There is no shortage of Google algorithm update timelines. What is genuinely scarce is a timeline that gets the classifications right.
Most published lists put "Medic," "Fred," "Phantom," and the March 2024 core update in the same column, as though they were the same kind of event. They aren't. Two of those are nicknames the SEO industry invented for ranking volatility it noticed but Google never named. One is a documented, dated, officially announced rollout with a published start and end. Treating them as equivalent is how teams end up rewriting content to recover from an "update" that was actually a manual policy enforcement action, or waiting out a "core update" that was really an indexing infrastructure change.
This guide separates them. It covers the named Google algorithm updates from Boston in February 2003 through the August 2026 spam update, sorted into seven classes, with rollout durations for the modern confirmed era. It also does something the standard timeline doesn't: it analyses the rollout data itself, because how long an update takes to deploy is a diagnostic signal most SEOs never use.
What counts as a Google algorithm update — and what doesn't
Before any timeline is useful, you need a taxonomy. Google runs thousands of changes to Search every year and announces only a small fraction of them. The things that get announced fall into distinct categories that call for completely different responses.
A ranking system is a persistent component of how Google ranks — RankBrain, BERT, PageRank, the spam-detection system, the reviews system. Systems run continuously.
An update is a change to one or more of those systems, deployed over a defined window with a start and completion date.
A core update is a broad revision to Google's overall ranking systems, affecting relevance and quality assessment across all types of queries. Core updates are not penalties. Nothing is "wrong" with a page that drops during one; Google has simply revised how it weighs usefulness.
A spam update improves the systems that detect and neutralise manipulative practices — scaled content abuse, link schemes, expired domain abuse, cloaking. Unlike core updates, spam updates do imply something is wrong, and the fix is policy compliance rather than general quality improvement.
A policy or enforcement change alters the written rules or how they're applied, sometimes through manual action rather than algorithmic ranking. The site reputation abuse enforcement of May 2024 is the clearest example: a policy applied initially by manual review, not an algorithm rollout.
An indexing or infrastructure change affects crawling, indexing, or serving rather than ranking directly. Caffeine in 2010 and Big Daddy in 2006 belong here, and confusing them with ranking updates has produced a great deal of bad advice over the years.
An industry-named update is a label the SEO community attached to observed volatility. Sometimes Google later confirmed a change occurred; often it confirmed only that "we make changes all the time." These names have real diagnostic value as shorthand, but they carry no official documentation, no confirmed date range, and no published guidance.
A ranking incident is a service issue Google logs on the same dashboard, under the same Ranking heading, titled "Ranking is experiencing an ongoing issue." It has a start date and a duration but no update name and no guidance, because nothing was deliberately shipped. This is the class almost every published timeline omits, and omitting it is expensive: both recorded incidents overlapped a named rollout, so anyone diagnosing traffic in those windows against the named update alone was working from an incomplete picture.
Keep those seven apart and the history below becomes far more usable. Collapse them and you get a list of names with no actionable meaning.
What rollout duration actually tells you
Here is the part almost every timeline leaves out. Google publishes how long each update took to deploy, and that number carries information most SEOs never extract from it. The full year-by-year tables follow further down; the analysis comes first because it is the part that changes what you do.
Across the 16 confirmed core rollouts from November 2021 through May 2026, the numbers look like this:
| Metric | Core updates | Spam updates |
| Rollouts analysed | 16 | 10 |
| Mean duration | 17.4 days | 8.7 days |
| Median duration | 14.5 days | 7.1 days |
| Shortest | 6 days, 4 hours (Dec 2024) | 19 hours, 30 min (Mar 2026) |
| Longest | 45 days (Mar 2024) | 26 days, 15 hours (Aug 2025) |
Strip out the March 2024 outlier and the core mean falls to 15.5 days with a median of 14 days. Two weeks is the realistic baseline for a core update. That single fact should govern your reporting cadence: if you draw conclusions from a five-day window, you are analysing a partially deployed system.
The trend over time is more interesting than the average.
Core updates are deploying faster. Grouped by era, mean core rollout duration ran 15.9 days across 2021–2023, rose to 20.4 days across 2024–2025, then dropped to 12.0 days in 2026. The March 2026 rollout (12 days, 4 hours) and the May 2026 rollout (11 days, 21 hours) are the two fastest core deployments since the anomalous December 2024 update.
Spam updates have collapsed in duration. This is the sharper signal. Spam rollouts averaged 8.5 days in 2021–2023 and 13.9 days in 2024–2025. In 2026 they average 1.8 days. The March 2026 spam update completed in 19 hours and 30 minutes — the shortest confirmed rollout anywhere in the dataset.
