For years the argument about sensitive data in programmatic advertising ran on hypotheticals. What would it actually cost a company if a health attribute leaked into an ad request? Nobody had a number. Now there is one.

Grindr agreed to pay £26 million to settle a group claim brought in the High Court of England and Wales by roughly 12,000 users, PPC Land reported on September 8. The money arrives in two equal instalments: £13 million due December 31, 2026, and £13 million due March 31, 2027, each worth about $17.6 million at the September 3 exchange rate. London firm Austen Hays, part of Gateley, represented the claimants. Divided evenly, the settlement works out to about £2,167 a person.

The conduct at issue is narrow and dated. Between July 20, 2018 and April 7, 2020, the claim alleged, Grindr passed HIV status alongside advertising identifiers, IP addresses, device specifications, GPS coordinates, age, gender and App IDs to advertising partners. The company worked with between seven and ten such partners over that window. The group action was issued in April 2024; Grindr was served in April 2025; settlement was reached on September 2, 2026. A Form 8-K signed by chief financial officer John North followed on September 4, and The Guardian reported the terms on September 7.

Grindr conceded nothing. The settlement carries no findings and no admission of liability, and the company continues to dispute the allegations while acknowledging distress and loss of trust among some UK users over that pre-2020 period. It described itself as remaining a safe space committed to transparency, user control and responsible data practices.

What makes the figure interesting is the comparison sitting next to it. Norway's data protection authority fined Grindr 65 million kroner for the same conduct over the same dates, roughly £4.8 million. The Oslo District Court upheld the fine on July 1, 2024, and the Borgarting Court of Appeal dismissed Grindr's appeal on October 21, 2025. The UK settlement prices the same exposure 5.4 times higher. Private litigation, not the regulator, set the ceiling.

The Norwegian ruling also did the technical work that makes the UK number legible. The appeal court found that sharing App IDs alone constituted special category data under Article 9 of the GDPR, because the identity of the app itself reveals sexual orientation regardless of what else travels with it. That is a finding about plumbing rather than intent. An app name in a bid request is not usually treated as a health or orientation signal by the people building the pipes. The Norwegian court treated it as one. The same ruling held that users faced a false choice between declining the app and accepting the entire privacy policy, which did not amount to voluntary agreement, and it called Grindr's public claim not to sell user data to third parties for advertising clearly misleading. The original 65 million kroner decision had already established the outline.

Neither settlement document names the advertising companies that received the data. Norwegian proceedings identified MoPub, which itself worked with 160 advertising partners, and AppNexus, which held rights to share onward with 4,000 more. AppsFlyer appeared through the original complaint. The arithmetic of downstream distribution is what turns a single tracking SDK integration into an unmappable exposure surface, and it is why a claim about one app's 2018 configuration can take eight years to price.

Context matters on scale. The conduct occurred under the ownership of Beijing Kunlun Tech, before US national security concerns forced a sale to San Vicente Acquisition in 2020 for $608 million. Grindr went public through a SPAC merger in 2022 at a $2.1 billion valuation and now carries a market value of about $2.65 billion, which puts the settlement at roughly 1.3% of the company. Its advertising business has grown throughout: full-year 2024 advertising revenue reached $53.7 million, up 56%, and second-quarter 2026 advertising revenue hit $25 million, up 44%. The penalty is real money and not an existential threat, which is precisely the shape of a cost that gets budgeted rather than avoided.

Comparable filings have been accumulating. A jury found Meta violated privacy law by collecting health data; Healthline settled the largest CCPA case to date for $1.55 million; a California data broker was fined for selling lists organised by medical condition; and dating app activity has continued to flow into commercial targeting systems. What the Grindr settlement adds is a per-claimant figure in a jurisdiction that permits group actions, which changes the calculation for anyone building an audience segment out of an app category.

A separate story published the same day, September 8, shows the other design path. AdExchanger profiled myGaru, a company that persuaded Ukraine's three rival mobile operators, Kyivstar, Vodafone Ukraine and lifecell, plus the landline operator Ukrtelecom, to collaborate on advertising identity without exchanging customer records. The platform builds pseudonymised signals from carrier subscribers and issues one-time tokens for targeting, so no personally identifiable information moves between the parties. Kyivstar's addressable reach went from 35% to 97% after integration. The company says it now reaches more than 115 million users across Ukraine, Vietnam and Azerbaijan. Yuriy Gorokhov, head of its ad platform, framed the ambition bluntly: "Basically, we want to accumulate more data signals than any walled garden while not being one ourselves." Chief executive Vitalii Morozenko was equally blunt about the category's record, noting that "there is no telecom in the world that succeeded with this."

