Google Exempts Europe From Site Reputation Hits

Estimated reading time: 28 minutes

Regulation reshaped Google’s own enforcement this week, and the numbers behind the August spam update finally arrived. From Sunday, site reputation abuse penalties stop applying to searchers inside the European Economic Area, a carve-out with a sharp edge for UK marketers. Elsewhere in this week’s search marketing news: SE Ranking measured how hard the spam update actually hit, Google confirmed the goto redirect we first flagged as a test in July, AI Mode learned to book hotels, and a study of 107 million AI answers explains why half your citations come from places you don’t own.

Site Reputation Penalties Stop Applying Inside the EEA

From 30 August, a manual action issued under Google’s site reputation abuse policy will behave differently depending on where the searcher is. Google published the change on Search Central, and it follows a European Commission investigation that required Google to stop enforcing that policy within the European Economic Area.

The split works like this:

Outside the EEA: a manual action directly affects results for the offending section of the site, with the rest of the site unaffected, exactly as before
Inside the EEA: the manual action’s impact does not apply. The affected section may instead be separated in Google’s systems so it ranks independently from the rest of the site over time

Read that second point carefully, because it is not an amnesty. Google keeps the underlying mechanism that treats a third-party content section as though it were its own site, which is the structural answer to parasite SEO. What disappears in the EEA is the manual action layered on top.

Now the part that matters most for readers here: the UK is not in the EEA. Britain left the EU and the EEA agreement no longer applies, so searchers in the UK sit outside this carve-out and manual actions will hit their results in the ordinary way. If you have been reading “Europe” and mentally filing this under “applies to us”, it doesn’t. What it does create is a genuine measurement oddity for any client selling across both territories, because the same penalised section can behave one way for a searcher in Dublin and another for a searcher in Manchester. Segment your reporting by country before anyone draws conclusions from a blended visibility chart this autumn.

The August Spam Update Hit 82% Harder Than Normal Churn

Last week we covered the August 2026 spam update starting to roll. SE Ranking has now measured what it did, and the answer is considerably more than routine turbulence.

Analysing US organic results for 100,000 keywords across 20 industries, comparing 17 to 22 August against a quiet baseline of 26 to 31 July, they found 16.71% of URLs that had ranked in the top ten fell beyond position 100 for the same keyword. In the baseline period that figure was 9.2%. In other words, an 82% increase, and a top-ten URL was roughly 1.8 times more likely to vanish from the top 100 entirely during the update.

Falling out of the top ten is ordinary. Falling past position 100 is a different category of event.

Volatility spanned all 20 niches, ranging from 74.64% in real estate to 85.55% in fashion and beauty, with YMYL categories including real estate and healthcare among the steadier. One honest limitation the researchers flag: they did not run a URL or domain-level analysis, so the data shows how much moved without showing what kind of site or tactic was hit. Useful for calibrating client expectations, not for diagnosing a specific loss.

Reports Point the Spam Update at Mass-Generated AI Content

What did it target? Reports gathered by Roger Montti suggest at least part of the August update went after AI-generated content produced expressly to rank, with practitioners describing sites that publish automatically via coding agents being filtered and dropped broadly.

The supporting clue is a Google research paper published recently: the Scalable Cluster Termination System, designed to terminate networks of mass-generated AI spam. Cluster is the operative word, and it echoes the account-level detection approach we covered in July when Google published research on catching AI video spam by grouping accounts rather than judging individual assets.

Montti makes the distinction that matters, and it deserves repeating to any client who asks whether AI content is now penalised. Content is not spam because AI made it. Content drifts towards spam when it is produced primarily to rank rather than to inform, and mass automation makes that drift both easier and more detectable. Treat this as evidence, not proof, since Google has confirmed no new policy types with this rollout. But if a site publishes at volume with minimal human judgement, this is the week to look honestly at what those pages actually offer.

Volatility Carries On After the Rollout Finishes

Tracking tools and forum chatter show ranking volatility continuing after the spam update’s rollout completed, which is normal and reliably misread.

Systems settle. Pages that lost visibility get reassessed, replacements shuffle, and the tools keep registering movement for a week or two afterwards. The practical risk is a client seeing a second wobble and concluding that a recovery attempt has failed, or worse, that a fresh penalty has landed. Neither conclusion is available from volatility data alone.

Hold the line on process: annotate the rollout window, wait for stability before drawing conclusions, and remember Google’s own guidance that recovery from a spam action can take months while its systems build confidence in a changed site. Reacting to week two of a settling period is how sites end up with three half-finished remediation strategies layered on top of each other.

