Google September 2026 Spam Update Completes After 14-Day Rollout

The Google September 2026 Spam Update officially finished rolling out on October 8, 2026 at 4:37 am ET, having started on September 28 — a total runtime of nearly two weeks. According to Search Engine Roundtable, this made it dramatically different from the three previous spam updates in 2026, which each took only two days. Search Engine Journal confirmed it was the longest-running spam update of the year, suggesting broader scope and more complex enforcement mechanisms.

Search Engine Roundtable had flagged the final phase as occurring between October 4–6, predicting one more spike in volatility before the update concluded — a call that proved accurate. All major sources (Search Engine Land, Search Engine Journal, Search Engine Roundtable) confirmed completion within the same window.

At the Google Search Central Live Deep Dive Europe event in Barcelona, Google revealed why spam updates have become so frequent in 2026: the explosion of AI-generated content has dramatically increased the volume of new content hitting the web, and Google is now using AI itself to detect and combat spam. This admission — reported via John Campbell's recap at We Are Roast — is a significant confirmation of an AI-vs-AI arms race in search quality enforcement.

Multiple SEOs are reporting a significant uptick in Google manual action penalties in the weeks following the spam update, with the increase appearing to accelerate more recently. Search Engine Roundtable notes the spike started several weeks ago but intensified around the update's conclusion. Google has not officially confirmed any policy change in manual action volume, but the pattern is consistent enough to warrant SEOs proactively checking Search Console for alerts and auditing site compliance.

Prominent SEO authority Lily Ray published a prediction on Search Engine Journal that a major Google algorithm update is around the corner, citing three converging signals: accelerated spam update cadence, recent changes to Google's documentation, and explicit statements from Google itself at Barcelona. Marie Haynes' newsletter similarly flagged signs of a big update emerging, directly linking Google's updated helpful content guidance to an imminent algorithmic shift.

Search Engine Journal's analysis of Google's recent signals suggests the next core update could deliver significant ranking changes — particularly for content quality and author authenticity. SEOs are advised not to make hasty site changes during early rollout volatility, as SEJ separately reported that early signals during updates do not reliably predict final ranking outcomes.

Google quietly updated its Helpful Content documentation to place significantly greater emphasis on the quality of a page's 'main content' (MC). As analyzed by Marie Haynes in Search Engine Journal, this revision reveals what Google's ranking systems are currently being trained to reward — making it a direct window into active quality signal weighting. The update also introduced new EOT (Expertise, Originality, Trustworthiness) and SA (Supplementary Content/Ads) sections.

Haynes argues that SEOs underutilize this documentation as a signal source, noting it explicitly articulates what Google considers high- vs. low-quality MC. The timing of this update — coinciding with spam update completion and imminent core update predictions — makes it especially significant.

Simultaneous with the helpful content documentation update, Google published an explicit warning to "avoid using deceptive authorship information" in content. According to Search Engine Roundtable, this is a direct policy signal targeting AI-generated bylines, fabricated author profiles, and other forms of authorship manipulation. The warning is tightly coupled with E-E-A-T enforcement — sites using fake authors or misleading attribution face increased scrutiny under the updated guidance.

Search Engine Journal's weekly SEO Pulse recap highlighted fake author profiles as a notable new content quality signal, framing it alongside the spam update conclusion as part of a broader enforcement push.

At the Google Search Central Live Deep Dive Europe event in Barcelona, Google's Gary Illyes presented rare internal data showing typical versus slowest timeframes for crawling, indexing, and serving — broken down by each process stage. According to both Search Engine Journal and Search Engine Roundtable, the presentation also covered timelines for site migrations (measured in months) and core update recovery — concrete figures that SEOs have long sought for client expectation management.

This kind of granular, data-backed transparency from Google is unusual. Practical implication: site move timelines are longer than many practitioners assume, and core update recovery is not a quick process even after the update itself concludes.

Three distinct technical crawling developments emerged this week. First, Google admitted that its 'couldn't fetch' sitemap error in Search Console may have nothing to do with whether Googlebot can actually access the sitemap — a significant admission that invalidates a common diagnostic assumption, per Search Engine Journal.

Second, Google updated its crawl rate documentation to add guidance on using the `Retry-After` HTTP header to urgently reduce Googlebot traffic during server emergencies, alongside broader organizational improvements to the documentation page.