Two approaches to the same raw material, then, published within hours of each other. One priced at £26 million after the fact. The other engineered so that the sensitive field never leaves the operator in the first place.

Eleven retail media networks, and the search for something other than scale

Retail media has spent five years selling the same proposition: purchase data, closed loop, measurable. That pitch works until every retailer makes it. At the Ascendant Network Showcase in New York on September 3, fifteen presentations from more than eleven networks laid out what each one now claims the others cannot do, and Digiday's Mitchell Parton catalogued them on September 9.

The pitches split along a clean line. Some networks sell a specific population; others sell a specific moment.

Home Depot's Orange Apron Media, which launched a self-service platform for advertisers earlier in its build-out, took the journey argument. Arun Ramaswamy, vice president of tech and product, and Kailey Emery, director of ad sales marketing, described visibility running from browsing through video viewing to chatbot interaction, and cited a 30% lift in brand trust, a 70% increase in recall and a 40% boost in purchase intent for brands appearing alongside the retailer. Those are attitudinal metrics, and the presentation did not disclose sample sizes, survey vendor or category mix.

Chase Media Solutions made the most direct claim on raw signal. Lauren Griewski, chief growth officer, pointed to $1.9 trillion in debit and credit sales in 2025 and characterised the resulting dataset as "actually a window into how consumers live, spend and make decisions." A bank is not a retailer, and the pitch depends on that: card data spans merchants rather than aisles.

PayPal Ads made the same cross-merchant argument from a different starting point. Mark Grether, senior vice president and general manager, described visibility across PayPal, Venmo, partner connected TV platforms and programmatic supply, with recently launched off-site ads carrying instant checkout. Instacart, through general manager of advertising Ali Miller, went further and declined the retailer label entirely, positioning itself as a retail technology company whose Carrot Ads stack powers other grocers' networks alongside its own marketplace, connected TV and in-store screens. Instacart's own advertising revenue reached $243 million in a recent quarter, a 7% year-on-year gain.

DoorDash sold timing. Tim Castree, chief marketing officer, together with Katie Daleo, general manager of CPG ads, and Peter Giordano, general manager of ads platform and growth services, described behavioural patterns visible in billions of annual orders: health food orders running 30% higher on Tuesdays than Fridays, toothbrush orders jumping nearly 30% at weekends. The platform now describes itself as a global commerce media business with 400,000 advertisers.

Then the audience specialists. Dick's Media, through vice president of media David Young, leaned on youth sports, citing a survey finding that 63% of US parents have children in youth sports and that 84% of those parents say youth sports influence their purchase decisions; the GameChanger streaming network, acquired in 2016, supplies the inventory, and the targets are quick-service restaurants, packaged goods brands and carmakers. Chewy's Frank Mulcahy pointed to a subscription base where 84% of sales are automatic, describing a cohort model rather than a reach model. H-E-B's Stacy Torstrick sold Texas: eight million households a week across almost 500 stores, with planning and forecasting tools built on insights competitors cannot reach.

Two networks argued for the upper funnel. Brian Monahan, senior vice president of retail media at Albertsons Media Collective, described a shift from item-and-price advertising toward content distribution, citing Rico's Tacos micro-sitcom ads that outperformed benchmarks by 200% and an America 250 campaign associated with a 24% increase in units sold for participating brands. Michael Krans, vice president of Macy's Media Network, made the sharpest version of the argument: "Most retail media strategies focus heavily on demand capture. At Macy's, we think the opportunity is larger. We help brands create demand, shape choice and capture demand."

Walgreens Advertising Group offered the least glamorous and possibly most consequential differentiator. John Storms, vice president of digital and retail media, who joined in March 2026 from Lowe's Media Network, described internal alignment across merchandising, media, marketing and digital teams, and announced digital screens across 1,200 stores rolling out from October 2026 in partnership with Looma. Organisational coherence is hard to demo and harder to copy.

