Google gives every site AI search reports, minus the clicks and queries

A dashboard is a claim about what matters. Whatever it counts becomes the thing people optimise; whatever it leaves out stops existing in the argument, however real it remains in the business.

On August 31, 2026 Google finished rolling its generative AI performance report and its AI feature opt-out control to every website in the world. The report says how often a page turned up inside an AI answer. It does not say whether anyone clicked, and it does not say what was asked. In the same week the company kept narrowing the routes by which anybody outside Google could count search results independently: a new verification screen standing between a browser and a results page, passthrough links that cost rank trackers hundreds of requests to resolve, and a lawsuit against the largest commercial scraper of its index. Then, on September 1, Google quietly unlinked an unknown number of Local Service Ads accounts from the Business Profiles that feed them, during the same migration that deletes years of historical reporting unless advertisers export it first.

Elsewhere the counting problem took less adversarial forms. Sony turned 125 million logged-in PlayStation accounts into a free ad-supported television service. The Weather Company spent a year in an fMRI lab trying to work out which neural pathways light up when a forecast and an advertisement arrive together.

One question runs underneath all of it: who is permitted to see the results, and who is permitted to see the numbers about the results.

The meter Google built for itself

Google rolled the Search Console generative AI performance report and the AI feature control out to all websites worldwide as of August 31, 2026, ending a phased release that began in the first week of June. The company's own wording was flat: as of August 31, 2026, the control has been rolled out to all websites worldwide.

The report itself arrived on June 3, 2026, announced by Hillel Maoz and Moshe Samet, and PPC Land covered it as the first time Search Console had given AI surfaces their own view. Five dimensions are available. Impressions record how often URLs appeared inside generative AI features in Search and Discover. Pages identify which URLs surfaced. Countries break visibility down geographically. Devices apply to Search only. Dates run at hourly, daily, weekly and monthly granularity.

Two dimensions are absent, and they are the two that convert visibility into revenue. There is no click data. There is no query data.

The gap is structural rather than accidental. Before June, appearances inside AI Overviews were folded into aggregate Web Search totals, which made attribution impossible in a different way: a publisher could see that impressions had risen and clicks had fallen, without being able to say which surface was responsible. The new tab does not create a separate silo. AI impressions remain inside the overall performance totals; the report simply carves out a dedicated view. So the number that has moved most sharply since AI Overviews launched in May 2024, the ratio between being shown and being visited, is still the number the tool declines to isolate. Research cited in PPC Land's coverage documented a 61 percent reduction in organic click-through rate for queries where AI Overviews appeared. An Ahrefs study put the figure at 58 percent for top-ranking pages.

The control that shipped alongside it has a clearer provenance. Google did not volunteer it. PPC Land reported in June that the toggle arrived as a response to a United Kingdom order, specifically the first binding conduct requirement imposed on the company under the Digital Markets, Competition and Consumers Act 2024. The Competition and Markets Authority designated Google with Strategic Market Status in September 2025, on the basis that it accounts for more than 90 percent of general search queries in the United Kingdom. The requirement obliges Google to let publishers exclude content from AI Overviews, AI Mode and AI features in Discover, either as links or as grounding material, without touching conventional rankings.

One clause in that requirement is worth reading closely, because it anticipates the obvious move. Google shall not seek to prevent publishers from using the controls, nor penalise them for doing so by reducing conventional search visibility. Regulators wrote an anti-retaliation provision into a remedy about a checkbox, which is a reasonable indication of how much trust the arrangement carries.

The compliance calendar is long. Main control obligations take effect on December 3, 2026. Page-level granularity, meaning the ability to exclude specific URLs rather than a whole domain, is not required until March 3, 2027. The first CMA compliance report is due on January 3, 2027. The requirement expires on October 10, 2030.

Between now and December, then, publishers hold a domain-wide switch, an impressions count, and no way to price the trade. Flipping it removes a site from the surface that Google is expanding fastest. Leaving it alone keeps the impressions, whose commercial value the reporting does not disclose.

