Key Takeaways
- Google released 4 confirmed updates in 2025 (vs. 9 in 2023) — fewer updates, higher per-update impact
- The December 2025 core update drove 40–70% overnight traffic drops; affiliate sites were hardest hit at 71%
- AI Overviews now appear for ~16% of all queries and reduce organic CTR by 58% (Ahrefs, December 2025)
- Being cited in an AI Overview delivers +35% more organic clicks — citation-worthiness is the new link building
- Recovery timelines: 2–6 months for e-commerce, 6–12 months for YMYL content
The Year Google Rewrote the Rules — Again
In 2022, Google ran 10 confirmed algorithm updates. In 2023, nine. In 2025? Four.
If you're reading that and thinking Google got less aggressive, you've misread the data entirely.
Fewer updates means each individual update carries more weight. The March 2026 core update — still rolling out as of this writing — has Semrush Sensor peaking at 9.5/10 volatility. The December 2025 update drove 67% of health content sites to suffer ranking drops. And layered on top of every core update, AI Overviews are systematically extracting organic traffic from even the sites that survived the algorithmic churn.
This article is a complete timeline of every confirmed Google algorithm update from early 2025 through the March 2026 core update, what each one targeted, who won, who lost, and — most importantly — what the pattern tells you about where Google is heading.
The 2025–2026 Update Timeline
Let's go chronologically. There's a logic to how these updates interconnect that only becomes visible when you see them in sequence.
March 2025 Core Update (March 13–27)
Google's 14-day March 2025 core update opened the year with volatility that Semrush's Mordy Oberstein described as "similar in size to the December 2024 core update — a mere .1 difference between the two when comparing volatility baselines." That might sound modest. It wasn't for the sites it hit.
The health sector saw the largest volatility spike of any vertical. Quora lost 22% visibility in the UK per SISTRIX data. Feature snippets underwent a quality overhaul — many previously featured sites lost their positions to newer, more comprehensive content.
Notable winners: flcgil.it (Italian trade union) gained +140.7% traffic (118,762 → 285,928 visits). Motors.co.uk and Dunelm.com saw significant UK gains. Teachmeanatomy.info recovered after being hit by the December 2024 spam update.
The pattern here: authoritative institutional content gained ground. Aggregator-style sites that thin-sliced content across broad topics continued losing.
June 2025 Core Update (June 30 – July 17)
The June update was, by most data providers' assessments, the most impactful individual core update of 2025. SISTRIX's Steve Paine described it as "one of the larger core updates seen in some time" with "at least two waves of effects across a wide range of domains and content types."
The critical development: this was the first update to deliver partial recoveries from the September 2023 Helpful Content Update — some sites saw their first significant ranking improvements in nearly two years. Google's Barry Schwartz noted the caveat that shaped recovery expectations for the rest of the year:
"Google's core systems need to learn and also confirm that the site on a whole level is now producing helpful, reliable people-first content — and it has to be in the long term."
Recovery was far from universal. Glenn Gabe (GSQi) documented extensively that many sites hit by the 2023 HCU did not recover in June 2025 — and some never will, because the content problems are structural, not fixable with incremental edits.
E-E-A-T also shifted category in this update. What had been a ranking factor became, in competitive niches, an effective indexing requirement. Content lacking demonstrable real-world experience was increasingly filtered out of competitive SERPs regardless of other signals.
August 2025 Spam Update (August 26 – September 22)
At 27 days, this was one of the longest spam update deployments in Google's history. Around September 9, a second volatility wave hit, extending an update that most sites had assumed was winding down.
Google's SpamBrain AI now detects 200x more spam sites than in previous versions. The primary targets remained consistent with the 2024 spam framework:
| Target | Definition | Detection Method |
|---|---|---|
| Scaled Content Abuse | Mass AI-generated pages with no quality value | Pattern analysis + engagement signals |
| Expired Domain Abuse | Aged domains repurposed for irrelevant content | Domain history vs. current content mismatch |
| Site Reputation Abuse | High-authority domains hosting third-party promotional content | Editorial independence signals |
| Content Duplication | Scraping and republishing without transformation | Near-duplicate detection |
Goong.com — identified by SISTRIX as having 27,443 ranking URLs — was removed from rankings entirely. The case made clear that SpamBrain can now identify and deindex sites with surgical precision, not just apply broad penalties.
