Local domains gained as Google’s May Core Update reset search intent

Google’s May 2026 Core Update did not simply lift “local sites” because they used country-code domains or localized copy. The more useful reading is sharper: Google appears to have rewarded pages and domains that were a better match for the user’s market, task, and expected result type. In the UK, that meant local retail and marketplace domains gained ground while some global .com versions lost visibility for the same broad commercial spaces. In reference, health, travel, jobs, forums, and ecommerce, the same pattern kept reappearing in different forms: the page that looked like the right destination for the query gained an advantage over the page that looked adjacent to the query.

The May update rewarded local fit, not local decoration

The official facts are narrow. Google’s Search Status Dashboard says the May 2026 core update began on May 21, 2026 at 08:40 PDT and ended on June 2, 2026 at 05:40 PDT, with completion posted shortly afterward. Google framed it as a regular core update and did not publish a separate update-specific explanation.

The market signal came from third-party data. Aleyda Solis analyzed SISTRIX visibility shifts in the United States and United Kingdom from May 26 to June 2, 2026. Her analysis found that the clearest gains often went to domains that were a stronger fit for the dominant intent, market, and expected result format of the query set. The UK ecommerce pattern was especially direct: local-market entities gained while some global .com domains lost ground in the UK index.

That finding matters because many site owners still read core updates through old categories: authority won, forums lost, aggregators lost, brands won, AI content lost, publishers were hit, ecommerce recovered. The May update resists those shortcuts. Some authoritative domains lost. Some marketplaces gained. Some user-generated sites declined. Some social and visual platforms held or rose. Some local domains benefited. The dividing line was not a single site type. The dividing line was whether Google’s systems seemed to prefer that source as the destination for that query in that market.

For SEO teams, publishers, retailers, marketplaces, SaaS brands, and local businesses, the practical implication is blunt. Ranking recovery after May cannot begin with a generic content-quality checklist. It has to start with the changed result set. If Google now prefers a local merchant page, a canonical source, a task-completion marketplace, or a patient-friendly medical explainer, a strong but mismatched page will struggle. The page may be accurate. The domain may be powerful. The content may be useful. None of that guarantees that it is still the right result type.

The known facts around the rollout

The May 2026 Core Update was the second Search core update Google confirmed in 2026, following the March 2026 Core Update. It also came after the March 2026 spam update and the February 2026 Discover update, making the first half of 2026 unusually dense for search teams trying to isolate causes of movement. Google’s ranking history page lists the May core update as lasting 11 days and 21 hours, just under the March core update’s 12 days and 4 hours.

Google’s own description was restrained. The Search Status Dashboard recorded that the update was released and that the rollout could take up to two weeks. Search Engine Land reported Google’s LinkedIn wording: the update was “regular” and meant to surface relevant, satisfying content from all types of sites. That wording is useful mostly because of what it does not say. It does not name a sector. It does not announce a new spam policy. It does not say local domains were targeted for gains. It does not state that forums, affiliates, AI content, or publishers were targeted for losses.

Google’s standing core update guidance says these updates are broad changes to its ranking systems, not actions against specific sites or individual pages. The company’s core update documentation also tells site owners to confirm that a rollout has finished, wait at least one full week, and then compare the week after completion with the week before the rollout began. For May, that means the first cleaner Google Search Console comparison window starts around June 9, 2026, not during the May 21 to June 2 rollout itself.

That waiting period is not editorial caution for its own sake. Core update movement often arrives in waves. Search Engine Journal reported that third-party tracking tools showed elevated volatility at several points during the rollout, and Search Engine Land noted movement around the first weekend, the second weekend, and the final day before completion. A domain that moved on May 23 may not have moved for the same reason as a domain that moved on June 2.

The cleanest verified statement is this: Google completed a broad core update on June 2, and third-party visibility data from the final rollout window shows a strong pattern of market and intent fit. Everything beyond that is analysis. Strong analysis can still be useful, but it must remain tied to the data window, the measured markets, and the limits of visibility tools.

The dataset behind the local-domain finding

The local-domain story comes mainly from SISTRIX visibility data interpreted by Aleyda Solis, not from an official Google announcement. That distinction matters. SISTRIX visibility is not a site’s traffic, revenue, Search Console click count, Discover reach, or AI Overview citation count. It is a third-party index designed to estimate organic visibility across monitored keyword sets. It is useful for seeing market-level patterns, but it does not prove what happened to every query, every country, or every site.

Solis measured domain-level shifts in the US and UK between May 26 and June 2, the final stretch of the rollout. Her central reading was that visibility shifted toward the source type that looked like a stronger fit for the dominant intent, user market, and result format. She called the pattern an intent-destination reset. That phrase is useful because it avoids a narrow “local SEO” interpretation. The gain was not merely about being local. It was about being the destination Google appeared to want for the user’s job.

The UK ecommerce findings carried the strongest local-market signal. Search Engine Journal’s coverage of the analysis reported that local retailers gained while the .com version of the same brand fell in the UK index. It cited the example of amazon.co.uk rising 21.3% while amazon.com fell 54.6% for UK users, with those same .com domains holding roughly flat in the US index.

A single example should not be turned into a universal rule. Amazon is not a normal retailer, and SISTRIX visibility shifts do not map one-to-one to sales. Still, the direction is commercially meaningful. If Google’s UK result set becomes less tolerant of wrong-market pages, then global sites that rely on a powerful .com domain to rank across markets face a tougher standard. The local page must be more than a translated copy. It has to prove that it is the right result for a UK searcher through inventory, currency, shipping, returns, compliance language, local customer support, entity recognition, internal linking, and technical localization.

SISTRIX’s own update post showed rising UK radar values as the rollout progressed and noted candidate domains appearing in the data from May 23 onward. It also observed dictionary-related movement early in the rollout and later listed UK winners and losers by percentage and absolute change. That adds confidence that May produced broad visible movement, while still leaving the cause open to interpretation.

The dataset therefore supports a careful conclusion: in the measured US and UK SISTRIX data, the May 2026 Core Update favored better-aligned destination types, with local-market ecommerce domains gaining in the UK against some global .com alternatives. It does not support a lazy conclusion that every local domain won, every global domain lost, or every site should migrate to a country-code top-level domain.

Market fit became the ranking story

Market fit is not a new SEO concept. International SEO has always involved country targeting, language targeting, currency, delivery rules, local trust cues, hreflang, canonicals, and country-specific content. The May 2026 pattern made the concept feel less like a technical hygiene item and more like a ranking interpretation problem. A page can be technically accessible in a country and still feel like the wrong market result.

For a UK user searching commercial terms, a global .com page may answer the keyword but fail the market. It may show prices in dollars, lack UK delivery terms, list unavailable inventory, bury local returns information, or use brand architecture that signals the wrong entity. Google can often crawl it, index it, and understand its content. The harder question is whether it is the best result to show in that country when a local version exists.

Google’s localized-version documentation says hreflang helps Google understand alternate language or locale versions of a page, and it lists HTML, HTTP headers, and sitemaps as accepted methods. It also says hreflang and the HTML lang attribute are not used to detect a page’s language; Google uses algorithms for that. This is a useful reminder that hreflang is a signal of relationship, not a substitute for local relevance.

The May update seems to have punished a common weakness in international sites: treating localization as duplication management rather than market proof. Hreflang tags can connect UK, US, Canadian, Australian, and Irish versions. Canonicals can reduce URL confusion. Local folders or ccTLDs can create a cleaner site architecture. None of these elements by themselves answer whether the UK page is genuinely better for a UK user.

That is why the SISTRIX pattern is so revealing. The gain for local domains in the UK did not occur in isolation. It appeared beside reference shifts, forum pullbacks, marketplace gains, and health-format sorting. The shared pattern was destination fit. For ecommerce, the right destination often means local product pages, local category pages, local availability, local checkout confidence, and local service promises. For jobs, it means active listings and application flow. For travel, it means booking or comparison capability. For health, it means a source type that fits the query’s risk and user need.

Market fit is also a brand problem. Google’s systems have to decide which entity the user likely expects. A user searching from the UK may expect the UK entity even if the global entity has more backlinks, more historical authority, and more content. The update suggests that global authority can lose to local appropriateness when the query carries a market expectation.

Local domains gained because the user market was visible

Local domains did not gain because geography became a magic ranking factor. They gained where the user market became clear enough that a local entity looked like the safer, cleaner, or more complete answer. That distinction protects teams from superficial fixes. Buying a local domain, adding “UK” to titles, or spinning regional landing pages will not replicate the advantage if the underlying experience still feels global, generic, or thin.

The local-market signal has at least four layers.

The first layer is technical clarity. Google needs to find the right version, crawl it, index it, and understand its relationship to alternates. Hreflang, canonical consistency, internal links, XML sitemaps, and localized URL patterns sit here. Google’s canonical documentation says redirects and rel=”canonical” annotations are strong signals, while sitemap inclusion is a weaker signal, and it warns that hreflang implementations should specify a canonical page in the same language where possible.

The second layer is content-market fit. A UK category page that uses UK sizing, VAT-inclusive pricing, local delivery thresholds, UK stock status, UK returns, and UK store information is more convincing than a copied US category page with a currency switcher. For service businesses, content-market fit includes local regulations, local terminology, local proof, and local customer objections. For publishers, it includes local reporting, local editorial context, and local audience assumptions.

The third layer is entity clarity. A global brand with local entities has to make those entities visible. Structured organization data may help where appropriate, but entity clarity is broader than schema. It includes footer information, About pages, store locators, legal pages, press mentions, Google Business Profile relationships, local social profiles, and consistent naming across the web. Google’s guidance for generative AI search also points site owners toward product feeds and Google Business Profiles where local business or product information is relevant.

The fourth layer is task completion. A user who lands on the local page must be able to do the thing the query implies. Searchers do not only need a page in their language. They need to buy, book, compare, apply, confirm opening hours, check delivery, read trusted local guidance, or complete research without hitting a market mismatch. The strongest local domains in a core update are usually not “localized content” sites. They are local task destinations.

This is the reason local gains are so difficult for global sites to copy quickly. A local domain advantage is often the visible edge of years of operational investment: warehouse coverage, payment methods, customer service, returns infrastructure, product feed quality, store networks, editorial staffing, and brand recognition in the market. SEO can expose or clarify that advantage. SEO cannot invent it overnight.

Authority was not enough on its own

The May data undermined a comforting old SEO belief: if a domain is authoritative enough, it can absorb almost any intent mismatch. Solis’s analysis found that several highly authoritative domains lost ground in the measured window, including major publishers, scientific publishers, government and institutional sources, and health authorities. Search Engine Journal’s summary noted that authority alone did not explain the winners.