Search Console's own recent record does not help. The September 2026 Google Webmaster Report lists a performance report incident around August 12 in which impression and click figures dropped without cause, crawl stats missing two days of data, and a favicon reporting bug since resolved. It also notes that AI Mode queries are now displayed, separately from the generative AI reports, which means the AI reporting surface is currently split across two places that measure different things. The same monthly summary records the August 2026 spam update rolling out between August 18 and 21, ranking volatility on August 5 and 6 and again on August 12 and 13, AI Overviews running on Gemini 3.7 Flash, and the shutdown of Google Assistant on mobile devices scheduled for September 4.

The turnstile

If the first story is about what Google chooses to report, the second is about what anybody else is still allowed to look at.

Search Engine Roundtable documented a new interstitial requiring users to click a Continue button before search results appear. The mechanics are deliberate. The button renders greyed out, becomes blue after a few seconds, and only then accepts a click. Glenn Gabe, who first shared the screenshot, described the pause exactly: the Continue button was greyed out for a few seconds before turning blue. Alexis Rylko observed the same screen in other language versions, which suggests a test running across multiple markets rather than a single-locale experiment. The delay is the point. It gives Google's back end time to complete a server-side check before the page becomes interactive, which defeats automation that expects a results page to be one request away.

This sits inside a pattern with an eighteen-month history. In September 2025 Google eliminated the num=100 parameter, which had let a single request return a hundred results, and forced a roughly tenfold increase in API calls for anyone tracking the top hundred positions. Semrush confirmed significant operational impact. Ahrefs eventually restored top-100 tracking at higher cost.

Then came the passthrough links. Alex Greenland observed the first google.com/goto URLs on June 23, 2026. Brodie Clark documented the pattern across both organic and paid listings on July 2. By August 26 the rollout was close to complete and Google had acknowledged it. Destination URLs in results are now encoded strings that resolve through a 302 redirect, and they cannot be decoded locally. HEAD requests are blocked. The only workaround is a GET request that stops at the redirect, which turns one lookup into one network round trip per link. Derek Perkins of Nozzle put the resulting arithmetic on the record: resolving all links for a single five-page ranking takes between 500 and 1,000 requests.

Google's public framing has been consistent and unspecific. The company said it has a long history of deploying technical measures against evolving forms of abuse. No target was named. No technical documentation followed, and no rate limits were published, which leaves vendors to infer the boundaries by hitting them.

The litigation runs alongside. Google sued SerpApi over scraping, lost a bid to treat search scraping as a Digital Millennium Copyright Act circumvention question, saw SerpApi file to dismiss, restricted the Light Fast API workaround to three results, and has now amended the complaint to build on licensed content, including Reddit's licensing terms. The September webmaster report records the amendment as a live development.

Put the two halves together and the shape is plain. The independent measurement of search results is getting more expensive, more legally exposed and technically harder, month by month. The first-party measurement of AI search results has just become universally available, free, and silent on clicks and queries. Neither development is a secret. Each has a defensible justification on its own terms. The combined effect is that the number of organisations able to say what Google search actually does is falling.

Google Discover supplied a small demonstration of the resulting dependency on September 1. Articles failed to load or loaded only after long delays from roughly 6:00 a.m. Eastern, with Valentin Pletzer flagging the problem on X and Damien Adell posting traffic charts showing a sharp drop from 10:00 Central European Summer Time. By 7:40 a.m. Eastern, Adell reported recovery. Roughly one hundred minutes of publisher referral traffic went missing, and no Google statement accompanied it. Publishers learned about it from each other.

Google unlinked its own products

The clearest failure of the day involved two Google systems losing track of one another.

On September 1, Local Service Ads accounts began detaching from the Google Business Profiles they depend on, without any action by the account holder. Jack Hepp reported an unlinked account that required support intervention to restore, alongside repeated error messages, platform slowness, and an inability to view Local Service Ads inside the associated Google Ads accounts. Ginny Marvin, Google's Ads Liaison, confirmed the problem in short form: a bug had been identified and was being actively worked on. She added that it was unrelated to the migration.