Google Confirms the goto Redirect Is Rolling Out for Real

In July we covered a curious test where Google appeared to be wrapping organic result links in a google.com/goto passthrough URL, and noted it was worth half an eye. It has now been confirmed and is rolling out, with Derek Perkins of Nozzle reporting close to 100% deployment across several residential IP providers.

“We have a long history of deploying technical measures against evolving forms of abuse, and we regularly take steps to protect our services and users.”
– Google spokesperson, on the goto rollout

Read alongside the SerpApi litigation, the purpose is not hard to infer. A redirect hop makes it materially harder for third-party tools, scrapers and AI engines to harvest clean destination URLs from the results page, which is precisely the capability Google failed to shut down in court. Having lost the legal argument, Google is winning the technical one.

The practical consequence lands on your reporting stack. Rank trackers, SERP monitoring tools and anything parsing result URLs may behave oddly over the coming weeks. If your data looks strange, check with the tool provider before assuming a client’s rankings moved, and expect some vendors to need time to adapt.

A Government Site Falls Out of the Results

The IRS website dropped heavily in Google Search this week, spotted and discussed across the SEO community.

Large institutional sites losing visibility abruptly is usually one of three things: a technical fault on their side, an indexing issue on Google’s, or genuine algorithmic reassessment. Whichever it turns out to be, the value for practitioners is as a live worked example. When a site this prominent moves, the community investigates in public, and the diagnostic path from symptom to cause gets documented in a way that no case study ever quite manages. Worth following the thread if you handle large or complex sites, and worth remembering the next time a client assumes big brands are immune to whatever just happened to them.

Reddit Owns 83.9% of the Discussions and Forums Feature

Glen Allsopp and Ahrefs’ data team have quantified something the industry has felt for two years. When Google’s Discussions and forums feature appears, it includes a Reddit link 83.9% of the time worldwide, rising to 87.8% in their SERP sample, with early August data suggesting 86.1%.

The concentration underneath that headline is the more useful finding:

Just three sites account for 76.75% of all Discussions and forums links worldwide
After Reddit come Facebook at 38.3%, Quora at 32%, JustAnswer at 8.2% and Mayo Clinic at 3.3%
Facebook overtook Quora for the first time in January 2026
98.7% of appearances are for informational-intent queries
The feature is most prominent in the US, Canada, the Philippines, Australia and the UK
No correlation was found between backlinks to individual threads and how often they appeared

That last point deserves a moment. The instinct to treat forum visibility as another link-driven ranking problem is wrong, and building links to threads is not the lever. The average top-100 subreddit was sixteen years old, so this is a feature that rewards established communities rather than newly minted ones.

For UK marketers specifically, note that Britain sits among the countries where the feature appears most. Which makes knowing the two or three communities that own your category, and being a genuine participant in them, a legitimate visibility task rather than a growth-hack sideline.

Chunkability: Why Some Content Gets Cited and Most Doesn’t

Lumar has published a useful technical framing for AI citation: chunkability, meaning how cleanly a page breaks into self-contained passages that a retrieval system can lift and use.

The concept explains a frustration many teams have hit. A genuinely excellent long-form guide can go uncited while a thinner competitor page gets pulled into answers repeatedly, because retrieval does not consume pages, it consumes passages. If your best insight only makes sense after three paragraphs of scene-setting, no chunk of it stands alone well enough to be quoted.

The remedies are structural rather than editorial: front-load the answer within each section, keep headings genuinely descriptive of what follows, make sure a passage lifted without its surroundings still states its subject, and stop relying on earlier paragraphs to carry meaning forward. None of that requires writing worse, which is the usual objection. It requires writing in units that survive being taken apart. Given how much of this year’s evidence points at retrieval rather than ranking as the gate, this is one of the more actionable technical ideas to land recently.

107 Million AI Answers Show Where Citations Really Come From

Trendos analysed 107 million AI answers this year and pulled the top ten citation sources for three industries across ChatGPT, Perplexity, Gemini and Google AI Overviews, sorting every cited domain into community and user-generated content, brand and retail, or independent editorial and reference.

One finding held steady across every industry examined: community and user-generated content makes up almost exactly half of the leading citations. Not high in one vertical and low in another, close to 50% throughout. Reddit, YouTube and forums sit under every category, which means a brand absent from public conversation is missing from half the sources AI reaches for.