This does not mean authority stopped mattering. It means authority looked conditional. A highly trusted page can lose if Google decides that a different source type better satisfies the query. A global medical authority may be the best source for an official disease definition but not for a patient-friendly symptom explainer, a drug price query, a local clinic query, or a treatment comparison written for lay users. A global newspaper may be an authoritative source for reported news but not the right answer for a persistent evergreen query where a specialist reference page or official source is cleaner.

That distinction is central to the local-domain pattern. A global .com can have authority, links, brand demand, and strong content. The local domain can still win when the query implies local fulfillment. Search engines are not only ranking “the best domain.” They are ranking a result for a job. A powerful domain can be too broad, too global, too indirect, or too far from the task.

This is especially relevant for enterprise SEO. Large brands often judge pages through internal authority metrics: domain authority, link equity, historic rankings, traffic, word count, template quality, and content freshness. Those metrics may still matter, but they do not answer the source-type question. If the SERP after May is dominated by local merchants, official pages, booking engines, specialist tools, or community visual platforms, then the affected page has to be compared against that new result mix.

Authority also becomes harder to assess in AI Search. Google’s AI features documentation says AI Overviews and AI Mode may use query fan-out, generating related searches across subtopics and data sources to form responses and show links. A domain may be authoritative for the original keyword but less competitive across the fan-out query space.

That expands the definition of fit. The page has to match the primary query and the likely subquestions around it. For local ecommerce, that can mean shipping, returns, size guides, availability, alternatives, reviews, and product comparison. For travel, it can mean dates, location, transport, cancellation rules, and attractions. For health, it can mean symptoms, diagnosis, treatment, risk, and when to seek care. The more Google interprets queries as tasks, the less raw authority can cover missing context.

Ecommerce showed the sharpest localization signal

Ecommerce was the clearest place where local domains gained visible ground. The UK examples reported from the Solis analysis point to a practical rule for international retailers: the page that ranks should look like the page that can sell, ship, support, and legally serve the user in that country. A global informational product page is not enough when the query carries purchase intent.

Retail localization has many moving parts. Currency and price formatting are the obvious ones, but they are not the whole job. Users care about stock, delivery dates, shipping cost, local returns, warranties, store pickup, payment methods, consumer rights, reviews from their market, customer support hours, and whether the product model is the same in their country. A page can rank before a core update despite weaknesses in these areas. After a recalibration, those weaknesses may matter more because a local competitor or local version offers a cleaner task path.

The Amazon example is useful because it shows the mirror effect. If amazon.co.uk rises for UK visibility while amazon.com falls there, and the .com stays roughly flat in the US, that does not look like a generic punishment of Amazon. It looks like a market-routing correction.

The same logic applies to fashion, electronics, home goods, beauty, grocery, furniture, automotive parts, and B2B ecommerce. A UK buyer searching for a product category often expects UK availability and service, even if they do not type “UK.” Google may infer market intent from location, language, query history, product type, legal context, and result interactions. A global page that once benefited from authority may lose if a local page better matches that implied market.

For retailers, the fix is not to create thinner local copies. Thin localization can make the problem worse. The better response is to audit market evidence page by page. Does the category have local filters? Are unavailable products suppressed or explained? Are delivery and returns visible before checkout? Are reviews market-specific where possible? Are local stores, local stock, and local service pages internally linked? Is the product feed aligned with the local website? Does the local page use the right canonical and hreflang signals?

Google’s generative AI guidance also notes that product listings, product information, and local business information can appear in generative AI responses, and it points merchants toward Merchant Center and Google Business Profiles where relevant. This matters because ecommerce visibility is now split across classic organic results, merchant surfaces, AI responses, local packs, images, and brand results.

The May update should push retailers away from one-domain global efficiency as the default mental model. Efficiency still matters. Duplicate content still matters. Crawl management still matters. But market-specific usefulness now looks like a ranking asset, not an operational luxury.

Reference sites exposed the source-type reset

Reference results showed that the May update was not only about local commerce. Solis’s analysis found sharp movement within reference and language-related domains. Canonical reference brands gained in aggregate, while pronunciation utilities, language Q&A sites, and dictionary aggregators were more exposed. The UK examples included cambridge.org up 40.9%, thesaurus.com up 39.7%, and merriam-webster.com up 33.3%, while youglish.com, forvo.com, hinative.com, onelook.com, and wordreference.com fell in the UK index.

The important lesson is not that dictionaries won. It is that a query class can contain several result types that appear similar from a keyword view but feel different to a ranking system. A definition page, a thesaurus page, a pronunciation tool, a translation forum, a Q&A thread, and a dictionary aggregator may all target overlapping terms. They do not serve the same need.

For years, SEO analysis often grouped sites by vertical. A “reference” SERP would be treated as one category. May suggests that Google may be sorting more finely. If the dominant intent is a canonical definition, a canonical dictionary gains. If the dominant intent is pronunciation, a tool might still be useful, but only if Google believes that is the expected result type. If the dominant intent is language learning from native speakers, a Q&A or forum page may fit. If the query is navigational or brand-led, the winner changes again.

That source-type reset helps explain why authority did not fully predict movement. Cambridge, Merriam-Webster, and other reference brands have strong authority, but the gain is best understood in context: they were also cleaner canonical destinations for many reference intents. A forum or aggregator might be useful, but it may look like a layer removed from the source when the user needs a direct answer.

This same mechanism extends beyond dictionaries. In software, an official documentation page may beat a tutorial aggregator for API syntax, while a practitioner tutorial may beat documentation for “real-world setup.” In finance, an official tax authority may beat a blog for compliance, while a calculator may beat both for a user trying to estimate a payment. In travel, an official attraction page may win for opening hours, while a booking marketplace wins for accommodation. The right source type changes with the implied task.

For content teams, the reference-site pattern is a warning against building pages around keywords alone. A page targeting a term has to declare its role. Is it the canonical definition? The comparison? The calculator? The tool? The marketplace? The local service page? The firsthand review? The official policy? The expert explanation? Pages that blur these roles can be vulnerable when Google tightens result-type selection.

Forums lost share while visual and social surfaces split

The May update also complicated the debate around user-generated content. Solis’s analysis found that forum, Q&A, and open-publishing surfaces declined in aggregate in the measured UK and US data. Reddit, Quora, and StackExchange all fell in both markets. The declines were not equal, and Reddit’s percentage drop was smaller than some others, but its absolute movement was large because it occupies so much visibility.

That matters because forums had gained unusual SEO prominence during the 2023 and 2024 period, when Google emphasized “hidden gems” and first-hand experience. Many searchers also added “Reddit” to queries because ordinary search results felt overproduced or affiliate-heavy. The May pullback suggests that Google may be recalibrating where forum content belongs. It does not mean forum content has no value. It means forum pages may be less likely to serve as broad default results when a better source type exists.

The more interesting detail is that this was not a universal user-generated-content decline. Solis reported that large social, video, and visual platforms were mixed to positive overall, with YouTube, X, and Facebook gaining in the US and Pinterest and Fandom gaining in both markets.

The distinction is format. A forum thread can be the right destination for lived experience, troubleshooting, opinions, unusual product feedback, or community comparison. It is less clean for queries that need canonical facts, fresh instructions, local availability, official rules, or a task flow. A visual platform can fit a style, design, recipe, craft, fandom, or image-led exploration query. A video platform can fit demonstrations, reviews, explainers, and walkthroughs. A social platform can fit real-time commentary or creator-led material.

For publishers and brands, the forum pullback should not be read as permission to ignore first-hand content. It is better read as a signal that first-hand content needs clearer packaging. A page with lived experience must still be edited, structured, attributed, and connected to the user task. Raw conversation is sometimes the best result. Often it is not.

For forum operators, the update raises deeper questions. Are thread pages indexable by default even when they contain shallow replies, outdated advice, duplicate questions, or thin content? Are canonical answers surfaced? Are old threads updated or marked as stale? Are high-quality contributors visible? Are category pages useful entry points? Are internal links helping Google find strong discussions rather than endless near-duplicates?

The May pattern does not kill community SEO. It raises the proof burden on community pages that rank for broad informational and commercial queries. They need to be the best answer format, not merely a relief from overpolished web content.

Jobs and travel proved aggregators were not the enemy

The May update did not produce a simple anti-aggregator story. In jobs and travel, category-defining transactional marketplaces gained in both markets, while more informational or utility-style aggregators were more exposed. Solis’s analysis found aggregated gains of 21% in the UK and 17% in the US for category-defining transactional marketplaces, with examples including Trip.com, Skyscanner, ZipRecruiter, Expedia, Glassdoor, Indeed, and Booking.com.

That result is important because SEO commentary often treats aggregators as a single group. In practice, users experience aggregators differently. Some aggregators are thin layers that collect snippets, repackage information, or interpose themselves between the user and the task. Others are primary destinations where users can compare options, filter choices, check availability, apply, book, save, review, or complete a transaction. Google appears to have favored the second type in May.

Jobs and travel are ideal test cases because the task is obvious. A user searching for flights, hotels, jobs, salaries, or travel accommodation often wants comparison and action. A marketplace or metasearch platform can be more useful than a single supplier or a generic article. The platform is not “thin” just because it aggregates. It becomes strong when it solves the task better than any one source.

That is the lesson affiliates and comparison sites should study. The question is not whether the site aggregates. The question is whether it earns its place as a destination. Does it have real filters? Fresh data? Transparent methodology? Original testing? Local availability? Clear trade-offs? User tools? Expert review? Price history? Direct task completion? Or does it merely summarize what other pages already say?

The May update appears less hostile to aggregation than to derivative layers. A derivative layer borrows the shape of usefulness without owning the data, the experience, the expertise, or the transaction. It may rank during periods when Google rewards breadth, links, or freshness. It is vulnerable when Google reweights destination fit.

Travel also shows why local and market signals are not limited to country domains. A travel marketplace may serve many countries on one domain, but it can still fit local intent if it handles local currency, language, availability, routes, regulations, and user expectations. A local domain can lose if it lacks task depth. A global marketplace can gain if it is the best task engine for the query.

The strategic implication is clear. Aggregators need to decide whether they are destination businesses or SEO wrappers. The former can survive core updates. The latter increasingly depend on gaps in Google’s current result selection.

Health results separated trust from format

Health and medical visibility after May showed another layer of source-type sorting. Solis reported that trusted health destinations generally held or rose, with WebMD up in both the UK and US and Cleveland Clinic, Healthline, and NHS broadly stable to positive. At the same time, GoodRx and UbieHealth fell sharply, and even authority sources such as WHO and NIH slipped in the measured data.

