Commission and will have its costs paid by Apple. One doctrinal question survives the judgment untouched: whether a decision to open a market investigation is, in general, a preparatory act incapable of independent challenge. The court dismissed the actions against the investigation-opening decision on the narrower ground that the classification produced no legal effect, and therefore never needed to reach the preparatory-act argument the Commission had also raised. Future litigants will test that gap.

A widening European enforcement front

The Apple judgment did not arrive in isolation. Six days earlier, on July 2, 2026, the same court dismissed Google's appeal against its 4.1 billion euro Android fine, ending eight years of litigation over Android search defaults and leaving Alphabet jointly liable for 1.52 billion euros of the total. Italy's competition authority opened a national investigation on June 9, 2026 into whether Apple withholds full-device backup interfaces from rival cloud providers on iPhone and iPad, a probe with a final ruling due by March 2027. The Commission opened DMA market investigations into Amazon Web Services and Microsoft Azure in November 2025. Each proceeding will eventually navigate the same designation-first, obligation-second, review-last sequence the Apple ruling has now confirmed.

France added its own pressure point on July 8. An interim order requires Meta to hand publishers its remuneration data within 15 days and to negotiate with press organizations that have gone unpaid for the use of their content since January 2025, an obligation that stands until the country's competition authority issues its final ruling, as PPC Land reported. The neighboring-rights fight has run for years in France, but the 15-day clock converts a slow structural dispute into an immediate compliance deadline for one of the two largest advertising platforms in Europe.

What has all this enforcement actually purchased? A study published July 8 offers a sobering interim answer. A University of Antwerp analysis found the DMA has shifted roughly six million EU users to Firefox while leaving Google with approximately 90 percent of the search market, and concluded that the main beneficiaries of the regulation's contestability provisions are Bing, Edge and Mozilla, American firms all, rather than any European challenger, findings PPC Land covered on July 8. The research raises a question Brussels will find uncomfortable: whether a regulation designed to open markets is redistributing share among American incumbents rather than cultivating European alternatives. For advertisers, the practical reading is simpler. Choice screens, browser switching and gatekeeper obligations have not yet moved the demand curves that determine where budgets go. The App Store rules that survived this week's judgment, alternative marketplaces, alternative payments, browser choice, changes Apple implemented for iOS to comply with the regulation, continue to reshape distribution economics on iOS. The search market, so far, absorbs everything thrown at it.

Google measures the platforms as publishers question the traffic

The search story of the week is really two stories moving in opposite directions. Google expanded what its measurement tooling can see, and publishers escalated their doubts about whether what it sends back is worth having. Both threads ran through the same 48 hours.

On Tuesday, July 7, 2026, Google introduced platform properties, a new Search Console property type that lets creators and site owners track how their Instagram, TikTok, X and YouTube posts perform inside Google Search and Discover, as PPC Land detailed on July 8. The announcement came through the Google Search Central Blog under Moshe Samet, Product Manager Lead for Search Console, with a companion LinkedIn post the same day. The structural change is larger than the feature list suggests. Search Console has served website owners since its inception; verification always meant proving control of a domain or URL prefix. Platform properties sever that dependency entirely. A creator with an Instagram or TikTok presence and no website at all can now verify an account through an authorization flow and receive dedicated reporting on which Google search queries drive impressions and clicks to that content.

Each platform property carries three reports. The Performance report shows total clicks, impressions and related metrics, filterable down to individual posts and queries, with export available for external analysis. The Insights report summarizes recent traffic trends, top posts and discovery patterns. Achievements marks milestones such as crossing click thresholds within a 28-day window, the default range applied across the product. The counting rules deserve attention from anyone who plans to reconcile these figures against native platform analytics. When an Instagram story appears in Google results, that counts as an impression; a click registers even when the content opens inside Google's own viewer without the user ever leaving Google's interface. The reports measure discovery through Google exclusively. A TikTok video that goes viral entirely inside TikTok's own recommendation feed shows nothing here, however many millions of native views it accumulates. Rollout proceeds gradually over the coming weeks, each account or channel must be verified as its own property, ownership is re-checked periodically, and a lapsed connection pauses reporting until re-verification, after which historical data returns immediately.