The other half is where the verticals diverge sharply, and B2B is the striking case. In IT and solutions services, brand-owned sources account for just 2% of leading citations, the lowest of the three industries, while independent editorial and reference climbs to 47%. Much of that is review directories: G2, Clutch, Gartner, GoodFirms, SourceForge. Ask an AI to recommend a software vendor and it declines to take the vendor’s word for it.

That reorders the B2B visibility roadmap quite aggressively. If your own site carries 2% of the citation weight in your category, the highest-return work is claiming and fully completing the two or three directory profiles your buyers actually shortlist from, with every field populated and category wording matched to how prospects phrase their questions. Unglamorous, cheap, and considerably more effective than another pillar page. It’s sponsored research, so treat the exact percentages with the usual caution, but the directional finding matches everything else we’ve seen this year.

AI Recommends Your Brand and Cites Somebody Else

New data shows AI tools recommending brands while citing entirely different sites as their sources, which quietly breaks the assumption underneath most AI visibility reporting.

We have circled this from several directions over the summer. Ahrefs found answers citing a brand’s own listicle then recommending a competitor from it. Search Engine Watch saw an AI Overview skip the brand that ranked itself first. Now the inverse: your brand gets named, and the citation credit goes to a review site, a forum thread or a publisher.

The reporting consequence is uncomfortable. Citation-share tools measure whose URLs get linked, which is not the same as whose products get recommended, and a brand can be winning the conversation while losing the citation count entirely. If your client dashboard has a single AI visibility number on it, that number is answering a narrower question than the client thinks. Track mentions and citations as separate lines, and be honest in the commentary about which one connects to revenue.

AI Mode Starts Booking Hotels and Tracking Flight Prices

Google has begun rolling out hotel booking inside AI Mode, alongside flight price tracking and mileage rate information, part of a set of three new travel features announced this week.

Booking. Inside the answer.

This is the agentic commerce thread arriving in a vertical with real money attached, and it follows the pattern established with Yelp reservations in ChatGPT earlier this month. Discovery, comparison and transaction collapse into one conversational surface, and the website that used to sit at the end of that journey becomes optional.

For travel and hospitality clients the strategic question has shifted permanently. It is no longer only whether you rank for the query, but whether your availability and rates are present in whatever inventory feeds the assistant. That makes your relationships with booking platforms and aggregators a search visibility asset, and your feed accuracy a ranking factor in everything but name. Anyone whose distribution strategy assumed direct booking would always be one click from a search result should be reviewing that assumption now rather than next planning cycle.

ChatGPT’s Browser Gains WebMCP Site Tools

OpenAI has added WebMCP support to the browser built into its ChatGPT desktop app, letting websites expose structured actions that ChatGPT and Codex can use directly from the page: searching documents, editing files, exploring dashboards, comparing travel options, updating shopping carts.

The distinction from existing MCP support matters. Server-based MCP requires setting up a connector; WebMCP lets the page itself offer its tools automatically to any agent visiting it. An arrow appears in the address bar when site tools are available, indicating whether a tool can read data or change it. Site tools need GPT-5.6 Sol or Terra and are not available in Enterprise or Edu workspaces.

Moz published a companion explainer on preparing a site to serve AI agents, which is the more useful read for most teams.

One caveat deserves emphasis, because it will get lost in the excitement: OpenAI explicitly does not connect WebMCP with search rankings, citations or recommendations. Implementing it will not make ChatGPT recommend you. What it does is make your site usable by an agent once someone is already there, which matters enormously if your site is a tool people operate and rather less if it is a brochure. Apply the cats.txt test from last month before anyone sells this as a visibility tactic.

ChatGPT Now Searches Individual Subreddits by Name

Suganthan Mohanadasan has found ChatGPT scoping its background searches to individual subreddits by name, a level of targeting that goes well beyond generic site-scoped queries.

Put this beside two other findings from the last fortnight and a coherent picture forms. ChatGPT’s use of the site: operator in fanout queries jumped roughly 46-fold in a single day earlier this month, meaning it increasingly goes deliberately to specific domains rather than searching the open web and seeing what returns. Now it is choosing specific communities within those domains.

The implication cuts against the panic that followed Reddit’s citation collapse. If the engine is selectively querying named subreddits, then the question stops being how much Reddit gets cited overall and becomes whether your category’s specific communities are among the ones it goes to. That is a far more tractable problem, and a considerably better use of a client’s time than reacting to a platform-level citation-share chart.