The migration is what gives the timing its edge. Google is retiring the standalone Local Services Ads dashboard and moving campaigns into a Performance Max variant inside Google Ads, starting in August 2026 with United States home service categories: plumbing, heating and air conditioning, electrical, roofing, pest control and related verticals. Service-area businesses and custom bidding accounts follow in late 2026. Non-United States accounts and the remaining verticals move in 2027.

Much of the account survives the trip. Pay-per-lead billing continues. Placement stays limited to Search and Maps. Targeting remains keywordless and category-based. Lead history, contact details, message threads and call recordings transfer into a new Lead Manager interface, and verified business badges carry across without re-verification.

Historical performance reporting does not survive. Impressions, clicks, weekly spending and ad-level breakdowns disappear at transition. Administrators receive fourteen days of notice before their individual migration date and must download the data manually inside that window. A Google product liaison put it as a strong recommendation that all advertisers download their historical data ahead of their transition timing. Manual bidding disappears, as do vertical-level target CPA settings and Better Business Bureau callouts. Weekly budgets convert to daily amounts. Business Profile core details, meaning name, address and hours, sync one way into Google Ads and can only be edited at the profile.

So the bug landed on accounts in the middle of a fourteen-day export window, on a platform returning errors, in a product whose historical record is about to be deleted, and it severed the link to the profile that the migration itself uses to establish identity. Google has said that accounts without a Business Profile link would have one created automatically during migration, which is a reasonable safeguard for an account that never had a link and a less reassuring one for an account that just lost the link it had.

October brings another change to the same product. From October 1, 2026, Google begins charging for unanswered calls that last more than 20 seconds during business hours, with an exception where call routing requires a key press, and begins charging for qualifying follow-up calls from a user whose first contact did not qualify. The billing definition widens in the same quarter that the reporting used to audit billing narrows. Google has referenced unspecified protections against robocalls and spam without publishing dispute procedures or timer mechanics.

Sony turns 125 million sign-ins into channels

New inventory arrived from an unexpected direction. AdExchanger's Tuesday roundup reported that Sony launched Live TV on PS5, a free ad-supported service carrying more than 100 channels on the PlayStation 5.

The asset is the install base. Sony counts roughly 125 million logged-in users, a number that arrives with authenticated identity, a payment relationship and a device that sits on the main television in the house. Programming spans sports, news, action, horror and community channels, with two new anime channels for Crunchyroll, which Sony owns. Ad serving runs through Publica and PubMatic.

The motivation is margin. Console manufacturing costs have risen, and chief executive Hiroki Totoki has said that marketing to new customers ranks below monetising the existing base. A console that already has the customer does not need to sell another console to make more money from that customer; it needs to sell the customer's attention.

Both ad tech partners have context worth attaching. Publica is owned by Integral Ad Science and handles ad pod decisioning, unified auctions and server-side ad insertion for streaming environments; its Samsung partnership was renewed as an exclusive global agreement on September 30, 2025, covering a Samsung TV Plus footprint of 88 million monthly active users. PubMatic reported revenue up 11 percent after three quarters of decline. The free ad-supported category they are selling into has been growing quickly, with channel counts up 42 percent as programmers moved beyond back-catalogue material and European penetration reaching 27 percent.

The audience is the uncertainty. Console gamers have historically been among the most advertising-averse segments in media, and they are underrepresented in mainstream connected television panels, which means the buyers evaluating this inventory will be doing so without much comparable measurement history. Nor has Sony published the details that would let anyone price it. No launch markets. No timing beyond the announcement. No CPM guidance, pod structure, frequency rules or measurement partners. No statement of whether the inventory clears programmatically, through direct deals, or both.

That last omission matters more than it used to. A new inventory class entering a market where four demand-side platforms held 85 percent of global programmatic spending in the first quarter of 2026 will be priced by mechanics that, for now, only the seller can see.

What a forecast does to a brain

The last story is a measurement experiment of a different kind, and it is the one furthest from any accepted currency.

AdExchanger reported that The Weather Company, which operates The Weather Channel app, has spent about a year working with the neuroanalytics firm Neuro-Insight on what it calls emotional targeting. The premise treats a forecast as an emotional trigger rather than a data point. Shannon Ferguson, the company's director of measurement, described the objective as understanding how weather triggers the subconscious mind, and specifically which neural pathways light up when particular advertising is processed alongside certain forecast data.