Sites with ads-to-content ratios above 25% were disproportionately affected. Posts under 500 words consistently underperformed across the update window.
December 2025 Core Update (December 11–29)
Glenn Gabe's characterization of this update as "The Core Before Christmas" is memorable, but the data behind it isn't festive.
ALM Corp analyzed 847 affected websites and produced the most granular impact breakdown of any 2025 update:
| Content Type | Impact Rate |
|---|---|
| Mass-produced AI without expert oversight | 87% negative |
| Thin affiliate content lacking original testing | 71% traffic drops |
| Generic keyword-optimized content | 63% ranking losses |
| Completely unedited AI at scale | 85–95% traffic losses |
| E-commerce sites (broad) | 52% affected |
| Health content sites | 67% affected |
| Affiliate sites | 71% affected — hardest hit category |
The Discover impact was severe and largely underreported. Multiple publishers documented drops from 90,000+ daily Discover clicks to zero. Some saw 98% overnight declines. Approximately 9,000 UK-based sites were disrupted. This was not a minor edge case — for many content publishers, Discover traffic dwarfed search traffic, and the update effectively eliminated it.
Core Web Vitals became a ranking differentiator in measurable ways for the first time at scale: - Sites with LCP above 3 seconds: 23% more traffic loss than faster competitors - Sites with INP above 300ms: 31% drops, especially on mobile - Sites with CLS above 0.15: 19% more traffic loss
Conversely, sites with deep content clusters (10–15 high-quality supporting articles) saw +23% average visibility gains.
Glenn Gabe's summary quote from his post-update analysis captures the core logic: *"Great pages can't save a weak site, but weak pages can hurt a strong site."*
March 2026 Core Update (February 24 – March 27, Ongoing)
The most significant update yet in this cycle — and notable for a first: Google simultaneously rolled out the first-ever dedicated Google Discover Core Update, separate from the main search ranking changes.
Semrush Sensor peaked at 9.5/10 volatility. A single-day high of 8.7 was recorded during peak rollout. Per ALM Corp's mid-rollout analysis: 55% of monitored sites saw ranking changes within the first two weeks. Daily organic traffic drops of 20–35% are being reported in webmaster forums.
The March 2026 update explicitly targets "scaled content abuse" using what Google's documentation references as the Gemini 4.0 Semantic Filter — a mechanism that evaluates whether content demonstrates genuine domain expertise versus SEO-optimized text assembly.
HubSpot has reportedly lost 70–80% of organic traffic through this update cycle — a data point worth sitting with. HubSpot has a domain authority of 90+, a massive content team, and years of SEO investment. If they're losing at this scale, the update isn't primarily about technical SEO. It's about content substance.
Recovery timelines from Google's own guidance and Gabe's post-analysis: 2–6 months for standard e-commerce, 6–12 months for YMYL topics.
The AI Overviews Problem (Separate From Core Updates)
Core updates and AI Overviews are two separate threats that operate simultaneously. Understanding both is essential.
AI Overviews have expanded dramatically and continuously:
| Metric | Value | Source |
|---|---|---|
| AIO prevalence (Jan 2025) | 6.49% of queries | Semrush |
| AIO prevalence (Mar 2025) | 13.14% of queries | Semrush |
| AIO prevalence (mid-2025 steady state) | ~16% of queries | Semrush |
| AIO queries that are informational | 88.1% | Semrush |
| Organic CTR with AIO present | 0.61% vs 1.76% without | Seer Interactive, Sep 2025 |
| CTR drop when AIO present (top rank) | 58% lower | Ahrefs, Dec 2025 |
| Top organic link CTR drop when AIO appears | ~79% | Authoritas |
| Zero-click searches (2025) | 60% of all searches | Bain & Company |
| News sites monthly organic visits lost (mid-2024 → May 2025) | 600 million+ | Multiple publishers |
This is the structural shift underlying the volatility: Google is extracting value from content it trained on and reducing the traffic it passes back to publishers. Every core update that rewards "helpful content" simultaneously makes that content more likely to be absorbed into AI Overviews.