AI Brand Preference Splits Sharply by Generation

Following last week’s YouGov consideration data, new figures show AI assistant preference dividing along generational lines, with Claude leading among Gen Z by more than seven to one relative to older cohorts.

Treat generational framing with the scepticism it usually deserves, as the analysis itself does. But there is a practical version of this finding that survives the caveats: different audiences are consulting different assistants, and the assistant doing the consulting shapes which brands get named. Our own coverage this summer found the models behave differently from one another, with Claude and ChatGPT tending to underclaim product features while Gemini overclaims them.

So if your client’s buyers skew young, your AI visibility testing should not be a ChatGPT-only exercise, and a monitoring setup that samples one assistant is measuring one audience. Worth checking which platforms your visibility tooling actually covers before presenting a number as though it described AI search generally.

A Judge Calls the Publisher Situation Really Unfair

A judge hearing a case involving Google and publishers described the situation as one that “seems really unfair”, as reported by Search Engine Journal.

Judicial asides are not rulings and should not be reported as though they were. What they occasionally reveal is which way the intuitive sympathy runs before the legal argument is settled, and on the question of publishers whose content trains and feeds AI answers that reduce their traffic, the sympathy appears to run towards the publishers.

Set it beside the EEA carve-out at the top of this post and a theme emerges for the autumn. Regulators and courts are now shaping the mechanics of search directly, not merely fining companies afterwards. Google is changing how enforcement works in Europe because a regulator required it. For agencies, that means jurisdiction is becoming a genuine variable in SEO strategy, which is a sentence I did not expect to write when this year started.

SerpApi Moves to Dismiss Google’s Amended Lawsuit

The SerpApi case continues. After a judge dismissed Google’s DMCA anti-circumvention claims in July and gave Google 21 days to amend the narrow portion tied to copyrighted Knowledge Panel images, Google amended. SerpApi has now moved to dismiss again.

Nothing is resolved, and the case may run for a long time yet. The reason to keep tracking it is that so much of our industry’s measurement infrastructure depends on the legality of reading public search results: rank trackers, SERP monitoring, competitive analysis, and a good deal of the data feeding AI visibility tools.

The two stories bracket each other neatly this week. Google is fighting scraping in court and simultaneously deploying goto redirects to make it harder technically. The legal outcome may end up mattering less than the engineering one.

Demand Gen Gains Messaging Ads, Travel Tools and AI Video

Google’s August Demand Gen drop adds messaging ads, travel-specific tools and AI video creation to the campaign type, continuing its build-out as the designated home for what used to be Display and video demand generation work.

Messaging ads are the interesting inclusion for UK advertisers, given how much customer contact runs through WhatsApp and similar channels in several client sectors. AI video creation lowers the production barrier that has kept many smaller accounts out of video advertising entirely, though it also guarantees a flood of generic AI-generated video into the same auctions, so early movers benefit most.

The travel tools land in the same week AI Mode learned to book hotels, which is unlikely to be coincidence. Google is assembling the paid and organic sides of travel search in parallel, and advertisers in that vertical should be looking at both halves together rather than treating this as a routine feature note.

What You Lose Moving From Display to Demand Gen

A companion piece sets out honestly what advertisers gain and lose migrating from standalone Display campaigns to Demand Gen, which is timely given Google sunset standalone Display earlier this year.

The honest accounting is the valuable part, because most migration guidance is written as though nothing is lost. Placement-level control, certain targeting granularity and some reporting detail change or disappear; reach into newer surfaces, better creative tooling and stronger automation arrive in exchange. Whether that trade favours a given account depends almost entirely on how much of its Display performance came from manual placement management.

If you still have accounts working through this migration, read it before the next optimisation session, and set client expectations about what will no longer be adjustable rather than discovering it mid-quarter.

Microsoft Adds a Landing Page Report for AI Max

Microsoft Advertising has added a search term landing page report for AI Max, showing which landing pages its final URL expansion actually sent traffic to.

This is the reporting that makes automated URL expansion auditable rather than a black box, and it arrives a week after Microsoft rolled AI Max out globally with its advertiser controls already in place. The contrast with Google’s approach continues to be instructive: Google is auto-upgrading campaigns into AI Max from 1 September with text customization enabled by default, while Microsoft ships controls and transparency alongside the feature.

If you run AI Max on Microsoft, this report is the first thing to open. Final URL expansion sending traffic to a page you would never have chosen is the failure mode worth catching early, and now you can see it happening rather than inferring it from odd conversion patterns.