What should you infer from that? Read it as an operational signal rather than a ranking signal. Short rollouts are consistent with deploying a retrained model rather than rebuilding scoring across the index. A sub-24-hour spam deployment suggests a system tuned for rapid iteration — which in practice means the window between publishing manipulative content and getting caught is narrowing from weeks to days.
There is a practical consequence. The old playbook of watching a spam update roll out over two or three weeks and adjusting mid-flight no longer applies. By the time volatility trackers register a 2026 spam update, it has usually finished deploying.
Two cautions on reading duration. A longer rollout does not mean a larger impact — the December 2024 core update ran six days and still moved rankings meaningfully. And nothing about duration tells you how many sites were affected. Duration describes deployment mechanics, not blast radius.
The consolidation pattern: fewer names, broader systems
Count the named updates per year and a clear arc appears.
| Year | Named updates | Distinct update families |
| 2021 | 11 | Core, spam, link spam, reviews, page experience, local, passage ranking |
| 2022 | 10 | Core, spam, link spam, helpful content, reviews, page experience |
| 2023 | 9 | Core, spam, helpful content, reviews |
| 2024 | 7 | Core, spam |
| 2025 | 4 | Core, spam |
| 2026 (to 29 Aug) | 6 | Core, spam, Discover core |
From eleven named updates across seven families in 2021 to four across two families in 2025. That is not a slowdown in engineering. It is consolidation: Google progressively folded specialised systems into its core ranking systems and stopped announcing them separately.
The absorption history is explicit:
- Penguin became part of core ranking systems in 2016 and has run in real time since.
- Panda was integrated into core ranking systems and stopped being versioned.
- The helpful content system was folded into core in March 2024.
- The reviews system moved to continuous, unannounced operation after November 2023.
- Page experience stopped being a named update family after the desktop rollout in February 2022.
The strategic implication is significant and widely misread. When a specialised system is absorbed into core, you no longer get a discrete recovery event. Under the standalone helpful content system, a site could be affected by a specific update and recover at the next refresh of that specific system. Once the system sits inside core, recovery arrives — if it arrives — at the next broad core update, and it is assessed alongside everything else. Recovery windows got longer and less legible at exactly the moment the update calendar got quieter.
The 2026 uptick to six updates in under eight months, already more than all of 2025, suggests the trough may have been temporary. But the composition has not reverted: it is still only core, spam, and a Discover-specific rollout. Nobody should expect the six-family calendar of 2022 to return.
The confirmed era: named updates since late 2021
Google's published ranking-updates history begins in late 2021. Everything from that point forward has an official name, a confirmed start date, and a recorded rollout duration. This is the only portion of the timeline where the data is precise enough to analyse quantitatively, and it is the data the analysis above is built on. Every row below was verified against Google's Search Status Dashboard on 29 August 2026.
2026 confirmed updates (through 29 August 2026)
| Start date | Update | Type | Rollout duration |
| 18 Aug 2026 | August 2026 Spam Update | Spam | 2 days, 16 hours |
| 24 Jun 2026 | June 2026 Spam Update | Spam | 2 days, 1 hour |
| 21 May 2026 | May 2026 Core Update | Core | 11 days, 21 hours |
| 27 Mar 2026 | March 2026 Core Update | Core | 12 days, 4 hours |
| 24 Mar 2026 | March 2026 Spam Update | Spam | 19 hours, 30 minutes |
| 5 Feb 2026 | February 2026 Discover Update | Discover core | 21 days, 17 hours |
Two things about 2026 deserve attention. First, the February update targeted Google Discover, not general web Search — a different surface with different behaviour, and it should never be pooled with web core updates when you're analysing traffic. Second, the balance has shifted. In every year from 2022 onward, core updates equalled or outnumbered spam updates. 2026 inverts that: three spam updates against two web core updates.
2025 confirmed updates
| Start date | Update | Type | Rollout duration |
| 11 Dec 2025 | December 2025 Core Update | Core | 18 days, 2 hours |
| 26 Aug 2025 | August 2025 Spam Update | Spam | 26 days, 15 hours |
| 30 Jun 2025 | June 2025 Core Update | Core | 16 days, 18 hours |
| 13 Mar 2025 | March 2025 Core Update | Core | 13 days, 21 hours |
2025 is the quietest year on record for named updates — four in total, and only two categories. That is not because Google stopped changing Search. It is because the specialised named systems had by then been folded into core.