Also presenting: AD Retail Media from Ahold Delhaize USA, Dollar General Media Network, Backpack Media from the student lender Sallie Mae, and Ace Hardware's RedVest Media.

Underneath the positioning, the buy-side question has not moved. Ramaswamy named it directly, listing measurement, specific outcomes and interoperability across channels and platforms, and reporting that incrementality is a key ask: prove that the investment truly works. The IAB has been pushing the same point, arguing that legacy measurement understates retail media's real value and publishing an incrementality framework for commerce media budgets. Eleven differentiated pitches still resolve to the same holdout study design when a budget holder asks what would have happened anyway.

The cord-cutters stopped leaving, and bundling is the reason

The pay-TV decline that has driven a decade of media strategy slowed markedly in the second quarter of 2026, and the cause appears to be packaging rather than affection. Tim Peterson set out the numbers in Digiday's Future of TV Briefing on September 9.

Charter lost 21,000 video subscribers, against 80,000 in the same quarter a year earlier. Comcast lost 280,000, down from 325,000. EchoStar lost 241,000, down from 261,000. Charter attributed the improvement in its quarterly announcement to "simplified pricing and packaging and benefits from the inclusion of programmers' streaming applications in Spectrum's expanded basic video packages."

That sentence describes the whole mechanism. ESPN Unlimited, Fox One, Paramount+, Peacock and Netflix now arrive inside expanded basic tiers. A household that would have cancelled to buy three streaming subscriptions separately finds the subscriptions already included, and the arithmetic that made cutting attractive stops working.

The behavioural data explains why the distributors reached for this lever. More than 30% of US households are now streaming-only. Among those who cut the cord, 72% already subscribe to streaming services before cancelling, and 31% sign up for a new streaming service within a month of cancelling, which means the saving is smaller than the decision implies. Paramount+'s premium tier ranked first among the services cord-cutters take up, with Netflix appearing in the second, fourth and fifth slots and Peacock's premium tier third. Cutting the cord does not mean leaving the pay ecosystem either: 47% of cord-cutters and cord-nevers hold a streaming pay-TV subscription of some kind, and 20% take broadcast television over free over-the-air signals.

For advertisers the consequence is not that linear inventory has been reprieved. It is that the boundary between linear and streaming inventory has become an accounting distinction inside a single bundle, at exactly the moment when the currency underneath it changed. Nielsen reported ad-supported viewing at 71.5% of US television consumption as it moved to a new currency basis at the end of August. Distribution consolidation has been running in parallel, with Fubo and Hulu + Live TV completing their merger to form the sixth-largest pay-TV provider, Comcast packaging a $70 sports tier, and Peacock Premium Plus launching on Prime Video as streaming prices rise. A connected TV plan built on the assumption of steady linear erosion now has to account for a base that is holding, inside packages the distributor controls, with the streaming services counted twice depending on which report is open.

Google says compliance broke its search, then publishes the manual for it

On September 8 a Google official told Reuters that changes made for the European Union's Digital Markets Act "mark the largest reduction in quality of service at the world's most popular internet search engine in its 29-year search history." Barry Schwartz reported the remark at Search Engine Roundtable the same day.

The claim is specific rather than rhetorical. The changes at issue concern vertical search services, the price comparison platforms such as Expedia and Booking.com that the Digital Markets Act requires Google to display prominently alongside carousels of hotels, airlines and restaurants. Restructuring the page to make room strips real-time pricing detail from the results. Google pointed back to 2024 testing which, on its account, showed lower user satisfaction, longer search times for hotels, more than 10% traffic loss for hotel websites, and hundreds of thousands of European hotels affected.

Every one of those figures comes from the party with the most to lose, and none of the underlying test methodology is public. That does not make the numbers wrong. It makes them unaudited, which is a different objection and the one worth holding.

The timing of the statement is the more revealing detail. On the same day, Google published documentation in Search Central explaining regional differences in the search experience, also covered by Search Engine Roundtable. The stated purpose was to help publishers, businesses and aggregators learn about the different regional search features available and understand the eligibility criteria and how to participate. The document goes further: "Google is evolving the search results page for certain types of queries to provide new opportunities for Vertical Search Services (VSS), Comparison Shopping Services (CSS), direct suppliers, and content providers." Named surfaces include aggregator units, supplier units and carousels, available in certain countries.