This is the clearest warning against reducing core updates to E-E-A-T language alone. Experience, expertise, authoritativeness, and trust still matter, especially in health. But high trust does not make all health pages interchangeable. A drug-price tool, symptom checker, hospital explainer, government guideline, research paper, emergency advice page, and local clinic page meet different needs. The user’s query determines the safer format.

A global health authority may be the best source for public health guidance, disease surveillance, and official policy. It may not be the most accessible answer for a patient searching symptoms in plain language. A drug-price utility may be useful in one market and less useful or less relevant in another. A symptom checker can be attractive for broad queries but risky if Google prefers a more authoritative explainer or clinical source.

This matters for local domains because health is both local and high-stakes. Searchers often need locally relevant care pathways, national health-system advice, insurance context, emergency numbers, regulatory rules, drug names, appointment availability, or clinic locations. A global medical page may rank for general knowledge. It may lose to a local public health page or hospital page where the query suggests local action.

Google’s helpful-content documentation asks creators to focus on people-first, reliable content and to assess whether content demonstrates experience and expertise. That guidance is stable, but the May health split shows that content quality must be judged against the current result type.

For health publishers, the practical audit starts with affected query groups. Did losses happen on symptom pages, drug pages, condition explainers, tool pages, doctor pages, or news pages? Did Google replace the site with official sources, hospitals, local health services, videos, forums, or AI Overview links? Did the query require freshness, local context, medical review, or a different reading level?

Trust is necessary in health search, but it is not enough when the page format is wrong. May made that visible by allowing trusted sites to move in opposite directions depending on the query class.

AI Search timing made the update harder to read

The May core update landed within days of Google announcing major AI Search changes at I/O. Google’s May 19 blog post described new AI features, agentic capabilities, and a redesigned Search experience that it called the biggest upgrade to the Search box in more than 25 years. Search Engine Journal noted that the core update overlapped with I/O week, making it harder for site owners to isolate whether changes came from ranking systems, interface changes, AI features, or a mix of them.

This overlap matters for interpretation. Organic visibility tools measure rankings and SERP features in specific ways. Google Search Console reports clicks, impressions, CTR, and average position. AI Overviews, AI Mode, product surfaces, local packs, images, videos, and Top Stories can all affect user behavior. A site can lose clicks without losing classic rankings. A site can gain visibility but lose traffic if AI answers satisfy the query. A site can be cited in AI features while receiving fewer standard organic clicks.

Google’s AI features documentation says AI Overviews and AI Mode surface relevant links and may use query fan-out, issuing multiple related searches across subtopics and data sources. It also says sites appearing in AI features are included in Search Console’s performance report under the Web search type.

That creates a measurement problem. If a page’s impressions rise because it appears in an AI feature, its CTR may fall because fewer people click. If a page disappears from a classic blue-link position but appears as a supporting AI link, its traffic impact may vary by query. If AI Mode uses longer conversational queries, pages built around short keyword patterns may lose relevance even without a classic ranking penalty.

Academic measurement work is already showing that AI search does not behave exactly like classic search. A 2026 arXiv study of Google AI Overviews reported that cited domains can differ from co-displayed first-page results, suggesting that AI source selection is not a simple copy of the organic top ten. Another study comparing Google Search, Gemini, and AI Overviews found substantial differences in retrieved sources across systems.

The timing does not mean AI Search caused the local-domain gains. It means the update happened inside a Search product that is becoming more task-based, conversational, and source-selective. The same logic that favors local task destinations in classic results may also matter in AI Search, but measurement will be messier.

Local domain gains do not mean ccTLDs are magic

Country-code top-level domains can help users and search engines understand market targeting, but the May update should not be reduced to a ccTLD ranking boost. A local ccTLD can still lose if it offers weak content, wrong canonicals, poor crawl paths, thin category pages, generic advice, missing product data, or an unconvincing local entity. A global .com can still win if it offers the strongest market-specific experience through subdirectories, subdomains, local feeds, local service pages, and strong internal localization.

Google’s international documentation explicitly allows alternate URLs to be on different domains, and it treats HTML, HTTP headers, and sitemaps as equivalent ways to indicate localized versions through hreflang. That means the architecture can vary. What matters is that Google can understand the relationship and that the user gets the right version.

The ccTLD conversation often distracts from harder questions. Does the local version have enough unique market value to deserve ranking? Are local inventory and pricing rendered in crawlable form? Are products that cannot be shipped to the user’s market excluded from indexable paths? Are legal pages and returns information localized? Are internal links from the global site pointing users and crawlers to the correct country version? Are backlinks and brand mentions split across domains in a way that weakens the local entity?

A ccTLD may make the market signal more obvious, but obvious is not the same as sufficient. Search engines can understand local subdirectories such as /uk/, local subdomains such as uk.example.com, and local domains such as example.co.uk. Each setup has trade-offs. A ccTLD may help local brand trust but fragment authority and operational work. A subdirectory may consolidate authority but require stronger signals to prevent wrong-market ranking. A subdomain sits between those models.

The May lesson is not “move to .co.uk.” It is “make the UK result unmistakably better for UK users.” If a global .com ranks where a local version should, that may be an opportunity for canonical, hreflang, internal-linking, and content-market improvements. If the local version ranks but loses, the issue may be deeper: competitors may serve the task better.

Local domain gains were evidence of market-fit preference, not proof of a domain-extension shortcut. The difference is expensive. Teams that chase the shortcut will create migration risk. Teams that fix market fit will build a more durable advantage.

Hreflang, canonicals and entity clarity became operational risks

International SEO problems often look harmless until a core update or ranking recalibration makes them visible. A site may have imperfect hreflang for years and still rank well because authority, links, and content breadth compensate. Then a core update changes the weighting of source fit, and the same imperfections become costly.

Hreflang errors are common: missing self-references, non-reciprocal tags, wrong country codes, mixed HTTP and HTTPS URLs, canonical tags pointing to another locale, old URLs left in sitemaps, and local pages blocked from crawling. Google says if two pages do not both point to each other with alternate tags, the tags will be ignored. That is a simple rule with large consequences for global sites.

Canonicals create another risk. Google’s canonical documentation says rel=”canonical” is a strong signal, but not an absolute directive. If a UK page canonicals to a US page, or if near-duplicate local pages all point to one global canonical, the site may be telling Google that the global URL is the preferred version. That can conflict with the market-fit goal.

Entity clarity sits above both. Search engines need to understand whether the local domain, subdomain, or folder represents a distinct market entity, a duplicate page set, a translated site, a franchise, a distributor, or the same global business. Entity ambiguity can show up in branded queries, knowledge panels, local packs, product results, and organic rankings. It can also affect user trust: a UK customer who lands on a US page with unclear delivery information may bounce back, sending poor satisfaction signals at scale.

The May local-domain pattern should push international brands to connect technical and brand governance. SEO teams cannot fix all localization issues in tags. Merchandising, legal, logistics, CRM, analytics, engineering, customer support, and content teams all own pieces of market clarity. The page that Google ranks is a public expression of that internal alignment.

A useful audit asks three questions for every affected country and category. Does Google crawl and index the correct local page? Does the page itself prove local usefulness? Does the broader web confirm that this local entity is trusted in that market? If any answer is weak, the local version may lose even when it technically exists.

The May update turned international SEO from a back-office correctness task into a competitive ranking task. The sites that treat hreflang and canonicals as clerical work will miss the broader issue. The sites that treat them as part of market identity will be better prepared for the next recalibration.

The wrong-market ranking problem

Wrong-market ranking is one of the most damaging international SEO problems because it can look like success in rank tracking and failure in business performance. A US page ranking in the UK may generate impressions and even clicks, but users often encounter the wrong currency, wrong shipping rules, wrong product availability, wrong legal context, or wrong customer support. The page ranks, but it does not satisfy the market.

The May data suggests Google became less forgiving of this problem in some UK ecommerce SERPs. When the local domain gained and the global .com lost, the system looked less interested in the strongest generic domain and more interested in routing users to the correct market result.

Wrong-market ranking can emerge from many causes. The global page may have more links. The local page may be newer, thinner, slower, or poorly linked. The hreflang cluster may be incomplete. The canonical may point to the wrong version. The local page may be blocked by faceted navigation rules. The product feed may use a different URL than the organic page. The site may redirect users based on IP in a way that confuses crawlers. The local page may be too similar to the global page to prove distinct value.

The fix starts with data segmentation. In Search Console, compare affected pages by country, query, page, device, and search type. Look for pages receiving impressions from countries they should not serve. Look for local pages with low impressions despite being the intended destination. Look for queries where two country versions compete. Look for canonicalized pages that still receive traffic. Look for local pages ranking for brand terms but not non-brand commercial terms.

Then compare SERPs manually from the market. If Google is now showing local retailers, local marketplaces, local official pages, or local review sources, a global page cannot be evaluated in isolation. The competitor set has changed. Your local page must be judged against local SERP expectations, not against the global template.

There is also a user-experience dimension. Wrong-market pages create friction that analytics may hide. Users may leave before checkout. They may search the brand again with “UK,” “Ireland,” “near me,” or “shipping.” They may click another result. They may convert through paid search later, making organic performance look less damaged than it is.

Wrong-market ranking is not only an SEO routing problem. It is a trust problem. The May update appears to have exposed that trust gap more clearly in some commercial result sets.

Query intent moved from keyword match to destination choice

The phrase “search intent” has become so common in SEO that it often loses meaning. Many teams reduce it to four buckets: informational, navigational, commercial, transactional. The May update shows why those buckets are too blunt. A query’s intent is not only about whether the user wants information or a transaction. It is about the destination type that best completes the job.

A query like “running shoes” can imply ecommerce category browsing, local store discovery, brand navigation, product comparison, image exploration, videos, reviews, or AI-assisted buying advice. A query like “flu symptoms” can imply a quick definition, a health-system page, a pediatric page, a symptom checker, local urgent care, or official public health guidance. A query like “software engineer jobs London” has location, task, freshness, marketplace, salary, employer, and application intent embedded in it.

The May findings make more sense when intent is treated as destination choice. In reference, canonical sources gained where the expected destination was a direct reference. In forums, Q&A declined where raw discussion may no longer have been the expected default. In jobs and travel, task-completion marketplaces gained. In UK ecommerce, local-market domains gained. In health, trusted but format-appropriate destinations held or rose.

Google’s AI feature documentation reinforces this broader view by describing query fan-out in AI Overviews and AI Mode. If Google’s systems generate multiple related searches to support an answer, the page is not judged only against the literal query. It may be judged against the cluster of subneeds behind that query.

For SEO strategy, the old page-level question “does this page target the keyword?” is no longer enough. The better question is “does this page deserve to be the destination for this user’s job in this market?” That question forces a different audit. It looks at format, evidence, data freshness, market specifics, task flow, author credibility, local proof, media, tools, comparisons, and user satisfaction.