The reporting carries reconciliation traps that agencies will encounter within the first week of use. The summary card at the top of the Insights page reflects all clicks across every Google surface, web, image, video and news search combined, while the detailed lists beneath it cover web search results specifically, which means individual cards can show totals lower than the headline figure above them, a discrepancy Google's documentation flags rather than leaving teams to discover on their own. Discover and Google News sections appear only for properties that actually receive traffic from those surfaces; a creator whose content has never surfaced in Discover simply sees no Discover report rather than an empty one. Newly created properties display empty charts for a few days while collection begins, and charts populate only from the connection date forward, with no backfill of activity predating verification. Operationally, the launch ships without a bulk-import mechanism, without a cross-platform aggregated dashboard, and without any API access point distinct from what exists for standard website properties, so an agency managing Instagram, TikTok and YouTube accounts for a single brand must verify and maintain three separate properties. Whether API support follows remains unaddressed in the launch materials.

Search Engine Roundtable's Barry Schwartz reported the launch on July 7, noting that Google had briefly published, then pulled, a help document named platform properties several weeks earlier, a premature disclosure that now has its explanation. Schwartz also flagged the genuinely novel element: Search Console now verifies content on domains the account holder neither owns nor develops. The feature grew out of a December 8, 2025 experiment that layered automatically detected social channel data into the Insights report for a limited set of websites, an approach that still required an existing verified site. That requirement is gone. The launch extends a year of steady Search Console expansion, from the June 30, 2025 integration of Insights into the main interface, through AI-powered query grouping on October 27, 2025, to custom annotations on November 17, 2025.

Why measurement is expanding while the traffic shrinks

The timing of the launch invites a less generous reading, and the surrounding week supplied it. Chartbeat data cited across PPC Land's publishing coverage shows small publishers losing 60 percent of their search referral traffic over two years, with page views from Google Search falling 34 percent between December 2024 and December 2025 alone. A randomized experiment published July 4 produced the first causal estimate of AI Overviews cutting outbound publisher clicks by 39.8 percent while lifting zero-click searches 34.5 percent. Separate research from Ahrefs associates AI Overviews with a 58 percent reduction in click-through rates for top-ranking pages. Against those numbers, a tool that lets creators watch their social posts surface in Google reads less like generosity and more like an effort to keep the creator economy inside Google's measurement layer as discovery fragments across platforms Google does not own.

The four supported platforms are not incidental choices, and neither is the surface pairing. Instagram, TikTok, X and YouTube collectively represent the destinations where audiences, particularly younger ones, increasingly begin product research and content discovery rather than starting from a search bar. Google's own Discover feed moved in the same direction months earlier, integrating content from X, Instagram and YouTube Shorts alongside a creator follow feature from September 17, 2025. Platform properties, in effect, hands the creators behind that integrated content a measurement layer for the distribution Google was already performing. The strategic logic runs both ways: every query answered inside TikTok's search box is a query Google never sees, and a reporting product that makes Google-driven discovery legible to creators gives those creators a reason to keep optimizing for Google's surfaces rather than treating the platforms as closed loops.

Publishers, meanwhile, are testing the limits of refusal. Cloudflare introduced Content Signals for robots.txt in early July, a mechanism that from September 15 will block bots designated for large language model training or agentic use by default while continuing to allow search crawlers. The catch emerged within days: neither Google nor any AI chatbot honors the classifications. Google search advocate John Mueller wrote on Reddit that the directives have "no effects whatsoever for any crawler or LLM," a statement Search Engine Roundtable reported on July 6. The asymmetry is structural rather than incidental. Google crawls with a single bot for both search indexing and AI-generated answers, which means a publisher cannot block AI Overviews or Gemini without simultaneously sabotaging its own search visibility. AdExchanger's July 8 roundup connected that bind to its logical endpoint: the chief executive of USA Today's parent, Mike Reed, told the Rebooting newsletter that the traditional Google search business model no longer functions for publishers, and that the company is weighing blocking Google's crawler entirely in favor of newsletters and social distribution. When the largest newspaper chain in the United States discusses de-indexing itself from the largest search engine, the relationship between the two industries has moved into new territory.