Google Ads Adds a Performance Fluctuation Troubleshooter

Google Ads is rolling out a new troubleshooting view for performance fluctuations and changes, aimed at explaining why a campaign’s numbers moved.

Useful, and worth approaching with the same caution as the Ask Advisor insights and missed-growth estimates we covered earlier in the summer: the tool explaining the fluctuation is built by the party that benefits from the resulting spending decision. An explanation that concludes “increase budget” is not wrong by definition, it simply needs verifying independently before it becomes a recommendation with your name on it.

Where it should genuinely help is triage speed. Account managers spend real hours reconstructing what changed and when, and anything that surfaces the change history and competitive shifts in one place saves that time for analysis. Use it to find the question faster, not to skip answering it.

Ask Advisor Is Getting Shared Memory Across Tools

Google’s Ask Advisor will soon carry shared memory across tools, so context established in one place persists into another rather than starting fresh each time.

Anyone who has used an AI assistant seriously knows how much of the friction is re-explaining the situation, so this is a real usability gain. It also quietly changes what the assistant is: a system that remembers your account’s context across Ads and Analytics is closer to a junior analyst than a help feature.

The governance question follows immediately, and agencies should settle it before the feature lands rather than after. If Ask Advisor accumulates context about a client account across sessions and users, whose assumptions is it carrying forward, and who checks them? A remembered misunderstanding is more durable than a fresh one.

Shopping Ads Push Deeper Into AI Answers on Both Platforms

Two moves in the same direction this week. Google is testing an inline shopping ads carousel placed within AI Mode content rather than beside it, and OpenAI has pinned shopping to the ChatGPT sidebar.

Inline is the word doing the work in the Google test. An ad unit inside the generated answer, rather than above or below it, is a different proposition for both advertiser and reader, and it continues the pattern where AI Mode functions as the laboratory for formats that later reach AI Overviews.

Retail clients should be planning for shopping placements inside AI answers as a standing channel rather than an experiment, which brings feed quality back to the centre of everything. It is the third consecutive week where the practical answer to an AI search question turns out to be product feed accuracy, which tells you where the real advantage genuinely sits for ecommerce accounts this year.

The Performance Max Brand Leak Audit

PPC Hero has published a method for measuring what it calls the brand leak in Performance Max: the portion of spend going to conversions you would have won anyway through branded search.

Every practitioner suspects this is happening. Few can quantify it, which is exactly why PMax reporting arguments go in circles. An audit method that puts a number on incremental versus cannibalised conversions turns a hunch into a testable claim, and it is the sort of analysis that changes what a client believes about their best-performing campaign.

It pairs neatly with the August bidding change we covered last week. If Smart Bidding now optimises strictly to your stated target regardless of budget constraint, then understanding which conversions were genuinely incremental becomes more valuable, because your targets are doing more of the work than the budget cap used to. Worth running on any account where PMax quietly became the largest line item.

Black Friday Arrives Into an Unusually Unstable Quarter

It’s late August, which means Black Friday planning is already late, and PPC Hero argues this year’s peak season carries unusual risk.

Look at what retail accounts are absorbing simultaneously. AI Max auto-upgrades land on 1 September with text customization on by default. The Smart Bidding change altered how budget-limited campaigns behave from 17 August. Shopping ads are appearing inside AI answers on two platforms. A spam update has just moved organic rankings more violently than usual. Every one of those changes affects peak trading, and none of them was scheduled around retail’s calendar.

The planning advice that follows is unromantic. Freeze what you can before November, get the AI Max decisions made in September rather than discovering them in October, bank your bid target choices now, and build the reporting that will let you separate a platform change from a trading problem when something moves in week two of peak. The teams who go into December with clean baselines will be the only ones able to explain what happened.

Strategic Direction: Where Search Marketing Is Heading

Jurisdiction became an SEO variable this week. Google is switching off a category of penalty for searchers inside the EEA because a regulator told it to, while keeping the same penalty live everywhere else, and the UK sits on the outside of that line. Agencies serving clients across both territories now have to segment reporting by geography before drawing any conclusion about enforcement, and the wider lesson is that “how Google works” is fragmenting into “how Google works here”. Expect more of this, not less, and expect the differences to be invisible in blended data.