This is also why local domains gained. Locality is one form of destination fit. If the user’s job includes buying in the UK, booking in the UK, reading UK advice, comparing UK services, or following UK rules, the local result becomes the better destination. If locality is irrelevant, a local domain has no automatic advantage.

The May update rewarded destination logic. Sites built around keyword coverage but weak destination identity are exposed. Sites that own a clear role in the user’s task are stronger.

Observed May 2026 visibility patterns

Compact table of source-type shifts

This table compresses the main third-party visibility findings into an operational reading. The repeated lesson is not that one sector won or lost. The repeated lesson is that Google appeared to sharpen its preference for the destination type that best matched intent, market, and task.

Publisher strategy after the local-market shift

News and editorial publishers should not read the local-domain gains as an ecommerce-only story. Local market fit affects publishers too, especially in news, finance, health, travel, culture, sport, public policy, and consumer advice. A global publisher can be authoritative, but a local publisher may be the better result for queries with local law, local politics, local prices, local institutions, local events, or local public impact.

Google’s May 27 blog post about original and quality content highlighted Preferred Sources in AI Search, a new carousel for timely articles and perspectives, and labels such as Highly Cited. The same post said users had selected more than 345,000 unique sources and that people were twice as likely to click through to a Preferred Source.

That is not part of the May core update itself, but it changes the environment in which publishers compete. If users can signal preferred sources in Top Stories, AI Overviews, and AI Mode, publisher loyalty becomes a search visibility factor at the interface level. Google’s Preferred Sources documentation says selected publications are more likely to appear in Top Stories with a preferred badge, and that content can be highlighted in AI Mode and AI Overviews where those features are available.

For local and national publishers, this creates two parallel paths. The first is algorithmic: publish original reporting, local expertise, clear authorship, topical depth, and durable explainers that deserve to rank. The second is audience-based: build direct reader trust so people actively choose the publication as a preferred source where the feature is available.

The May local-domain pattern supports that second path indirectly. A publication with a clear market identity is easier for users and search systems to understand. A UK outlet covering UK consumer rights, UK politics, UK health policy, UK sport, or UK travel disruption has a stronger local role than a generic international explainer rewritten for search. A city publisher with verified local reporting has a stronger role for municipal issues than a national site summarizing the same story from afar.

For global publishers, the challenge is to avoid parachute content. A global explainer about a local regulation may rank for a time if the domain is strong. It is vulnerable when Google decides that a local legal, government, or specialist publisher better fits the user market. Original reporting and local expertise are not just editorial values. They are search-market assets.

Publishers should audit where they lost visibility after May by topic and market. Did local news pages replace evergreen explainers? Did official sources replace reported summaries? Did AI features reduce clicks on quick-answer articles? Did forums or social platforms lose share in topics where publishers can now regain ground with reported, structured, firsthand material? These are different problems and require different fixes.

Ecommerce strategy after UK domains gained ground

For ecommerce teams, May should trigger a market-by-market category audit, not a panic rewrite. The first question is whether losses happened on wrong-market pages, local pages, or both. A global .com losing UK visibility while the local domain gains is a routing and market-fit story. A local domain losing to another local retailer is a competitive usefulness story. A local category losing to marketplaces is a task-completion story.

The category page is often the battlefield. Product detail pages can rank for exact products and long-tail terms, but category pages carry broader commercial intent. A category page that lists products is no longer enough. It has to guide choice, expose relevant filters, show availability, answer local purchase concerns, link to buying guides where useful, and connect to internal search, reviews, and support.

Google’s ecommerce-related AI guidance suggests that product information and local business information can appear in AI responses and points to Merchant Center and Google Business Profiles. This creates a wider surface area for ecommerce visibility. A retailer may need consistency across organic pages, product feeds, merchant listings, local store data, images, and structured data.

After May, the local ecommerce audit should include:

Price and availability consistency across page, feed, structured data, and checkout. Local delivery promises visible on crawlable pages, not only in checkout modals. Country-specific returns and warranty details linked from product and category templates. Local reviews or at least review filtering by country where relevant. Internal links that prioritize the local version for local users. Hreflang and canonical signals checked against live indexed URLs. Product variants mapped correctly by market.

The strongest ecommerce response is not to inflate copy. It is to reduce market uncertainty. Many category pages contain long SEO text that few users read. That text may not solve the core problem. A UK user choosing a washing machine, running shoe, sofa, laptop, or skincare product needs filters, comparison cues, availability, delivery timing, returns confidence, and local proof. A few generic paragraphs at the bottom of the page will not fix a weak task path.

International brands also need to review how global campaigns interact with local search. A global product launch may create backlinks and brand demand for the .com page, causing it to outrank local versions temporarily. If May-style market routing becomes more common, brands will need launch plans that strengthen local URLs from the start.

The ecommerce winners after core updates are usually the pages that remove purchase uncertainty. In May, local market uncertainty appears to have become more expensive.

Affiliate and comparison sites face a harder proof burden

Affiliate and comparison sites sit close to the fault line exposed by May. Some are genuinely useful. They test products, explain trade-offs, maintain data, compare prices, track availability, interview experts, and help users make decisions. Others are derivative pages built to capture search demand between the user and the actual destination. The May update’s marketplace pattern suggests that Google is not against intermediaries. It is against weak intermediaries when a stronger destination exists.

The jobs and travel gains prove this point. Booking.com, Indeed, Expedia, Skyscanner, Trip.com, and similar platforms aggregate information, but they also let users act. They have data, filters, depth, inventory, and task flow. Thin affiliate sites often lack those assets. They describe choices but do not own them. They list “best” products but do not test them. They summarize features but do not resolve local availability. They cite reviews but do not add original evidence.

Google’s guidance for generative AI search warns against recycled content and commodity content. It says creators should avoid simply restating what is already available and should create content based on what they know about the topic. It also warns against creating many pages mainly to manipulate rankings or AI responses.

For affiliate teams, May should shift the strategy from page expansion to proof expansion. Add original photos, testing notes, price history, local availability checks, expert commentary, failure cases, methodology, comparison tools, user decision trees, and post-purchase outcomes. Remove or consolidate pages that repeat the same advice across near-identical query variants. Stop treating “best X for Y” as a template that can be scaled endlessly.

Locality makes this harder and more useful. A comparison of the best credit cards, broadband providers, mattresses, accountants, gyms, or travel insurance products is market-specific. If the page fails to handle local regulations, provider availability, currency, delivery, tax, eligibility, or consumer rights, it is not the best result for that market. A local official source, local marketplace, or local expert publisher may deserve the ranking instead.

The same applies to B2B comparison pages. Software availability, pricing, data residency, support hours, integrations, procurement rules, and legal compliance vary by market. A generic global comparison may lose to a local specialist if the query implies local buying risk.

Affiliate recovery after May is less about adding more words and more about proving that the page earns its place between the user and the destination.

News publishers need market identity and original reporting

The local-domain gain story intersects with a larger publisher problem: search is moving from links toward answers, summaries, preferred sources, and AI-mediated discovery. News publishers therefore need two forms of defensibility. They need content that Google’s ranking systems identify as the right result, and they need a brand relationship strong enough that users seek or select them directly.

Original reporting is central to both. Google’s May 27 blog post discussed features meant to highlight original and quality content, including Preferred Sources in AI Search and discovery aids for timely articles and perspectives.

For local publishers, this is an opening. Many broad publishers produce thin local summaries because they have authority and scale. A local newsroom can produce reporting that includes primary documents, named local sources, local history, public-meeting context, maps, timelines, service information, and follow-up coverage. That material is harder for a global or national site to copy convincingly.

For national publishers, market identity matters when covering regulated, local, or country-specific topics. A UK publisher covering UK tax, NHS policy, Ofcom rules, rail strikes, education changes, or supermarket pricing carries market relevance that a generic global article may lack. The same logic applies in the US, Canada, Australia, India, Germany, France, and smaller markets where local institutions shape the answer.

The May update also suggests that publishers should stop expecting authority to carry all evergreen content. A strong publication may rank for news about a health policy change but lose evergreen symptom queries to medical destinations. It may rank for reporting on a travel disruption but lose hotel booking queries to marketplaces. It may rank for analysis of a consumer issue but lose product category terms to retailers. That is healthy if the result type is better for users.

Publisher SEO should therefore separate roles. Reported news pages, explainers, live blogs, service journalism, investigations, reviews, local guides, data tools, and opinion pieces should not be judged by one template. Each has a different destination role. A core update may hit one role and reward another.

The publisher that wins after May is not the publisher with the most pages. It is the publisher with the clearest claim to being the right source for the query. That claim can come from locality, originality, expertise, data, audience loyalty, or task usefulness. It rarely comes from rewriting what everyone else already published.

Small local businesses should read the update differently

The phrase “local domains gained” can mislead small businesses. A local plumber, dentist, restaurant, lawyer, shop, clinic, or consultant is usually competing in local packs, maps, organic results, review ecosystems, and branded search. The May SISTRIX finding focused heavily on domain-level organic visibility in US and UK indexes, with a strong example in UK ecommerce. That is not the same as saying every local business website gained after the core update.

Still, the principle matters. Google’s systems appear to reward results that fit the user’s market and task. Small local businesses can benefit when their website and local presence make that fit obvious. That means clear service areas, local proof, real photos, named staff where appropriate, opening hours, pricing or quote expectations, booking paths, local reviews, Google Business Profile accuracy, and content that addresses local customer needs rather than generic national advice.

Google’s generative AI guidance explicitly mentions Google Business Profiles for local business visibility in AI responses and other Search results.

Small businesses should avoid the wrong lesson. They do not need thousands of location pages. They do not need AI-generated blog posts for every suburb. They do not need to chase national keyword volume if their business serves a small area. Scaled local content can become a quality risk, especially if it creates near-identical pages with swapped city names. Google’s generative AI content guidance warns that using AI tools to generate many pages without adding value may violate scaled content abuse policies.

A better local strategy is narrower and stronger. Build the pages that match real services and real locations. Show evidence of work. Answer questions customers ask before buying. Use local terminology. Mark up business information accurately where relevant. Keep technical basics clean. Earn local mentions and links from real relationships. Make the conversion path easy.

For service businesses, market fit includes proximity and trust. A national directory may rank for a broad query, but a local provider can be the better destination for a user ready to act. A small business site will not beat a marketplace on every query. It can win where the query implies local service, local credibility, and direct contact.

The May update’s lesson for small businesses is not scale. It is specificity. A small site with real local proof may be more useful than a broad site with generic advice and no operational connection to the searcher’s market.