The advertising layer follows the queries

Advertising infrastructure is chasing the same shift in behavior, with uneven results. OpenAI began rolling out an Audiences targeting option inside ChatGPT Ads, letting advertisers upload raw or hashed emails and phone numbers as campaign filters, a change Search Engine Roundtable documented on July 7 after specialist Craig Graham surfaced screenshots on LinkedIn. Customer-list targeting is the connective tissue of mature ad platforms, the mechanism behind Google's Customer Match and Meta's Custom Audiences, and its arrival in ChatGPT Ads signals a platform building for direct-response budgets rather than experimental brand spend.

The demand side, at least in some markets, has yet to show up. Adthena logged zero paid placements across 169,560 ChatGPT result scrapes in the United Kingdom during June, even as Google's AI Overviews appeared on 23 percent of tracked UK queries, figures PPC Land reported on July 8. The gap between infrastructure and inventory is instructive. Targeting tools, measurement pixels and format experiments are arriving faster than actual ad load, which suggests OpenAI is assembling the machinery of an advertising business ahead of switching on supply at scale, a sequencing choice with obvious echoes of earlier platform build-outs. The two data points in the Adthena study also invert the usual competitive framing. Google's generative surface already touches nearly a quarter of tracked UK queries, giving marketers an unavoidable organic-visibility problem, while OpenAI's surface offers advertisers no purchasable placements at all in that market, giving them nothing to buy even where budgets are willing. British advertisers, in other words, currently face maximum disruption from AI answers and minimum access to AI ad inventory simultaneously, a combination that helps explain why measurement vendors are finding an eager audience for scrape-based visibility data in the interim.

Google's own auction mechanics shifted inside the same window, with consequences arriving on a fixed calendar. The Bid Target Adjustment Tool went live on July 6, opening a 42-day period before campaigns constrained by budget begin bidding toward their stated CPA and ROAS targets rather than the more efficient levels automation historically delivered, a change PPC Land explained on July 7. A campaign paying 5 dollars per acquisition against a stated 10 dollar target will drift toward the stated figure from August 17. The alteration rewards advertisers who kept their declared targets honest and penalizes those who treated the target field as a bargaining position, and the six-week runway that opened this week is the only adjustment period on offer. Search, in other words, is changing at every layer at once: what gets measured, what gets crawled, what gets answered without a click, and what each click costs.

Europe rewrites the anonymity test that ad tech was built on

The least noisy story of the window may prove the most durable. The European Data Protection Board opened a public consultation on July 8, 2026 on Guidelines 02/2026 on Anonymisation, a 33-page framework adopted the previous day that replaces the three-criteria test the Article 29 Working Party established in 2014 and adds a fourth element aimed squarely at inference-based re-identification, as PPC Land reported on July 8. Comments are due by 23:59 CET on October 30, 2026. Twelve years of case law, EU-wide data spaces and advances in artificial intelligence, the board states, made the 2014 criteria obsolete.

The commercial stakes explain why every identity vendor, clean room operator and retail media network in Europe should read all 33 pages. Anonymised data falls entirely outside the scope of the General Data Protection Regulation. Data that qualifies escapes consent requirements, purpose limitations, transfer restrictions and the rest of the regulation's machinery, which creates a powerful incentive across advertising technology to demonstrate that a given dataset clears the bar. The penalty for getting the assessment wrong runs in the other direction with equal force: a controller whose anonymisation fails becomes retroactively accountable for processing personal data without a lawful basis for the entire period the data was wrongly treated as anonymous.

A two-question test, viewed from the recipient's chair

The updated framework rests on two questions. Does the information relate to a natural person, and if so, is that person identified or identifiable? A negative answer to either question renders the data anonymous. The complication, which the guidelines spend considerable space unpacking, is that the answers can differ depending on who holds the data. Information may plainly relate to an individual while only one organisation possesses the means to identify that person, in which case the data is personal for that organisation and anonymous for everyone else. This relativity principle comes directly from the Court of Justice's September 4, 2025 judgment in EDPS v Single Resolution Board, case C-413/23 P, which held that whether information constitutes personal data must be assessed from the recipient's perspective, a ruling PPC Land covered when it was delivered. The new guidelines cite that judgment more than a dozen times.