Third, Google's John Mueller revealed that AI training crawlers — which unlike Googlebot do not offer sitemap submission tools — are actively accessing default sitemap filenames and RSS feeds as observed in his own server logs. Mueller's recommendation: use standard sitemap filenames (e.g., `/sitemap.xml`) or maintain an RSS feed to maximize discoverability by AI crawlers.

Google's John Mueller addressed a specific scenario this week: why Google crawls thousands of pages on large product aggregator sites but indexes very few — sometimes not even the homepage. Mueller categorized these sites as "large product aggregators" and explained that Google's evaluation criteria differ for this site type, implying that aggregator-style content faces a higher bar for perceived unique value. SEOs managing large comparison sites or product directories should note this framing when diagnosing thin indexing issues.

New tracking data reported by Search Engine Journal shows that AI Overviews (AIOs) now appear on the majority of big-brand search queries — a significant expansion from earlier, more topic-based deployments. Marie Haynes' newsletter flagged this directly, raising the question of whether brand information surfaced in AIOs is accurate, and urging brand managers to audit what AI Overviews say about their company.

Google also added an experimental 'Add to Calendar' feature within AI Overviews for event-related queries, though Search Engine Roundtable noted it was inconsistent and not reliably replicable. A legal challenge to AI Overviews was dismissed by a court this week, per Marie Haynes' newsletter.

Search Engine Journal's Duane Forrester reported on a Google DeepMind development that could fundamentally reshape SEO: a single AI model has reportedly been demonstrated to rank search results without relying on a separate traditional index. If this technology ships into production, Forrester argues that SEO visibility would shift from a positional model to a binary inclusion/exclusion model — you're either in the AI's knowledge or you're not, with no position 1-10 to optimize for. This is currently theoretical, but the mathematical proof-of-concept exists.

Two Ahrefs pieces this week tackled the attribution problem head-on. On visibility measurement, Ahrefs cited an example where AI drove more signups than YouTube — ranking second only to Google as a signup source — yet traditional referral attribution badly undersold it because most AI influence happens before the user visits the site at all. On the ROI front, Ahrefs cited Forrester data showing 94% of B2B buyers now use AI search in purchasing decisions, rating it above vendor websites, product experts, and sales reps. NielsenIQ puts AI-assisted consumer research at 42% of shoppers.

Kevin Indig's Growth Memo complemented this by arguing that traditional traffic-to-revenue SEO targets no longer make sense for most brands in 2027 planning, urging a shift toward capacity-based SEO budgeting rather than traffic targets — a direct consequence of AI-driven dark traffic. Eli Schwartz similarly argued in his newsletter that SEO still matters in a zero-click world, but its value must be reframed as visibility and influence rather than direct traffic generation.

Anthropic updated Claude's usage policy this week to explicitly prohibit seeding search engines and AI answer systems with misleading content designed to manipulate AI-generated responses — a direct shot at emerging GEO manipulation tactics. Notably, the update also drops automated publishing from its high-risk category, per Search Engine Journal.

Separately, Search Engine Journal reported that Google is actively working to block SERP tracking tools, and the proliferation of AI agents is expected to make this clampdown more aggressive. As AI handles more searches in non-traditional interaction patterns, rank tracking as practitioners know it faces structural pressure.

Google has shipped a long-requested feature: Search Console's Performance report now supports multi-select country filters, allowing users to combine and compare traffic data across multiple countries simultaneously in a single view. Previously, country analysis required switching between individual country filters. The update is confirmed live and directly benefits international SEO practitioners managing multi-regional site performance.

Google Search Console has begun sending emails to site owners reporting how many users have selected their domain as a 'preferred source' in Google Search. The email surfaces a specific count of 'Preferred source selections,' indicating Google is building out a personalization and following layer in search. This metric is new to most webmasters and may eventually factor into brand visibility signals. Search Engine Roundtable also confirmed that dark mode is coming to Search Console, as announced at the Barcelona event.

Google updated its Search Console generative AI control page to clarify how child property settings can override parent-level AI inclusion/exclusion controls. This is particularly relevant for publishers and site owners managing complex multi-domain or subdomain structures in Search Console, where AI content settings may behave differently than expected across property hierarchies.