Data discipline matters before any recovery plan

Core updates tempt teams into fast action. Traffic drops, dashboards turn red, executives ask for explanations, and agencies prepare recovery plans before the rollout is even finished. Google’s guidance argues against that reflex. Its core update documentation says to confirm completion, wait at least a full week, compare the right dates, review pages and queries, and separate small drops from large drops.

For the May update, the rollout ended on June 2. A clean comparison should use data after roughly June 9 compared with the week before May 21. Even then, teams should account for seasonality, news cycles, product demand, tracking changes, SERP feature shifts, and AI Overview behavior. Search Engine Journal reported that May showed multiple volatility points, which makes single-day comparisons especially risky.

Search Console analysis should begin with affected pages and queries, not with total organic clicks. A site-wide drop may hide a small number of high-traffic pages. A category-wide drop may hide wrong-market routing. A click drop may come from lower rankings, lower impressions, lower CTR, or changes in the result page. Google’s traffic-drop documentation says impressions and clicks patterns should be interpreted differently, and that small ranking fluctuations can happen without requiring radical changes.

The first recovery meeting after May should therefore answer these questions.

Which countries moved? Which search types moved? Which pages lost impressions? Which pages lost clicks while impressions stayed stable? Which queries lost average position? Which queries gained impressions but lost CTR? Which local versions gained or lost? Which competitors replaced us? Which result types appeared after the update? Which losses are brand, non-brand, product, category, article, guide, forum, or local?

Only then should teams propose changes. A page that fell from position two to four may not need rewriting. A page that fell from the top ten to page three across many queries needs a deeper review. Google’s traffic-drop documentation uses similar examples and says larger sustained drops call for whole-site and affected-page assessment.

The worst post-core-update mistake is treating every decline as a content problem. Some declines are market routing problems. Some are SERP feature problems. Some are demand problems. Some are technical problems. Some are genuine source-fit problems. The recovery plan depends on the diagnosis.

Recovery work begins with affected queries

A May recovery plan should begin at query level because the update appears to have changed which destination type Google preferred. Page-level traffic charts are useful for prioritization, but queries reveal the intent shift. If a page lost ten queries and held fifty, the lost queries may share a destination type that the page no longer satisfies.

Start by exporting Search Console data by page and query for the week before the update and the week after the waiting period. Segment by country. For international sites, this is not optional. The local-domain finding was market-specific. A page may hold in the US and drop in the UK, or the opposite. Mixed country data will hide the cause.

Next, classify lost queries by task. Do they imply buying, comparing, booking, applying, defining, troubleshooting, visiting, verifying, calculating, or reading news? Then examine the post-update SERP manually. Which result types replaced your page? Local domains? Official sources? Marketplaces? Forums? Videos? Visual platforms? Product grids? AI Overviews? News carousels?

If local domains replaced global pages, the recovery path is localization and routing. If marketplaces replaced affiliate pages, the recovery path is task depth. If official sources replaced explainers, the recovery path may be authority, citation, or role adjustment. If videos replaced text pages, the recovery path may include media. If forums declined and your publisher gained, the opportunity is to strengthen first-hand reporting or expert material.

This work is slow, but it prevents bad edits. Many sites respond to core updates by rewriting affected pages with longer introductions, more FAQs, more schema, and more generic “expert” language. If the result type changed, that work may not matter. A global page does not become a local destination because it adds 500 words. A weak affiliate page does not become a marketplace because it adds a comparison table. A forum thread does not become an official guide because it ranks for the same keyword.

The query is the evidence. The new SERP is the verdict. Recovery work should close the gap between the affected page and the result type Google now rewards.

Content pruning is the wrong first reflex

After core updates, many sites rush to delete, noindex, or consolidate content. Sometimes that is correct. Often it is premature. Google’s guidance on core updates and traffic drops does not recommend mass pruning as a first response. It recommends assessment, comparison, and improvement where needed. Google’s traffic-drop documentation also warns that changes can take days or months to show effect and that improved rankings are not guaranteed if other content is more deserving.

The May update makes pruning especially risky because losses may reflect destination mismatch rather than poor quality. A page may be useful but not for the query it lost. Deleting it could remove value for other users, internal links, topical support, long-tail traffic, or AI Search eligibility. A better decision may be to retarget the page, strengthen local cues, create a more suitable destination, or adjust internal linking.

Pruning makes sense when pages have no clear user purpose, duplicate the same intent, contain outdated or unsafe advice, exist only for keyword variations, or cannot be improved. It also makes sense when low-quality programmatic pages consume crawl attention and dilute site quality. But pruning should be based on evidence, not fear.

A useful content decision framework has four outcomes.

Keep and monitor pages that moved slightly or still satisfy their role. Improve pages that remain the right destination but are weaker than competitors. Consolidate pages that split the same intent across near-duplicates. Remove or noindex pages that have no useful purpose and no realistic improvement path.

For local domains, consolidation decisions must be market-sensitive. Do not consolidate UK and US pages just because they look similar if they serve different markets. Instead, make each version locally useful. Conversely, do not create separate pages for every city, county, or country if the content is identical and the business has no real local presence there.

Google’s guidance on generative AI content is relevant here because AI tools make scaled page creation easy. If a site created hundreds of thin local or AI-assisted pages before May, pruning or consolidation may be necessary. But the goal is not to reduce page count for its own sake. The goal is to align indexed pages with real user value.

A core update is not a command to delete. It is a reason to prove which pages still deserve to exist.

Technical SEO still decides which local version Google can trust

The May analysis focuses on intent and market fit, but technical SEO remains the gatekeeper. A local page cannot win if Google cannot crawl it, index it, trust its canonical, or understand its relationship to other versions. Technical clarity is the difference between having local content and having a local result Google can safely show.

International sites should check five areas after May.

First, crawlability. Local pages blocked by robots.txt, noindex tags, JavaScript rendering issues, CDN rules, or faceted-navigation traps cannot compete. Google’s AI features documentation says supporting links in AI Overviews or AI Mode must be indexed and eligible to appear in Google Search with a snippet.

Second, canonical consistency. The local page should not canonicalize to a global page unless that is truly the preferred version. Google treats canonicals as strong signals, and conflicting signals can lead to unwanted URL selection.

Third, hreflang clusters. Each language or locale version should reference itself and the other alternates. Broken reciprocity can cause Google to ignore the annotations. For large international sites, hreflang validation should be automated because manual checks miss template-level and sitemap-level drift.

Fourth, internal linking. Users and crawlers should reach the local version naturally from navigation, country selectors, category pages, related content, and product paths. If global navigation links mostly to the .com version, the site may be strengthening the wrong URL. Local domains need local link equity inside the site.

Fifth, structured data and feeds. Product, organization, local business, breadcrumb, review, and merchant data should match visible content. Google’s AI feature guidance says structured data should match visible text, and its generative AI guidance points merchants toward product and local business data channels.

Technical work alone will not create market fit. A perfectly tagged thin page is still weak. But technical errors can suppress a strong page or send Google to the wrong version. The May local-domain pattern makes these errors more costly because market routing appears to be part of the competitive story.

Technical SEO is not separate from content quality in international search. It is the infrastructure that lets the right content represent the right market.

Brand demand and preferred sources now matter beside rankings

The May core update happened in a search environment where user preference signals at the interface level are becoming more visible. Preferred Sources is the most direct example. Google’s documentation says users can select domain-level or subdomain-level publications as preferred sources, and those sources can receive a preferred badge in Top Stories, AI Mode, and AI Overviews where available.

For publishers and brands, this creates a second competition beside rankings. A site can earn algorithmic visibility by matching intent and market. It can also earn user-driven visibility when people choose it as a preferred source or search for it by name. This does not replace SEO. It raises the value of brand trust inside Search.

Local domains may benefit from this more than generic global sites because local identity is easier for users to feel. A reader may choose a local newspaper for city news, a national consumer publication for local regulations, a trusted health system for medical explainers, or a specialist retailer for product advice. A generic content site has a weaker claim to preference.

Brand demand also protects against volatility. A site that depends entirely on non-brand rankings is exposed to source-type resets. A site with direct demand, returning users, newsletter subscribers, app users, social followings, and preferred-source selections has more paths back to its audience. Google Discover, Top Stories, AI Overviews, AI Mode, and classic organic results may all change, but audience memory remains a defensible asset.

This does not mean brand solves poor search fit. A beloved publisher can still lose a query to an official source if the official source is the better result. A strong retailer can still lose to a local competitor if its page lacks local availability. Brand is a support beam, not a pass.

For local and niche brands, the practical work includes consistent naming, clear About pages, author and expert profiles, local proof, social and newsletter prompts, branded search encouragement, and content worth returning to. For publishers, Preferred Sources should be treated as part of audience development, not a technical SEO trick.

The May update rewarded destination fit. Preferred Sources rewards reader choice. The brands that combine both are harder to displace.

AI Overviews change the value of being the source

AI Overviews and AI Mode change the economics of search visibility because users may get synthesized answers before clicking. For some queries, being cited or linked inside an AI response may matter. For others, classic ranking may still drive the click. For quick facts, the click may shrink. For complex tasks, AI may create more exploratory paths and more links.

Google says AI Overviews and AI Mode rely on core Search ranking and quality systems, use AI techniques such as retrieval-augmented generation and query fan-out, and continue to depend on indexed, crawlable content. It also says there are no special technical requirements beyond eligibility for Google Search with a snippet.

Independent research suggests the picture is more complicated. A 2026 study on Google AI Overviews found that cited domains can differ from the first-page organic results and that a portion of claims were unsupported by cited pages. Another study found that generative search systems can have source-selection biases and different citation patterns from classic search.

For local domains, this matters because AI systems often need grounded, specific, task-ready facts. A local ecommerce page with clear delivery, stock, pricing, and returns may be easier to use as a source than a global page with vague availability. A local publisher with original reporting may be a better citation than a national summary. A local health system may be more relevant for country-specific care pathways than a global overview.

The value of being the source increases when AI interfaces summarize the web. Derivative pages become less useful because AI can summarize them too. Original data, firsthand experience, local facts, tools, images, videos, and primary reporting become more defensible. That is the same direction suggested by the May core update’s destination-fit pattern.

AI Search also changes query length. Search Engine Journal reported Google’s own AI Mode usage claims, including longer queries and more follow-up behavior, while noting the data was internal and not publicly verifiable. Longer queries can carry richer market and task signals.

The local-domain gain is a classic Search story, but it points toward an AI Search reality: sources with clear market-specific facts and task value are more likely to matter than generic pages built only to rank.

The commercial meaning of local visibility gains

For businesses, local visibility gains are not only SEO movement. They change acquisition cost, channel mix, and competitive pressure. If local domains gain organic visibility for commercial queries, paid search may become less dominant for those brands. If global .com pages lose local visibility, paid search, marketplaces, affiliates, and brand campaigns may have to cover the gap.