Two assessment methods are on offer, and the choice between them is itself a risk decision. The contextual approach weighs the actual capabilities of each entity that might attempt re-identification, reflecting the full nuance of the legal standard and permitting the conclusion that data is anonymous for some parties while remaining personal for others. The simplified approach ignores those differences and asks only whether re-identification is theoretically possible by anyone, an approach that can exceed the legal standard and lead a controller to treat data as personal even where it would qualify as anonymous for specific recipients. What the simplified method sacrifices in precision, it returns in caution. The guidelines suggest a combination will often prove most practical: run the simplified test first, and escalate to contextual analysis only where a theoretical vulnerability appears.

Record isolation, linkage, and the new inference criterion

The technical core consists of three criteria, all of which a dataset must satisfy for the guidelines to presume it anonymous. No Record Isolation asks whether the data contains a unique combination of attribute values tied to a single individual; the guidelines illustrate with five patient records combining sex, date of birth, postcode and diagnosis, every combination of which proves unique, failing the criterion outright. The larger a record and the more attributes it carries, the higher the likelihood of uniqueness. No Linkage examines whether a record could be matched with certainty or high likelihood to a record about the same person in a separate dataset; the worked example involves a board game shop whose de-identified purchase records could still be matched against a public website where customers log the games they own, failing the criterion even though no name, email or phone number appears anywhere in the retailer's file.

No Inference is the addition that reaches furthest into advertising practice. The criterion fails where a relevant entity could draw a specific and meaningful conclusion about an individual from the data, without needing to isolate a record or link it externally. The guidelines draw a careful boundary between inferences that are merely specific, relating to one identified person, and those that are also meaningful, meaning they depend on the underlying dataset rather than general population knowledge. Anonymised bank loan data that supports a default-risk prediction about a new applicant does not violate the criterion when the prediction reflects population-level correlations that would hold regardless of whether that applicant's data ever entered the file. The AI-specific passages sharpen the point further. Membership inference, the technique of determining that a specific person's data was included in a model's training set, can itself violate the criterion without extracting any other detail about the person. De-aggregation attacks capable of pulling data out of models presumed anonymous receive their own warning, and the guidelines note that agentic AI is likely to reduce both the time and the cost of increasingly sophisticated re-identification techniques.

Failure of a criterion does not automatically doom a dataset. The framework directs controllers toward further analysis rather than an immediate finding of personal data. A cinema survey with thirty free-text questions per respondent produces nearly unique response sets across the board, violating Record Isolation, yet nothing in the survey ties any answer set to an identifiable person, so the dataset can still pass overall. The compiled results table spanning all three criteria and both assessment approaches formalizes this escalation logic.

The presumption that people follow the law, and when it breaks

One clarification will reshape how legal teams draft data-sharing agreements. Prior case law held that re-identification means need not be considered reasonably likely where they are prohibited by law and the prohibition is genuinely effective. The guidelines refuse to treat that as an automatic exemption. The general assumption of legal compliance can be rebutted with evidence of a concrete risk that unlawful means will be used anyway: a prohibition that goes unmonitored and unenforced, potential gains that outweigh the costs and risks of breach, or a history of comparable prohibitions being circumvented in similar situations. A worked example features a media company relying on a legal ban on third-party server access, only to find that its out-of-date network security renders the ban meaningless as a practical barrier. The guidelines separately warn against leaning on an entity's presumed lack of motivation, which is hard to demonstrate objectively and may not survive contact with changed circumstances; identification can happen through accident or negligence as easily as intent.

The roster of entities a controller must consider extends well past its contractual counterparties. Receiving entities and rogue employees appear alongside investigative journalists, particularly for data about people in or approaching public life, domestic and foreign law enforcement and intelligence agencies, companies willing to exploit legal grey areas, and cybercriminals who disregard the law entirely. Where anonymisation succeeds for a recipient but not for the party performing it, the anonymising controller must continue treating the data as personal for its own purposes, GDPR obligations intact, even as the identical dataset counts as anonymous in the recipient's hands.

What this means for fingerprints, clean rooms and lookalikes

The guidelines speak to advertising with unusual directness. A website that identifies visitors through a unique combination of browser, operating system, screen resolution and time zone, then tracks them page to page, is using an identifier in the same legal sense as a name or ID number. The example escalates: several websites sharing fingerprint-derived pseudonyms with a common advertising provider leave the individuals identified or identifiable from that provider's perspective, because the third party can link data across sites and decide which advertisements to show, despite never interacting with the people directly. Identity resolution platforms, clean rooms and cross-context measurement systems all process exactly the intermediate material, hashed emails, pseudonymous identifiers, aggregated segments, that the framework's spectrum is designed to sort. The inference criterion sets a technical bar that lookalike modelling and propensity scoring will need testing against, since targeted advertising leans on inferred attributes far more than declared ones. And the recommendation of periodic reassessment carries standing operational cost: re-identification techniques and auxiliary information both grow more available over time, so a dataset that passed yesterday may fail tomorrow. Datasets anonymised under the 2014 criteria need not be retested against the new framework, per the transitional note, though the document describes periodic review as good practice regardless.