The citation research keeps converging on the same uncomfortable answer. Half of AI citations across every industry examined come from community and user-generated content, brand-owned sources carry as little as 2% in B2B, and Reddit alone appears in 83.9% of Google’s forum results. Meanwhile AI recommends brands while citing other sites entirely, so the metric most tools report is measuring an adjacent thing to the outcome clients care about. The work this points to is genuinely different from classic SEO: showing up credibly in places you do not control, completing directory profiles you have ignored for years, and writing in passages that survive being lifted out of context.

And the transaction keeps moving inside the answer. Hotel booking in AI Mode, shopping pinned in ChatGPT’s sidebar, inline shopping carousels inside generated content, WebMCP letting agents operate your site directly. The click was the unit of measurement for twenty years and it is quietly ceasing to be the unit of value. What replaces it, for most businesses, is presence in the feeds, inventories and communities the assistant consults before it ever mentions your name.

Key Takeaways

Remember the UK is not in the EEA: site reputation manual actions still apply here in full from 30 August, and cross-territory clients need country-segmented reporting to see the difference
Calibrate spam update conversations with SE Ranking’s finding that 16.71% of top-ten URLs fell past position 100, an 82% increase on normal churn
Audit high-volume AI-assisted content honestly, since reports and Google’s own cluster-termination research both point the update at content built to rank rather than to inform
Warn clients that rank tracking data may wobble as goto redirects roll out, and check with tool vendors before treating odd numbers as ranking movement
For B2B clients, claim and fully complete G2, Clutch and Gartner profiles: brand-owned sources carried just 2% of citations in that vertical against 47% for independent directories
Identify the two or three communities that own your category rather than chasing Reddit generally, now that ChatGPT queries individual subreddits by name
Restructure key pages for chunkability so passages stand alone, because retrieval consumes passages rather than pages
Get AI Max decisions made in September, before peak trading collides with the 1 September auto-upgrade and the new bidding behaviour
Treat WebMCP as a usability feature, not a visibility one: OpenAI explicitly does not connect it to rankings, citations or recommendations

Frequently Asked Questions

Does the EEA site reputation change apply to UK websites?

No. The UK is not part of the European Economic Area, so searchers in the UK fall outside the carve-out and manual actions under the site reputation policy will affect results as they always have. The change is defined by where the searcher is, not where the site is hosted, so a UK business with EEA customers may see the same penalised section behave differently across markets.

Was the August spam update about AI content?

Partly, according to practitioner reports, and Google’s recently published Scalable Cluster Termination System research on terminating networks of mass-generated AI spam supports that reading. Google has confirmed no new policy types with this rollout though, so treat it as strong evidence rather than proof. The safer framing for clients is that content built primarily to rank is the risk, whether or not AI produced it.

Will goto redirects break our rank tracking?

Possibly, at least temporarily. Google now routes result clicks through a google.com/goto redirect, which makes result URLs harder for third-party tools to parse, and one measurement firm reports close to full rollout across residential IPs. If tracking data looks strange in the coming weeks, contact your provider before concluding rankings have moved.

Where should B2B brands focus to get cited by AI?

On third-party validation rather than owned content. In the Trendos analysis of IT and solutions services, brand-owned sources accounted for 2% of leading citations while independent editorial and reference sources took 47%, much of it review directories. Claim the two or three directories your buyers actually shortlist from and complete every field, matching your category wording to how prospects phrase questions to AI.

Should we implement WebMCP?

Only if your site is something people operate rather than read. WebMCP lets agents perform structured actions on your pages, such as editing files or updating a cart, and OpenAI states plainly that it has no connection to rankings, citations or recommendations. It is a usability investment for tool-like sites, not a visibility tactic, and anyone selling it as the latter should be asked for evidence.

Conclusions

This week put a border through the middle of Google’s enforcement, put a number on the spam update’s damage, and confirmed a redirect that will keep measurement vendors busy for a month. Underneath the news, the same instruction keeps repeating in different accents: the things that decide your visibility increasingly sit outside your own website, in directories, communities, feeds and inventories you influence rather than own. So check whether your reporting can tell an EEA searcher from a UK one, get the September ad decisions made before peak, complete the directory profiles nobody has touched since 2023, and keep asking what each shiny new standard actually changes. Unglamorous work, reliably, is what separates the teams who explain the quarter from the teams who are surprised by it.

Need help adapting your search strategy for the AI era? Contact the Anicca team for expert SEO and PPC guidance.

Table of Contents

SEO & Google Search Updates

AI Search & Visibility

PPC & Paid Media

Wrap-Up