The biggest impact is often at category level. Product detail pages capture users who know what they want. Category and comparison queries capture users still choosing. If a local domain gains those queries, it captures demand earlier. That affects remarketing pools, email acquisition, conversion paths, store visits, and brand consideration.

For global retailers, a wrong-market organic loss may expose an operating model problem. Centralized SEO teams often manage global templates and content, while local teams manage merchandising and promotions. If the local page lacks enough authority or uniqueness to rank, the business may need to invest locally rather than ask the SEO team to rewrite metadata.

For local retailers, the gain is an opportunity but not a guarantee. Visibility gains must be converted. If pages are slow, filters are poor, stock is missing, returns are unclear, or checkout is weak, rankings will not become revenue. A core update can send more users. It cannot fix the business.

Marketplaces face a different commercial implication. If Google favors task-completion destinations, marketplaces with strong data and UX may gain leverage against affiliates and publishers. But they also become more accountable. Outdated listings, poor filters, fake reviews, thin vendor pages, or weak local coverage can undermine destination status.

For publishers, local visibility can support subscriptions, memberships, event revenue, and advertising, but only if content creates reader loyalty. A one-time click from a local explainer has less value than a repeat relationship. Preferred Sources and AI Search make that relationship more visible in Search itself.

The money is not in “local domains” as a label. The money is in owning the local user’s task at the moment they need a trustworthy destination.

The risks of over-reading one SISTRIX window

The strongest May local-domain evidence comes from a specific dataset, a specific tool, two countries, and a short window at the end of the rollout. That does not weaken the finding; it defines its boundary. SEO teams should use it as a hypothesis generator, not as a universal law.

SISTRIX visibility is useful because it allows cross-domain comparison at scale. It is not the same as Search Console data. It does not know every query that matters to a business. It does not reflect conversion value. It can overweight or underweight query sets relative to a specific site’s revenue. It can show visibility movement that does not map to clicks if SERP features change.

The May window also overlapped with AI Search announcements, volatility waves, and other interface changes. Google completed the update on June 2, but rank tracking and click data may need additional time to stabilize. Google’s own guidance recommends waiting at least a week after completion before drawing conclusions in Search Console.

There is also survivorship bias in winner-and-loser analysis. The domains that move most are easiest to discuss. Smaller changes across many sites may be more representative but less visible. A handful of examples can illustrate a pattern without proving that pattern explains every movement.

The right use of Solis’s analysis is comparative. Take the pattern and test it against your data. Did local pages gain where global pages lost? Did task-completion destinations replace informational layers? Did reference or official sources replace tools or forums? Did social and visual formats gain where text pages lost? Did health queries split by format? If yes, the May interpretation may explain your movement. If no, keep looking.

This caution also protects against harmful SEO narratives. “Local domains won” can become a simplistic sales pitch for ccTLD migrations. “Forums lost” can become a reason to remove community features. “Aggregators lost” can become a reason to abandon comparison tools. The actual data says something subtler: source-type fit varied by query and market.

Good SEO analysis preserves uncertainty without becoming useless. The May evidence is strong enough to guide audits. It is not strong enough to justify blanket site migrations or mass content deletion.

A practical framework for the next core update

The May update gives search teams a better framework for future core updates. Instead of asking only whether content is “helpful,” ask whether each priority page is the expected destination type for the query in the market.

The framework has six parts.

First, identify the market. Do not analyze global traffic first. Segment by country, language, and device. Local-domain gains can disappear in blended data.

Second, identify the task. Classify affected queries by what the user is trying to do. Buying, booking, applying, comparing, defining, troubleshooting, verifying, reading news, getting local service, or learning from experience all imply different destinations.

Third, identify the current winning source type. Look at the post-update SERP. Are winners local domains, official sources, marketplaces, videos, forums, publishers, tools, images, social platforms, or brand pages?

Fourth, compare your page’s role. Does it actually perform the same role as the winners? If not, decide whether to change the page, create a better destination, or stop chasing that query.

Fifth, verify technical routing. For international and local pages, check indexability, canonicals, hreflang, internal links, sitemaps, structured data, and feed alignment.

Sixth, measure business impact. Rankings matter, but revenue, leads, subscriptions, local actions, and qualified traffic decide priority. A large visibility loss on low-value informational queries may be less urgent than a smaller loss on local commercial terms.

This framework works because it respects what core updates actually are. Google says core updates are broad changes to ranking systems and do not target specific sites or pages. That means recovery is not a single fix. It is a process of becoming more deserving relative to the current result set.

For local domains, the framework prevents both complacency and panic. A local domain that gained should still ask why. Was it better market fit, or temporary volatility? Did competitors lose because of wrong-market routing? Are gains concentrated in categories with weak conversion? A global domain that lost should ask whether the right local page exists and whether users can complete the task there.

The next core update will not repeat May exactly. But the discipline of matching market, task, and destination type will remain useful.

Post-update diagnosis framework

Compact table for site owners

This table should not replace manual SERP review. It is a triage tool. The correct recovery action depends on which result type replaced the affected page and whether your page can credibly become that type.

The real lesson from May 2026

The May 2026 Core Update is best understood as a destination-fit update in the observed data, not because Google officially named it that, but because the visible shifts point in that direction. Local domains gained where local market fit mattered. Canonical reference brands gained where canonical answers fit. Task-completion marketplaces gained where users needed to act. Forums declined where raw discussion was less suitable. Health results split by trust and format.

That reading is useful because it gives teams something concrete to do. It moves the conversation away from vague quality talk and toward evidence. Which market? Which task? Which source type? Which page role? Which technical signals? Which user proof?

For global brands, the lesson is humility. A powerful .com cannot assume it deserves every market. If a local version exists, it must be strong enough to stand on its own. If it does not exist, the global page must still satisfy local expectations. Authority is not a license to be imprecise.

For local brands, the lesson is opportunity with responsibility. Local relevance can matter, but only when it is real. A local domain, local address, or local landing page is not enough. The site must solve local user needs better than broader competitors.

For publishers, the lesson is role clarity. Original reporting, local expertise, preferred-source loyalty, and durable explainers matter more than broad rewrite coverage. Search is becoming more selective about when it wants an article, an official page, a tool, a marketplace, a forum, or an AI-supported answer.

For affiliates and aggregators, the lesson is proof. If users can complete the task elsewhere, the intermediary must add something the destination cannot. That may be testing, comparison, filtering, local context, pricing history, or expert interpretation. Without that, the page is a layer, not a destination.

For SEO teams, the lesson is discipline. Wait for stable data. Segment by market. Compare SERPs. Diagnose source-type changes. Fix confirmed technical problems. Improve pages where they remain the right destination. Do not rewrite everything. Do not delete in panic. Do not chase domain-extension myths.

Local domains gained in Google’s May Core Update because, in the measured data, Google appeared to reward the result that belonged to the user’s market and task. That is a bigger lesson than local SEO. It is the direction of search itself.

The long arc from core updates to destination ranking

Core updates have always forced the web to reprice relevance. A site that once looked useful can become less competitive because better content emerges, user expectations shift, product formats change, or Google’s systems learn to identify quality differently. Google’s own restaurant-list analogy in its core update documentation makes this point: rankings can change because the list is reassessed, not because a downgraded item became bad.

The May update fits that long arc. A global page that ranked for years may not have become worse on May 21. A local page may not have become brilliant overnight. The system may have become more confident that the local page was the better answer for that market. That is a different kind of loss. It is not a penalty. It is a relative demotion caused by a better-matched alternative.

This is why core update recovery often feels frustrating. Site owners ask what they did wrong. Sometimes the answer is nothing specific. The competitor set changed. The result type changed. The user expectation changed. The ranking system changed how it weighs those signals. The path forward is not confession. It is adaptation.

Destination ranking is a useful phrase because it captures the shift from content as an asset to content as an interface for a task. Search engines do not rank documents in a vacuum. They rank destinations for users. A page is a destination when it satisfies the user’s intent, reduces uncertainty, supports the next action, and fits the market context. A page is merely content when it talks about the topic without owning the task.

Local domains gain naturally under destination ranking when the task is market-bound. A UK shopping query, local health query, national regulation query, regional travel query, or city service query has local destination logic built in. The user may not say it. The query context says it.

AI Search may strengthen this direction because AI systems need reliable grounding. A vague global article is less useful as grounding than a page with explicit local facts, current data, and clear task details. Google says its generative AI features rely on core Search ranking and quality systems and use retrieval techniques to find relevant, current pages from the Search index.

The future advantage belongs to sites that are not merely about topics, but are clearly useful destinations for specific users in specific contexts. May was not the start of that shift. It was a visible reminder.

The local signal inside global search

Global search is not borderless in the way many websites wish it were. Language, law, logistics, currency, culture, trust, and service delivery all create borders inside user intent. Google has spent years improving its ability to infer those borders. The May local-domain gains suggest that, at least in some UK commercial SERPs, those borders became more visible.

A global search system has to balance two truths. The web is international, and users often benefit from global sources. But many queries are local even when they lack local words. A user searching “best broadband deals” from Manchester, “mattress delivery,” “return policy,” “tax deadline,” “flu vaccine,” or “train delay compensation” carries a market context. Showing a global page may be technically relevant and practically wrong.

Country-specific domains, local folders, and localized content are ways to express that context. But Google also looks beyond the URL. User location, language, query phrasing, domain history, page content, structured data, links, brand entity, and behavior signals may all contribute. Google does not publish a simple formula, and site owners should be wary of anyone who claims one.

The operational takeaway is to make the local signal redundant. Do not rely on one cue. Use clean URL architecture, correct hreflang, local internal links, localized metadata, visible local service information, local feeds, country-specific legal and support pages, local reviews, local media, local backlinks, and consistent entity data. If one signal is weak, others should confirm the intended market.

Global sites also need to avoid forced redirects that block crawlers or frustrate users. A country selector can help, but it should not hide content or prevent Google from accessing alternate versions. Local relevance should be expressed on crawlable pages, not only through scripts, modals, or checkout logic.

A local signal is strongest when the entire experience agrees. The URL, content, data, links, checkout, support, and brand presence should all say the same thing: this is the right page for this market.

The hidden cost of generic global templates

Generic global templates are efficient. They allow brands to launch thousands of pages across countries with shared design, shared copy blocks, shared components, shared product data, and shared SEO logic. They also create a subtle risk: the page may look technically localized but feel operationally generic.

The May update’s local-domain pattern exposes that risk. A local page built from a global template may carry the correct hreflang and currency but still lack local depth. The category copy may be identical. The buying guide may ignore local regulations or sizing. The product selection may include unavailable items. Reviews may come mostly from another country. Delivery information may sit behind a modal. Returns may be written for a global policy rather than a local promise.