The counterintuitive finding: being cited *within* an AI Overview delivers +35% more organic clicks and +91% more paid clicks versus not being cited. The new SEO game is not just ranking — it's becoming the source Google cites.
What Google Is Consistently Rewarding
The pattern across all 2025–2026 updates is remarkably consistent:
### Content That Survives Updates - Genuine first-hand experience: Not just expertise claimed in an author bio, but demonstrated in content — specific product test results, original methodologies, real case study data - Topical depth over breadth: Sites covering a narrow subject comprehensively outperform generalist sites covering the same keyword - Content clusters: 10–15 supporting articles around each core topic; sites with these structures saw +23% visibility gains in December 2025 - Technical performance: LCP under 2 seconds, INP under 200ms, CLS under 0.15 — these are now measurable differentiators, not hygiene factors - AI-assisted content with editorial oversight: John Mueller (November 2025): "Our systems don't care if content is created by AI or humans. We care if it's helpful, accurate, and created to serve users rather than just manipulate search rankings."
### Content That Gets Hit - Mass AI generation without expert review (87% negative impact in December 2025) - Thin affiliate content without original product testing (71% drops) - "Fake freshness" — updating publish dates without substantively updating content - Expired domain repurposing - YMYL content without verifiable professional credentials
The E-E-A-T Shift You Can't Ignore
Experience, Expertise, Authoritativeness, Trustworthiness has been a Google ranking concept since the Quality Rater Guidelines update in 2018. What changed in 2025 is its application scope.
Pre-2025: E-E-A-T primarily mattered for YMYL (Your Money, Your Life) content — health, finance, legal, safety topics.
Post-June 2025: E-E-A-T requirements have expanded beyond YMYL to virtually all competitive queries. A review of kitchen equipment now requires demonstrable hands-on testing evidence. A tutorial on software configuration now benefits from author credentials showing they've actually used the tool professionally.
The practical implication: author bylines, author bios linking to portfolios, first-person experience signals in the content, and entity-level authority signals (brand mentions, authoritative backlinks) are now ranking factors in categories they previously weren't.
You can audit your backlink profile and authority signals with Backlynk's analyzer — start with your referring domain quality and anchor text distribution before addressing content E-E-A-T signals.
Recovery Framework: What the Data Supports
One site documented by ALM Corp enhanced its top 20 pages (2–3 hours per page adding depth, original data, and expert commentary) and recovered approximately 70% of lost traffic within 6 weeks following the December 2025 update. This is the fastest documented recovery in the post-update research.
The framework that emerges from that case and others:
### Phase 1: Diagnosis (Week 1–2) - Identify affected pages via Google Search Console (filter by date range around update) - Segment losses: are they in one content category or site-wide? - Site-wide losses suggest a domain-level quality signal problem; category losses suggest content quality in that topic cluster - Run Core Web Vitals assessment via PageSpeed Insights — address any LCP > 3s, INP > 300ms, CLS > 0.15 issues first
### Phase 2: Content Triage (Week 3–8) - Identify top 20 pages by pre-update traffic - For each: does it demonstrate actual experience with the subject? Does it contain data unavailable elsewhere? Is the author verifiably credentialed? - Pages failing all three criteria: rewrite substantially or consolidate into a stronger parent page - Pages passing two of three: upgrade the weakest signal first
### Phase 3: Authority Building (Month 2–6) - Build topical content clusters around each core topic (10–15 supporting articles minimum) - Acquire links from topically relevant domains — not just high-DA generalists - Submit to relevant industry directories to build citation footprint and branded search volume - Establish author entities: bylines, social profiles, professional portfolios linked from author bios
### Phase 4: Monitor and Iterate - Watch for the next core update (typically 2–5 months after each update) - Core updates are evaluation points, not finish lines — improvements made between updates are scored at the *next* update
How to Position for AI Overview Citations
Given that AI Overview citations now drive significantly more traffic than organic rankings for cited pages, this is the highest-leverage SEO activity in 2026.