The consultation lands in the middle of a live legislative fight over the same terrain. The European Commission's Digital Omnibus package, presented on November 19, 2025, proposes amending Article 4 of the GDPR so that information is not personal data for an entity lacking reasonably likely means of re-identification, an attempted codification of the SRB judgment. The EDPB and the European Data Protection Supervisor pushed back in a joint opinion on February 10, 2026, warning that the proposed text goes beyond a mere codification of the case law, particularly a clause providing that data does not become personal merely because a subsequent recipient holds identification means, which the authorities read as contradicting the SRB ruling itself. Germany, for its part, pressed the Commission for legislative clarity on anonymisation and pseudonymisation in an October 23, 2025 proposal, nine months before the board produced its own answer. A stakeholder event on December 12, 2025 drew 115 participants into the drafting process. By anchoring its technical framework so heavily in the SRB judgment, the board has effectively staked out its interpretation of the case law before the legislature finishes rewriting the statute, a sequencing move that guarantees the two texts will be read against each other whatever the Digital Omnibus negotiations produce. The consultation mechanics themselves reward attention. Submissions go through the board's provided form, will be published directly on the EDPB website, and may separately be subject to access requests under Regulation 1049/2001 governing public access to EU institutional documents; the secretariat screens replies only to block spam, leaving the substance untouched. Industry associations on both sides of the argument now have a formal channel and a hard deadline, and the positions filed by the end of October will double as the record each camp cites when the Digital Omnibus text reaches trilogue. Whether the final guidelines track the consultation draft closely is genuinely uncertain, precisely because the underlying statutory definition is being contested in parallel. Feedback closes October 30. The definition of personal data in Europe, the foundation on which the entire consent and identity apparatus of digital advertising rests, is now in motion at both the interpretive and the legislative level simultaneously.

The thread running through the week

A pattern connects these five stories, and it is worth stating plainly. Ownership, measurement and legal definition are all being renegotiated at once, and faster than operating practice can absorb. Private capital is repricing the middle of the ad tech stack while the public markets look away; the Criteo bid and the IAS leadership change are two expressions of the same appetite. IAB Europe's figures show demand consolidating into video, social and retail media, formats where a handful of platforms and retailers set the measurement terms. Europe's courts confirmed that gatekeeper obligations survive contact with the continent's best-funded legal departments, even as new research questions whether the enforcement changes market shares at all. Google extended its measurement perimeter to platforms it does not own in the same week publishers openly debated leaving its index. And the regulators in Brussels quietly reopened the most basic question in the entire system: what counts as data about a person. None of these threads resolves this quarter. All of them will price into budgets, contracts and valuations before the year ends.

Also noted

  • July 7, 2026: Dentsu agreed an API partnership with Meta to integrate Creator Marketplace and Partnership Ads into its dentsu.connect operating system, allowing social listening, creator selection and paid activation from a single dashboard, Digiday reported.
  • July 8, 2026: Google Shopping data covering 500 European advertisers shows Temu's auction presence halving since March as the EU's new 3 euro per-item parcel fee bites, with SHEIN approaching a full exit from the channel ahead of an early Prime Day, per PPC Land.
  • July 8, 2026: An Adobe survey of more than 1,000 consumers found just 17 percent of shoppers recall advertisements 24 hours after exposure, tying forgettable creative to irrelevance and mistrust, PPC Land reported.
  • July 8, 2026: New Jersey enacted a ban on the sale of precise location data, extending the state-level privacy patchwork that governs American ad targeting, MediaPost reported.
  • July 8, 2026: Warner Bros. Discovery is consolidating its siloed advertising units onto a single platform built on six named AWS services, including Bedrock AgentCore and SageMaker, with unified planning due in the third quarter and order management to follow in the fourth, per PPC Land.