Users notice these gaps. Search systems may notice them indirectly through content, data consistency, result satisfaction, and comparison with stronger local competitors. A local domain with a weaker global template may outperform a global .com because the market signal is clearer. But a local domain with a generic template is still vulnerable to a competitor that owns the local experience more deeply.

Template efficiency should therefore be separated from content sameness. A brand can use a shared technical framework while allowing local modules to vary. Local merchandising blocks, local buying advice, local FAQ content, local store and delivery data, local expert commentary, and local imagery can all sit inside a global design system. The template remains efficient, but the page becomes market-specific.

This is not only ecommerce. SaaS sites use generic templates for country pages that ignore data residency, local integrations, procurement rules, and support coverage. Education sites duplicate course pages without local accreditation context. Travel sites reuse destination pages without local transport changes. Finance sites reuse explainers without local tax or regulatory detail. Health sites reuse symptom pages without local care pathways.

Generic templates become a ranking risk when they prevent the page from proving local usefulness. May did not create that risk. It made the risk more visible.

Search Console interpretation after local gains and losses

Search Console is the most important owned dataset after a core update, but it must be read carefully. Google’s core update guidance says to compare the week after the update’s completion with the week before the rollout began, and to analyze different search types separately.

For May, that means teams should build comparisons around three periods: pre-rollout baseline before May 21, rollout period from May 21 to June 2, and post-stabilization period after the first full week following completion. The rollout period should be labeled as volatile and not used as the sole basis for permanent decisions.

Country segmentation is central. If local domains gained in the UK but not the US, global Search Console data may show a mild decline while the UK business feels a large impact. Segmenting by country reveals whether losses are market-specific. Segmenting by page reveals whether the wrong local version is ranking. Segmenting by query reveals whether the intent changed.

CTR analysis is also critical. AI Overviews, rich results, product grids, local packs, images, videos, and other features can change click behavior without a matching ranking loss. Google’s traffic-drop documentation says if impressions remain the same but clicks drop, titles, snippets, and competing result features may be part of the issue.

For local pages, compare clicks and impressions by country against availability and revenue. A page that gained UK impressions but not UK sales may have a conversion or market trust issue. A page that lost impressions but held conversion rate may have a pure visibility problem. A page that held rankings but lost CTR may face SERP changes.

Average position should be treated cautiously. Google itself warns site owners not to focus too much on absolute average position and to look for dramatic, persistent drops before deeper self-assessment. Average position can be distorted by country mix, query mix, SERP features, and impressions across long-tail variants.

The purpose of post-update data analysis is not to prove that Google was unfair. It is to find the exact point where your page stopped being the preferred destination.

SERP reconstruction as the missing SEO skill

Many recovery audits fail because they compare the affected page with its old self rather than with the new winners. SERP reconstruction fixes that. It asks what the result page looked like before, what it looks like after, and which source types replaced the affected URL.

This is especially useful after May because the observed pattern was source-type movement. If a local domain replaced a global domain, the new SERP is telling you something. If a marketplace replaced a guide, the new SERP is telling you something. If a canonical reference brand replaced a tool, the new SERP is telling you something.

SERP reconstruction should capture:

The top organic results before and after. The domains and URL patterns. Country domains versus global domains. SERP features. AI Overviews or AI Mode exposure where measurable. Product, image, video, news, and local elements. The user task implied by winners. The page formats that lost.

Historical rank tracking tools can help, but manual review is still needed. A ranking table will tell you who moved. It will not tell you whether the new winner has local inventory, a better booking engine, original reporting, official status, or a more useful format.

For international SEO, reconstruct SERPs from the affected country. Do not judge a UK loss from a US browser. Use reliable localized rank tracking and manual checks where possible. Consider language, personalization, and device. Mobile SERPs may express local and AI features differently from desktop.

The output of SERP reconstruction should be a page-role decision. If your page can become the role Google now prefers, improve it. If it cannot, create or strengthen a different page. If the query no longer fits your business, stop chasing it. If the loss is temporary volatility, monitor.

SERP reconstruction turns core update analysis from guesswork into competitive reading. May rewarded teams that can read the result page as a product decision, not only as a ranking list.

Local entities and the trust layer

Local domains gained where they represented more than a URL. They represented a local entity. Entity trust is difficult to quantify, but it is visible in the web’s signals: company registrations, local addresses, reviews, news mentions, customer service details, social profiles, store pages, citations, partnerships, policies, and user demand.

A local entity can be a retailer, publisher, hospital, university, government body, marketplace, service business, or community platform. The common thread is that users in the market recognize it as relevant. Google’s systems may use many signals to infer that recognition. SEOs do not need to know every signal to improve the underlying reality.

Entity clarity is especially important when a global brand operates multiple local domains. Are they all the same company? Different subsidiaries? Franchisees? Distributors? Regional publishers? Separate legal entities? The website should make this clear. Ambiguity can undermine trust for users and search systems.

Local entity pages should include accurate About information, legal details, contact options, local leadership or editorial information where relevant, service areas, store or office data, and relationships to parent brands. For publishers, author pages, editorial policies, corrections policies, and ownership information matter. For health sites, medical review and institutional identity matter. For ecommerce, returns, warranty, and support details matter.

Preferred Sources adds a user-facing layer to entity trust for publishers. Users can choose a publication at the domain or subdomain level, and Google may label it in Top Stories, AI Overviews, and AI Mode. That makes the domain itself a unit of preference.

A local domain without a local entity is just a domain. A local domain with a trusted local entity is a destination. The May update appears to have favored the second condition in market-sensitive result sets.

Programmatic local content faces a stricter test

Programmatic SEO can be useful when it turns structured data into genuinely useful pages. It becomes risky when it generates thousands of near-identical pages with minimal local value. The May update’s local-market signal should not be used as justification for more programmatic local pages. It should be used as a reason to raise the standard.

A travel site can create useful programmatic pages for routes, hotels, destinations, and dates if the data is fresh, the filters work, and the page helps users act. A real estate site can create useful pages for neighborhoods if it has listings, prices, maps, schools, transport, and local market data. A job site can create city-role pages if listings are fresh and filters are helpful. These are task destinations.

A lead-generation site that creates “best plumber in [city]” pages for every town without real vetting, local proof, or service depth is not a strong destination. A SaaS site that creates “CRM software in [country]” pages with identical copy and no local pricing, compliance, or customer evidence is weak. A publisher that creates AI-generated local explainers without reporting is exposed.

Google’s generative AI content guidance is explicit that using AI or similar tools to generate many pages without adding value may violate scaled content abuse policies.

The programmatic test after May is simple: does the page contain market-specific data or functionality that could not be replaced by a generic template? If yes, it may be worth indexing. If no, it may dilute quality or compete with better local pages.

Programmatic pages also need lifecycle management. Local data changes. Listings expire. Prices move. Stores close. Regulations update. If the page cannot stay current, it may not remain a good result. Search engines are likely to prefer fresher task destinations where freshness is central to the query.

Programmatic local SEO is not dead. Weak programmatic local SEO is easier to identify when Google cares more about destination fit.

The role of media and visual proof in local relevance

The May data showed visual and social platforms performing differently from forums and Q&A. Pinterest and Fandom gained in both measured markets, while YouTube, X, and Facebook were mixed to positive in the US.

This matters for local domains because media can prove local relevance in ways text cannot. A retailer can show local store photos, product videos, size demonstrations, installation images, delivery examples, and user-generated images. A restaurant can show current menu photos, interior shots, accessibility details, and neighborhood context. A local publisher can use maps, charts, embedded documents, photos, and short video explainers. A service business can show completed work, staff, vehicles, certifications, and before-and-after examples.

Google’s generative AI guidance says images and video can create more opportunities for a website to appear beyond text links, and its AI feature guidance lists images and videos among continuing SEO best practices where applicable.

Visual proof is not decoration. It helps users decide whether the page is real, current, local, and trustworthy. It can also support Google Images, video results, AI features, Discover, and product surfaces. For local domains that gained after May, strengthening visual assets can help defend the advantage.

The key is relevance. Stock imagery and generic product images add little. Market-specific visuals add more. A UK retailer showing UK packaging, local delivery context, store pickup, or real customer use is more convincing than a generic manufacturer image. A local health provider showing facilities, clinicians, accessibility, and service routes can reduce uncertainty. A publisher showing original photographs from a local event proves reporting presence.

Local relevance is easier to believe when users can see it. Text explains. Media verifies.

The competitive reset for global .com domains

Global .com domains still have major advantages: brand recognition, backlinks, crawl frequency, content scale, product breadth, engineering resources, and cross-market authority. The May update did not erase those advantages. It showed that they can be overridden when market fit is poor.

For global .com teams, the response should not be defensive. It should be architectural. Decide which queries the global domain should own and which queries local domains or folders should own. Then align internal links, canonicals, hreflang, content, feeds, and analytics with that decision.

A common mistake is letting the strongest URL win internally. If the .com page has more links, better content, and stronger templates, it may outrank local pages even when the business wants local pages to rank. That can feel like an SEO win until a core update corrects it or users fail to convert. The better approach is to build local pages strong enough that Google does not have to choose between authority and market fit.

Global domains should also create rules for cross-market content reuse. Which blocks can be global? Which must be local? Product specifications may be global. Delivery, returns, pricing, legal claims, reviews, and buying advice often need localization. Editorial explainers may share a base but require local context. Health content may require country-specific care pathways. Finance content may require local regulation.

For brands that cannot maintain many local versions, the answer may be to narrow indexation. It is better to have fewer strong market pages than many weak alternates. If a business does not serve a market, it should not create the impression that it does.

The global .com is not dead. The lazy global .com is weaker. May rewarded brands that could express global authority through local usefulness.

Core updates and the myth of one recovery lever

Core update recovery advice often becomes too simple: improve content, build E-E-A-T, fix technical SEO, remove low-quality pages, get links, improve UX. All of these can matter. None of them is the one lever.

The May update shows why. A local ecommerce loss may need international routing and local merchandising. A reference loss may need source-role clarity. A forum loss may need content governance. A marketplace gain may come from task completion. A health loss may need format and trust alignment. A publisher loss may need original reporting or a different query target.

Google’s core update documentation says negative impact may not mean something is wrong with pages, and its traffic-drop documentation says algorithmic updates can change how pages perform without a fundamental content flaw.

Recovery therefore requires multiple lenses. Content quality asks whether the page is useful. Intent fit asks whether it is the right type of page. Market fit asks whether it serves the right country or locality. Technical SEO asks whether Google can understand and show it. Business fit asks whether the page supports the task. Brand fit asks whether users trust it. SERP fit asks whether it resembles the winners for the query.