So the company argued to a news agency that the change degraded search more than anything in twenty-nine years, and simultaneously told publishers how to qualify for the units the change created. Both statements can be true at once. Read together they describe a firm that has stopped expecting the obligation to be reversed and has begun operating the new layout as product.

The regulatory backdrop is unambiguous. The European Commission fined Google 890 million euros and gave it 60 days to fix search, a decision that also carried cease-and-desist orders touching nine services and drew a company response arguing the penalty would damage Search itself. Eighteen industry groups had earlier warned the Commission to act on search non-compliance. The dispute has always turned on self-preferencing, and the new documentation is the first artefact that treats the remedy as a permanent surface with entry rules rather than a temporary imposition.

For search marketers working European markets, the practical content sits in the eligibility criteria, not in the quality argument. Aggregator and supplier units are inventory. Somebody will be inside them.

Reporting goes dark in two Google products while a new AI reporting layer ships

Three items from September 8 belong in one paragraph, though they were published separately.

Google Business Profiles insights have shown no data at all for September 2026. Barry Schwartz, who runs RustyBrick alongside Search Engine Roundtable, documented the gap from his own profile: "Google Business Profiles insights analytics report is missing data for the whole month of September 2026. Normally, these reports can be a few days delayed, but now we are 8 days into the month, and I still do not see any data." He described the dashboard as showing a whole lot of nothing. Complaints have appeared on Reddit but remain limited. Google has not commented, and the expectation is a backfill.

The Search Console links report has been stale for about a month. Fresh link data last landed around August 8, 2026. The report carries no refresh timestamp, so the delay is only visible to somebody tracking it deliberately, which Schwartz has recently started doing. Google has not commented on that either.

Meanwhile the Gemini-powered dashboards Google showed in May 2026 and said in August would roll out shortly have started appearing in live accounts. Thomas Eccel posted screenshots to LinkedIn reporting them "live in some accounts! Spotted them and they are amazing!" The feature accepts a text prompt and returns interactive charts, graphs and tables built from account data, with written summaries explaining the patterns. PPC Land covered the text-prompt dashboard build when it was announced.

The sequencing is worth stating plainly. An interpretation layer that explains why the numbers moved is shipping across the advertising product at the same time as two adjacent reporting surfaces are silently failing to update. Those are different products with different teams, and no causal link is implied. The dependency runs one way, though: a natural-language summary is a function of the underlying feed, and it has no mechanism for announcing that the feed is a month old.

There is precedent for the gap staying open a long time. A Search Console impressions bug ran for nearly a year before anyone noticed. Performance data froze for an extended period from October 19 in an earlier incident, and the platform has acknowledged latency in performance reports before. Business Profile data, meanwhile, has grown more consequential rather than less, having become a data layer feeding AI surfaces, Maps and Search rather than a listing dashboard. A month of missing local performance data now propagates further than it did when the report was a curiosity.

Also noted

  • September 8: Amazon launched Business Deals, a promotional mechanism visible only to registered business customers across 10 markets and more than 11 million business organisations, requiring a minimum 10% discount from Business Price and 15% during high-velocity events such as Prime Big Deal Days and Cyber Monday. PPC Land
  • September 8: Google Shopping ads have been showing sales-volume badges reading along the lines of "200+ sold last week," spotted by Arpan Banerjee, who noted the format appearing more aggressively of late. Search Engine Roundtable
  • September 8: Google is testing paginated follow-up questions in AI Mode on both mobile and desktop, with a skip button that lets a user escape the automatically generated next query, spotted by Sachin Patel. Search Engine Roundtable
  • September 8: L'Oréal's Australia-New Zealand chief digital and marketing officer Georgia Hack described using Flint, a proprietary model from the Australian startup Springboards, to generate 40 campaign ideas over two weeks for a CeraVe push at male audiences, landing on the tagline "CeraVe, Your Skin Barrier's MVP." AdExchanger
  • September 8: Samsung's US marketing chief Allison Stransky is leaving to become chief marketing officer at Boston Beer Company. Adweek