2024 confirmed updates
| Start date | Update | Type | Rollout duration | Ran concurrently with |
| 19 Dec 2024 | December 2024 Spam Update | Spam | 7 days, 2 hours | — |
| 12 Dec 2024 | December 2024 Core Update | Core | 6 days, 4 hours | — |
| 11 Nov 2024 | November 2024 Core Update | Core | 23 days, 13 hours | — |
| 15 Aug 2024 | August 2024 Core Update | Core | 19 days, 4 hours | Ranking incident (4 d 11 h) |
| 20 Jun 2024 | June 2024 Spam Update | Spam | 7 days, 1 hour | — |
| 5 Mar 2024 | March 2024 Spam Update | Spam | 14 days, 21 hours | March 2024 Core Update |
| 5 Mar 2024 | March 2024 Core Update | Core | 45 days | March 2024 Spam Update |
Policy and enforcement changes in 2024 — stored separately, deliberately:
| Date | Change | Correct classification |
| 6 May 2024 | Site reputation abuse enforcement begins | Manual / policy enforcement |
| 31 Jul 2024 | Ranking changes for explicit fake and deepfake content | Specialised ranking change |
| 19 Nov 2024 | Site reputation abuse policy expanded | Spam policy / enforcement |
March 2024 is the most consequential entry in the entire modern table. Google announced it alongside a spam update on the same day and used it to absorb the standalone helpful content system into its core ranking systems. That is why the rollout ran 45 days — the longest in the confirmed record by a wide margin. It also explains why "helpful content update" stopped appearing as a separate line item afterwards.
2023 confirmed updates
| Start date | Update | Type | Rollout duration | Ran concurrently with |
| 8 Nov 2023 | November 2023 Reviews Update | Reviews | 29 days | November 2023 Core Update |
| 2 Nov 2023 | November 2023 Core Update | Core | 25 days, 21 hours | November 2023 Reviews Update |
| 5 Oct 2023 | October 2023 Core Update | Core | 13 days, 23 hours | Oct 2023 Spam Update + ranking incident |
| 4 Oct 2023 | October 2023 Spam Update | Spam | 15 days, 12 hours | Oct 2023 Core Update + ranking incident |
| 14 Sep 2023 | September 2023 Helpful Content Update | Helpful content | 13 days, 11 hours | — |
| 22 Aug 2023 | August 2023 Core Update | Core | 16 days, 3 hours | — |
| 12 Apr 2023 | April 2023 Reviews Update | Reviews | 13 days, 2 hours | — |
| 15 Mar 2023 | March 2023 Core Update | Core | 13 days, 7 hours | — |
| 21 Feb 2023 | February 2023 Product Reviews Update | Product reviews | 14 days | — |
The November 2023 reviews update was the last separately announced reviews update. Google subsequently moved reviews improvements onto a continuous, unannounced schedule. If you publish review content and you're waiting for the next reviews update to recover, you're waiting for something that no longer exists as a discrete event.
Note also the October 2023 overlap: a spam update starting 4 October and a core update starting 5 October, running concurrently. Overlapping rollouts make attribution genuinely difficult, and any analysis that assigns a single cause to traffic movement in that window is guessing.
2022 confirmed updates
| Start date | Update | Type | Rollout duration | Ran concurrently with |
| 14 Dec 2022 | December 2022 Link Spam Update | Link spam | 29 days | December 2022 Helpful Content Update |
| 5 Dec 2022 | December 2022 Helpful Content Update | Helpful content | 38 days | December 2022 Link Spam Update |
| 19 Oct 2022 | October 2022 Spam Update | Spam | 2 days | — |
| 20 Sep 2022 | September 2022 Product Reviews Update | Reviews | 6 days | September 2022 Core Update |
| 12 Sep 2022 | September 2022 Core Update | Core | 14 days | September 2022 Product Reviews Update |
| 25 Aug 2022 | August 2022 Helpful Content Update | Helpful content | 15 days | — |
| 27 Jul 2022 | July 2022 Product Reviews Update | Reviews | 6 days | — |
| 25 May 2022 | May 2022 Core Update | Core | 15 days | — |
| 23 Mar 2022 | March 2022 Product Reviews Update | Reviews | 14 days | — |
| 22 Feb 2022 | Desktop Page Experience Update | Page experience | 9 days | — |
2022 was the year of maximum fragmentation: six distinct update families in twelve months. The August 2022 helpful content update introduced the standalone helpful content system — the one that would be absorbed into core nineteen months later.