The one lever is not a tactic. It is alignment. The page, market, task, format, entity, and technical signals must agree. When they do, a site is resilient. When they do not, a core update can expose the gap.

May did not reward a checklist. It rewarded coherence.

Search visibility now depends on being less replaceable

The web is full of replaceable pages. Generic explainers, copied product descriptions, shallow comparison lists, location pages with swapped city names, AI-assisted summaries, rewritten press releases, and thin affiliate guides are easy to produce and easy to replace. Core updates tend to make replaceability more expensive.

Local domains can be less replaceable when they own facts and services tied to a market. A local retailer’s live inventory is not easily replaced by a global article. A local publisher’s original reporting is not easily replaced by a summary. A local health system’s care pathway is not easily replaced by a generic medical overview. A marketplace’s live listings and filters are not easily replaced by a static guide.

Google’s generative AI guidance uses the concept of non-commodity content and warns against recycling what others have already said.

The replaceability test is harsh but useful. If another site, or an AI system, could produce a similar page from public information in a few minutes, the page has weak defensibility. If the page contains proprietary data, firsthand evidence, local operations, expert judgment, unique visuals, tools, or original reporting, it has stronger defensibility.

For SEO teams, this shifts investment away from volume. Publishing more pages may increase coverage, but if the pages are replaceable, they add risk. Improving fewer pages with harder-to-copy value may support both classic rankings and AI citation potential.

The local domains that gained were not rewarded for being local in name. They were rewarded where locality made them less replaceable as search destinations.

The role of links after a destination-fit update

Links still matter, but the May pattern shows that links do not settle every ranking question. A global .com may have more links than a local domain. A major publisher may have more links than a specialist page. A government or academic site may have more authority than a patient-friendly health destination. Yet the better source type can win.

This does not mean link building is irrelevant. It means links should support the correct destination. For local domains, market-relevant links are especially useful because they confirm local entity value. Local newspapers, industry associations, chambers of commerce, suppliers, partners, universities, events, charities, review sites, and community organizations can all provide genuine market signals.

Global link equity can also be routed through internal linking. If a global brand has strong authority on the .com domain but wants local pages to rank, it should link prominently to local versions. Country selectors hidden in footers may not be enough. Category, product, support, and content pages should guide users and crawlers to the relevant local destination.

Links also interact with content roles. A guide with strong links may lose to a marketplace because the query now prefers a task-completion result. The guide may still rank for research queries. The marketplace may win for transactional queries. Link strategy should respect that split.

For publishers, original reporting can attract links that generic rewrites cannot. Google’s features such as Highly Cited are part of a broader effort to make original sources easier to identify, according to its May 27 Search blog.

Links amplify relevance. They do not always define it. After May, the better question is whether links are strengthening the page that truly fits the market and task.

Local domain gains and the future of GEO

Generative engine optimization, answer engine optimization, and AI search visibility have become popular labels, but Google’s own guidance says that, from its perspective, optimizing for generative AI search is still SEO because AI features are rooted in core Search ranking and quality systems.

The May update supports that view. The same qualities that helped local domains gain in classic visibility may matter in AI-mediated retrieval: clear market identity, task usefulness, original facts, crawlable content, structured technical foundations, and trustworthy entity signals. The tactics may change, but the underlying source quality does not become separate.

AI systems also make local specificity more valuable. A conversational query may include constraints that classic keyword pages never addressed: “best laptop for a UK student under £700 with next-day delivery,” “private dentist open Saturday near Bristol,” “rights if my flight from Heathrow is delayed,” or “running shoes for flat feet available in size 10 UK.” A generic page may struggle. A local destination with precise data can answer.

Research on AI Overviews and generative search suggests source selection can differ from traditional search results and that interface design affects publisher and platform impact.

For GEO strategy, the May lesson is to avoid hacks. Google’s AI guidance says sites do not need special markup such as llms.txt, do not need artificial chunking, and do not need to rewrite content just for AI systems. It recommends the same fundamentals: useful content, technical accessibility, media where relevant, and user-focused pages.

The best GEO strategy after May is not to write for machines. It is to become the most reliable, specific, crawlable source for a real user task.

Strategic priorities for the next ninety days

Teams affected by the May update should prioritize work by risk, evidence, and business value. The first ninety days should not be spent rewriting every page. They should be spent identifying which losses matter, which patterns repeat, and which fixes are likely to change the site’s destination fit.

The first priority is measurement. Build post-update comparisons after the waiting period. Segment by country, page, query, device, and search type. Mark pages affected by AI features or major SERP layout changes where possible. Separate ranking loss from CTR loss.

The second priority is market routing. For international ecommerce, travel, SaaS, finance, and publisher sites, audit wrong-market rankings. Fix hreflang, canonicals, internal links, sitemaps, local feeds, and local page strength. Do not migrate domains without evidence.

The third priority is source-role clarity. For each affected page, define whether it should be a local service page, product category, guide, tool, marketplace, reference, news article, review, forum thread, or official resource. If it is trying to be several at once, split or refocus carefully.

The fourth priority is proof. Add local evidence, original data, expert input, firsthand experience, visuals, tools, and market-specific details. Remove generic filler. Google’s guidance on helpful and generative AI content both point away from recycled or commodity material.

The fifth priority is audience. Publishers and brands should encourage direct relationships: newsletters, apps, accounts, communities, preferred-source selection where relevant, and branded search. Search visibility is more durable when users know who they want.

The sixth priority is governance. Core updates expose organizational weaknesses. International SEO needs content, engineering, legal, merchandising, customer support, and analytics alignment. AI search needs editorial, technical, and product teams working together. Local SEO needs operations and reputation management, not only website edits.

The May winners were not only better pages. They were often better-aligned businesses.

The final read on local domains and Google’s May Core Update

The headline “local domains gained” is true in the most visible part of the analysis, but it is only the entry point. The deeper story is that Google’s May 2026 Core Update appears to have sharpened the relationship between query, market, and destination type. Local domains gained where the market mattered. Canonical sources gained where direct reference mattered. Marketplaces gained where task completion mattered. Forums declined where raw discussion was less suitable. Health results split by trust and format.

That makes May a useful update for serious SEO teams because it rewards better questions. Do we own the market? Do we own the task? Do users recognize us as the right destination? Does our local page prove its value? Are our technical signals helping Google choose correctly? Are we building assets that AI Search and classic Search can both use? Are we less replaceable than the pages we want to outrank?

The answer will differ by sector. A local retailer may need stronger category pages and feed consistency. A global brand may need to stop letting the .com outrank local versions. A publisher may need more original reporting and clearer local identity. An affiliate site may need proof and tools. A forum may need content governance. A health site may need format-specific trust.

Google will not provide a site-by-site explanation. Core updates rarely work that way. But the visible data gives enough direction to act wisely. The safest strategic response is to build pages and domains that are unmistakably the right destination for a specific user, in a specific market, with a specific task.

That is why local domains gained in May. Not because local is a shortcut. Because, when search becomes more precise, being the right local answer is a powerful advantage.

Practical questions about Google’s May Core Update and local domains

Author:
Jan Bielik
CEO & Founder of Webiano Digital & Marketing Agency

This article is an original analysis supported by the sources cited below

May 2026 core update
Google’s official Search Status Dashboard entry confirming the May 2026 core update start time, completion time, and rollout status.

All incidents reported for Ranking
Google’s ranking update history showing confirmed Search updates, including the May 2026 core update and other recent ranking events.

Google Search’s core updates and your website
Google Search Central documentation explaining how broad core updates work and how site owners should analyze ranking changes.

Creating helpful, reliable, people-first content
Google Search Central guidance on assessing content quality, usefulness, reliability, and people-first publishing standards.

Optimizing your website for generative AI features on Google Search
Google’s guidance for site owners on AI Overviews, AI Mode, query fan-out, retrieval-augmented generation, and SEO fundamentals for generative AI search.

AI features and your website
Google Search Central documentation explaining how AI Overviews and AI Mode work from a site owner’s perspective.

Guide to Preferred Sources in Google Search for web publishers
Google documentation explaining Preferred Sources, domain-level eligibility, and how preferred source labels can appear in Top Stories, AI Mode, and AI Overviews.

How Google Search helps you find original, quality content
Google’s May 2026 blog post about Preferred Sources in AI Search, original content discovery, and publisher visibility features.

A new era for AI Search
Google’s I/O 2026 announcement describing major AI Search updates, agentic search features, and changes to the Search experience.

May 2026 Core Update visibility analysis and data updates
SISTRIX’s visibility tracking and data updates during the May 2026 core update rollout.

Google May 2026 Core Update analysis
Aleyda Solis’s analysis of US and UK SISTRIX visibility shifts showing patterns around intent, market fit, source type, local ecommerce domains, forums, marketplaces, and health results.

Google’s May Core Update favored pages that match intent
Search Engine Journal’s coverage of Aleyda Solis’s SISTRIX analysis and the finding that local UK domains gained while some global .com domains lost ground.

Google’s May Core Update complete after volatile rollout
Search Engine Journal’s report on completion of the May 2026 core update, rollout timing, volatility observations, and post-update analysis guidance.

Google May 2026 core update rolling out now
Search Engine Land’s report on the launch of the May 2026 core update and Google’s public wording at rollout.

Google May 2026 core update rollout is now complete
Search Engine Land’s report confirming completion of the May 2026 core update and summarizing volatility during the rollout.

Google May 2026 Core Update has completed rolling out
Search Engine Roundtable’s coverage of the completed rollout, timing, observed volatility, and quick facts around the May 2026 broad core update.

Debug Google Search traffic drops
Google Search Central documentation for diagnosing traffic drops, distinguishing algorithmic changes, technical issues, CTR shifts, and demand changes.

Localized versions of your pages
Google Search Central documentation on hreflang, localized versions, alternate URLs, reciprocal annotations, and international SEO signals.

How to specify a canonical URL with rel canonical and other methods
Google Search Central guidance on canonicalization, duplicate URLs, rel=”canonical”, redirects, sitemap signals, and hreflang-canonical consistency.

Google Search’s guidance on generative AI content on your website
Google guidance explaining acceptable use of generative AI in content workflows and warning against scaled content that lacks user value.

Measuring Google AI Overviews
Academic measurement study examining Google AI Overviews activation, cited domains, claim fidelity, and publisher implications.

The impact of AI Search on the online content ecosystem
Academic study examining how Google AI Overviews and AI Mode affect Reddit engagement and the relationship between AI search interfaces and online content platforms.

How generative AI disrupts search
Academic study comparing Google Search, Gemini, and AI Overviews across source retrieval, generated answers, and implications for web visibility.

Answer bubbles
Academic paper examining information exposure, source selection, and fidelity in AI-mediated search systems.