Semrush's analysis found 88.1% of AI Overview queries are informational — meaning informational content is where the citation opportunity concentrates. Structure your informational content as the definitive reference:
- Define terms precisely in the first 100 words — AIOs frequently pull exact definitions
- Use structured data (FAQ schema, How-To schema, Article schema) — Semrush found structured markup correlates strongly with AIO citation likelihood
- Include named statistics with source attribution — AIOs favor citable, verifiable claims
- Write succinct, standalone-quality paragraphs — AIOs extract paragraphs, not entire articles
- Earn links from authoritative domains — AIO citation weighting correlates with backlink authority
You can audit your backlink profile to identify authority gaps that may be limiting both traditional rankings and AIO citation likelihood.
FAQ: Google Algorithm Updates
What was the biggest Google algorithm update of 2025?
By impact breadth, the June 2025 core update was described by SISTRIX's Steve Paine as "one of the larger core updates seen in some time" with multiple waves. By individual-site impact, the December 2025 update was more severe — driving 87% negative outcomes for mass AI content and 71% traffic drops for thin affiliate sites.
How long does it take to recover from a Google core update?
Recovery timelines from Google's documentation and independent analysis: 2–6 months for standard e-commerce sites that address content quality issues; 6–12 months for YMYL sites (health, finance, legal). Recovery only registers at the next core update — improving content between updates doesn't produce gradual recovery; it positions you for the next evaluation.
Does Google penalize AI-generated content?
Not categorically. John Mueller stated in November 2025: "Our systems don't care if content is created by AI or humans. We care if it's helpful, accurate, and created to serve users rather than just manipulate search rankings." What Google penalizes is mass-produced content without expert oversight — which describes most AI content at scale. Semrush's study of 20,000 articles found the penalization signal is quality and expertise absence, not AI involvement per se.
Are Core Web Vitals now a ranking factor?
Yes, with measurable impact documented for the first time at scale in December 2025. Sites with LCP above 3 seconds suffered 23% more traffic loss than faster peers. Sites with INP above 300ms saw 31% drops. Notably, 43% of sites still fail the 200ms INP threshold as of 2026, and only 55.7% of sites pass all three Core Web Vitals — meaning strong technical performance is still a genuine competitive differentiator.
What is the Helpful Content System and is it still separate?
The Helpful Content System (HCS) was introduced as a standalone signal in August 2022. By end of 2025, it was fully integrated into Google's core ranking logic — no longer a separate, identifiable signal. Recovery from HCS-related penalties now depends on demonstrating sustained content quality improvement across the full site, not just fixing individual pages.
What percentage of searches are zero-click in 2026?
Bain & Company's 2025 research found 60% of all Google searches end without a click. AI Overviews specifically: Pew Research Center found 26% of searches with AIOs ended with zero clicks (vs. 16% for traditional results). Zero-click share has risen from approximately 56% to 69% over the 2024–2025 period as AIOs expanded.
How do I know if my site was hit by a Google update?
Check Google Search Console for traffic drops correlating with confirmed update dates. Cross-reference with Google's official update announcements on the Google Search Status Dashboard and the Search Engine Land updates log. A site-wide traffic drop (all pages affected proportionally) suggests a domain-level signal. A category-specific drop (e.g., only product review pages) suggests content quality signals in that topic area.
Should I disavow backlinks after an algorithm update?
Only if you have a documented history of building spammy links and have received a manual action — not for algorithmic penalties. Google's systems are sophisticated enough to ignore low-quality links in most cases. Disavowing legitimate links you earned can harm your rankings. If you're uncertain about your link profile quality, audit your referring domains before submitting any disavow file.
Historical Google Algorithm Updates (2003–2024)
For SEO context, every major Google algorithm update from 2003 forward shaped the rules competitors play by today. This timeline summarizes the most consequential updates that still influence current ranking signals.
### 2003: Florida Update (November) The first widely-recognized core update. Targeted keyword stuffing and over-optimized commercial pages. Marked the start of Google's anti-manipulation arc. Many e-commerce sites disappeared overnight in the run-up to the 2003 holiday shopping season.
### 2005: Jagger Update Three-phase rollout (Sep–Nov 2005) targeting reciprocal link schemes, paid links, and low-quality directories. The first major update penalizing the "link exchange" tactic that had dominated SEO for the prior decade.
### 2009: Vince Update (February) Boosted brand authority signals — large recognized brands gained ranking advantages over small affiliates. Introduced the "brand bias" thesis still relevant in 2026 E-E-A-T signals.