The commercial test ran with S.C. Johnson. Because weather varies by region, different creative executions were served according to local conditions, with rain in some markets and clear conditions in others. That is a familiar dynamic creative setup dressed in unfamiliar clothes; what is new is the claim that the underlying variable is affective state rather than product relevance.

The measurement apparatus around it is conventional, and deliberately so. The Weather Company maintains in-house attribution linking impressions to actions on advertiser websites, runs brand studies, and certifies vendors for California Consumer Privacy Act and General Data Protection Regulation compliance. Ferguson said the company avoids proprietary frameworks and works inside clients' existing measurement stacks, which is a notable choice in a market where most attention and outcome vendors sell a currency of their own. The company has also built internal AI sales tools that pull performance data, campaign metadata and takeaways from prior campaigns to generate recommendations, with human review retained.

What has not been published is the evidence. No sample size, no effect size, no third-party verification, no indication of how many advertisements were tested or how the neural signal maps to any outcome an advertiser buys. Neuro-Insight's method is proprietary. On present disclosure the work is a hypothesis with a client attached.

Academic neuromarketing offers a benchmark for what published results in this area look like. A study led by Hang-Yee Chan of King's College London, with collaborators at Rotterdam School of Management, Temple University and the University of Amsterdam, used functional magnetic resonance imaging on 113 participants across two countries against 85 video advertisements, validated with behavioural studies on 302 United Kingdom participants. The findings were specific about timing: emotion and memory signals predicted liking within three seconds, social cognition, language and executive function became predictive at five to seven seconds, and perception signals took around ten. Social cognition was the strongest sustained predictor. Emotion peaked early and fell away. Executive function correlated negatively with preference, meaning advertisements that required effortful processing were liked less. The paper was submitted on February 25, 2021 and published in the Journal of Marketing Research on July 28, 2023, a two-and-a-half-year gap that illustrates the distance between laboratory rigour and campaign timelines.

A parallel line of work aims at the same goal by simulation rather than scanning. PyMC Labs and Colgate-Palmolive published research on October 27, 2025 in which large language models reached 90 percent of human test-retest reliability in replicating purchase intent across 57 personal care surveys and 9,300 human responses, using semantic similarity mapping rather than direct numerical ratings. The authors were explicit about the limits: one product category, one company, unvalidated generalisation.

Three methods, one ambition. Each promises to explain response without depending on what a person says afterwards, and each currently sits outside the currencies that money moves through. The dashboards remain the only agreed record, which returns the day to where it started.

Also noted

• September 1: Google Ads sponsored results on Search have begun carrying a block headed "What customers love" and labelled as AI-generated from the store rating reviews, spotted by Sachin Patel, layering machine summarisation on top of store rating summaries that are themselves machine-generated. Search Engine Roundtable
• September 1: Google added a notice to its AI Max help documentation stating that broad keywords are not treated as exact match keywords for prioritisation when the broad match campaign setting is disabled or a campaign moves to AI Max, and that converting broad keywords to exact removes their historical performance data. Search Engine Roundtable
• September 2: Simon Property Group is now pitching its retail media network as an audience product built on consent-based Wi-Fi location data, Simon+ loyalty records and ShopSimon transactions rather than as a screen inventory sale, with chief revenue officer Jared Blechman citing more than 200 centres, nearly 4,000 digital screens, over 2 billion annual visits and campaign measurement against control groups. Digiday
• August 31: Zappi ran the withdrawn Callaway and Good Good driver advertisement through 150 AI synthetic respondents trained on consumer response data and recorded the lowest possible score of 1 out of 5, with shock levels eight times normal and nearly a third of respondents flagging offensive content, after Callaway chief executive Chip Brewer said the video was produced by Good Good but signed off by Callaway and that the approval should never have happened. Adweek
• September 1: Sandhya Devanathan, previously Meta's vice-president for India and Southeast Asia, joined OpenAI to run growth, partnerships, compliance and operations across Southeast Asia and Australia, as OpenAI also began letting some large customers pay for advertising on a performance basis. AdExchanger