OpenAI has begun rolling out GPT-6 in ChatGPT alongside a new 'Intelligent UI' that dynamically adds visuals, charts, and interactive tools to answers depending on the query. According to Search Engine Journal, this is a significant product leap that reshapes how users interact with AI-generated answers — moving them closer to a rich, visual experience that competes more directly with traditional web browsing and search.

OpenAI announced it will test a new visual ad format inside ChatGPT that surfaces specifically during AI image generation. The format is being tested initially with a select group of US advertisers, per both Search Engine Roundtable and Search Engine Journal. This marks a concrete step toward ChatGPT becoming an advertising platform — and raises early questions for brands about visibility and GEO strategy in AI-native ad environments. New advertiser measurement tools are also being introduced alongside the ad test.

OpenAI is also testing descriptive product labels on ChatGPT shopping results — tags like 'best all-rounder' or 'best for beginners' — mirroring Google's editorial-style product recommendations and potentially influencing how brands optimize for AI-driven commerce.

OpenAI is rolling out invisible watermarking for ChatGPT and Codex-generated text in the EU, with an API opt-in available for developers. Internal testing shows editing can weaken the watermark, raising questions about its robustness. This is part of a broader wave of AI content provenance initiatives — Google also opened its SynthID watermark detector to the public this week, allowing anyone to check images, video, and audio for AI-generated content markers.

Google revealed its UGC Fresh Data Program — an API-based initiative allowing platforms like Reddit to push fresh, high-quality, and authentic user-generated content directly to Google Search for faster indexing and surfacing across search features. Search Engine Roundtable describes it as essentially an API that UGC platforms can use to accelerate content discovery. This has structural implications for how community-driven platforms are crawled and ranked, and may inform content strategies on UGC-heavy sites.

Google is rolling out Search Profiles — allowing creators and publishers to establish a verified presence directly in Search results — to users in India and Canada, with Brazil to follow shortly. Both Search Engine Journal and Search Engine Roundtable confirmed the expansion, which may signal an imminent global rollout.

In a notable SERP change going the other direction, Google has stopped displaying review rich snippets for healthcare web pages even when review schema or structured data is correctly implemented. Search Engine Roundtable reports the change appears to be deliberate, likely reflecting Google's ongoing sensitivity around YMYL health content. Healthcare SEOs relying on star ratings to boost CTR should audit their structured data strategy.

Two smaller but confirmed Google platform changes: Search Console will add dark mode support, as announced by a Senior Software Engineer at Search Central Live Deep Dive Europe 2026 in Barcelona. And Google is now emailing site owners with their 'preferred source' subscriber counts — the number of users who have explicitly followed or selected their domain as a preferred source in Google Search — hinting at a growing social/personalization layer within search.

Google updated its Merchant Center help documentation with a significant new policy: products that change more than once per day may be disapproved instead of updated if Google detects a data mismatch. This directly affects e-commerce sites with dynamic pricing, frequent inventory updates, or flash sales — scenarios where data feeds can lag behind real-time product states. Merchants and e-commerce SEOs should audit their feed update frequency and consider whether their current setup risks triggering this new disapproval logic.

Search Engine Journal detailed how Google's hidden data layer can cause outdated product images and misleading sale badges to appear in search results even after feed updates have been submitted. The piece offers diagnostic steps going beyond the current product feed — checking cached data, structured data rendering, and Google's internal data refresh cycles — to identify and resolve discrepancies that standard feed management would miss.

Google is now displaying WhatsApp contact buttons in local search panels, appearing alongside or in place of existing action buttons (call, directions, website, share). Search Engine Roundtable confirmed the new button is live, and it could influence local engagement metrics for businesses that use WhatsApp as a primary customer communication channel.

Separately, Google updated its Google Maps and Business Profiles prohibited content policy, specifically revising the advertising and solicitation section to explicitly list what merchants cannot post in listings and posts. Local SEO practitioners managing Google Business Profiles should review the updated guidelines to avoid penalties.

Two WordPress developments this week. First, WordPress 7.1.3 was released, patching seven security vulnerabilities including a critical bug that causes fatal errors during image uploads. Sites running older versions should update promptly to avoid potential downtime or exploitation.

Second, WordPress officially released a canonical MCP (Model Context Protocol) Adapter plugin, enabling AI agents to connect with WordPress sites while maintaining compatibility across MCP versions. As AI agents increasingly interact with web content for summarization and answer generation, this plugin positions WordPress sites to participate in those workflows — a forward-looking integration worth monitoring.