### 2010: May Day + Caffeine May Day targeted long-tail informational queries from low-quality content farms. Caffeine (June 2010) was an indexing infrastructure overhaul, not a ranking update — enabled freshness signals to scale.
### 2011: Panda (February) The watershed content quality update. Targeted "content farms" (Demand Media, eHow, Suite101) and thin/duplicate content. Roughly 12% of search results changed. Panda became part of the core algorithm in January 2016. The Panda playbook (E-E-A-T, helpful content) is a direct ancestor of the December 2022 Helpful Content Update.
### 2012: Penguin (April) Targeted unnatural link profiles, especially exact-match anchor text and link networks. Sites that bought guest posts on low-quality blogs or built private blog networks (PBNs) lost rankings overnight. Penguin 4.0 (September 2016) made it real-time and granular (per-page rather than site-wide).
### 2013: Hummingbird (August) Complete algorithm rewrite enabling semantic / conversational query understanding. Marked the shift from exact-match keyword matching to intent matching. Foundation for natural language search and the later transformer models (RankBrain, BERT).
### 2014: Pigeon (July) Local search overhaul — tightened the link between local pack rankings and traditional organic ranking signals. Local SEO citations and Google Business Profile setup became core requirements.
### 2015: RankBrain (October) + Mobile-Friendly Update (April) RankBrain — Google's first machine learning ranking signal — moved beyond rule-based ranking to context understanding. Mobile-Friendly (April 2015) made mobile-responsive design a ranking signal; affected ~40% of queries.
### 2016: Possum (September) Local pack overhaul — diversified local results so a single business couldn't dominate multiple locations. Plus AdWords overhaul (4 ads above organic).
### 2017: Fred (March) Targeted thin, ad-heavy content. The original "Helpful Content" prototype.
### 2018: Mobile-First Indexing Google switched primary index from desktop to mobile versions of pages. Sites with mobile rendering issues lost rankings. Final phase rolled out 2020–2023.
### 2018: Medic Update (August) Focused on health/wellness/finance (YMYL — Your Money Your Life) categories. Required higher author expertise signals. The August 2018 update is the foundation of E-A-T (later E-E-A-T) doctrine.
### 2019: BERT (October) Bidirectional Encoder Representations from Transformers. Improved understanding of conversational queries (especially prepositional context). Affected ~10% of queries — large for a single update.
### 2020–2021: Multiple Core Updates + Page Experience Page Experience update (June 2021) added Core Web Vitals (LCP, FID, CLS) as ranking signals. Core updates in May 2020, December 2020, June 2021, July 2021, November 2021 reshuffled YMYL rankings.
### 2022: Helpful Content Update (August + December) Site-wide signal targeting "people-first content vs. SEO-first content." First version had narrow effect; December 2022 version was harsher. Originally a separate signal; integrated into core algorithm in March 2024.
### 2023: Spam Updates + Link Spam Update + 4 Core Updates 2023 saw 9 confirmed updates including March, August, October, November core updates plus Reviews update consolidations. Link Spam Update (December 2022 + ongoing) targeted PBNs, sponsored posts without rel="sponsored", and link-exchange schemes.
### 2024: March 2024 Core Update + Helpful Content Integration The March 2024 core update was a 45-day rollout (longest ever) that integrated Helpful Content into core algorithm. ~40% of low-quality results were removed. Sites recovering from earlier Helpful Content penalties got a second chance.
### 2024: November 2024 Core Update Last update of the heavy 2024 cycle. Set the stage for the AI Overviews focus that dominated 2025.
For SEO recovery from any historical update, the universal recipe applies: audit content quality (Helpful Content lens), audit backlink profile (Penguin lens), audit technical health (Page Experience), audit topical authority (BERT/MUM understanding). Run a full backlink audit as a starting point.
*Google's algorithm is increasingly rewarding sites with strong authority signals — broad citation footprints, topically relevant links, and high-quality referring domains. If your site lost traffic in a recent update, the recovery path starts with understanding your link profile. Analyze your backlinks to identify gaps, then plan vetted directory submissions to rebuild a cleaner authority foundation.*