2021 confirmed updates and other major 2021 changes
| Start date | Update | Type | Rollout duration |
| 1 Dec 2021 | December 2021 Product Reviews Update | Reviews | 20 days |
| 17 Nov 2021 | November 2021 Core Update | Core | 13 days |
| 3 Nov 2021 | November 2021 Spam Update | Spam | 8 days, 1 hour |
Earlier 2021 changes, which predate the published rollout-duration record:
| Date | Update / change |
| Nov 2021 | November Local Search Update |
| Jul 2021 | Link Spam Update |
| Jul 2021 | July Core Update |
| Jun 2021 | June Spam Updates |
| Jun 2021 | Page Experience / Core Web Vitals rollout |
| Jun 2021 | June Core Update |
| Apr 2021 | Product Reviews Update |
| Feb 2021 | Passage Ranking |
Ranking incidents are not updates
This section exists because, as far as a survey of the ranking pages for this topic could establish, no other published timeline includes it.
Google's Search Status Dashboard records more than named updates. Under the same Ranking heading it also logs service issues, titled "Ranking is experiencing an ongoing issue." There are two in the confirmed record.
| Start date | Dashboard entry | Duration | Overlapped |
| 15 Aug 2024 | Ranking is experiencing an ongoing issue | 4 days, 11 hours | August 2024 Core Update — same start date |
| 5 Oct 2023 | Ranking is experiencing an ongoing issue | 26 days | October 2023 Core Update and October 2023 Spam Update |
Neither has an update name, neither came with guidance, and neither was something Google deliberately shipped. They were malfunctions.
The October 2023 entry deserves particular attention. For roughly two weeks that month, three separate dashboard entries were live at once: a spam update, a core update, and a ranking incident. Every retrospective that attributes an October 2023 traffic movement to "the October core update" is asserting more than the record supports. There were three candidate causes and no way to separate them.
The practical rule follows directly. When you check a traffic drop against the update calendar, check the incident log in the same pass. A four-day unexplained dip in August 2024 was far more likely to be the incident than the core update that started the same day, and the correct response to an incident is to wait, not to rewrite anything.
A note on concurrency
Thirteen confirmed entries overlapped at least one other. The tables above mark them, and the pattern is worth internalising:
- October 2023 — three concurrent entries, roughly 14 days of overlap
- December 2022 — link spam and helpful content updates overlapping for about 29 days
- November 2023 — core and reviews updates overlapping for about 20 days
- March 2024 — core and spam updates launched the same day, overlapping about 15 days
- September 2022 — core and product reviews updates overlapping about 6 days
- August 2024 — core update and a ranking incident overlapping about 4.5 days
When rollouts overlap, single-cause attribution is not available. The honest move in a client report is to say so rather than pick the more interesting explanation. Overlap lengths here are derived from published start dates and durations; Google publishes dates but not start times, so treat each as approximate to within a day.
The historical era: 2003–2020, organised by family
Chronological lists of pre-2021 updates are hard to learn from, because a date-ordered sequence hides the fact that most of these updates belong to a handful of long-running programmes. Grouped by family, the history explains itself.
What Google actually said about impact
Modern updates come with no impact figures at all. Older ones often did, and those numbers are useful context that most timelines drop. Each of these is a figure Google itself published at the time.
| Update | Google's stated impact |
| Panda (Feb 2011) | 11.8% of US search results |
| Freshness (Nov 2011) | 35% of searches |
| Penguin (Apr 2012) | ~3% of search queries |
| Exact Match Domain (Sep 2012) | 0.6% of English-US queries |
| Panda 4.0 (May 2014) | 7.5% of English queries |
| Panda 4.2 (Jul 2015) | 2–3% of English-language queries |
| BERT (Oct 2019) | ~10% of English queries in the US at launch |
| Passage ranking (Feb 2021) | ~7% of queries across all languages |
Two things stand out. First, the numbers are far smaller than the industry reaction to each update suggested — Penguin reshaped link building for a decade while touching about three percent of queries. Second, Google stopped publishing them. No core update since 2021 has come with a percentage, which is part of why rollout duration has become the only quantitative handle available on modern updates.
The quality family: Panda and its descendants
Panda launched in February 2011 and ran as a versioned, periodically refreshed filter targeting thin, duplicated, and low-value content.
| Date | Update | Date | Update |
| Feb 2011 | Original Panda | Apr 2012 | Panda 3.5–3.6 |
| Apr 2011 | Panda 2.0 | Jun 2012 | Panda 3.7–3.8 |
| May 2011 | Panda 2.1 | Jul 2012 | Panda 3.9 |
| Jun 2011 | Panda 2.2 | Sep 2012 | Panda 20 |
| Jul 2011 | Panda 2.3 | Nov 2012 | Panda 21 |
| Aug 2011 | Panda 2.4 | Nov 2012 | Panda 22 |
| Sep 2011 | Panda 2.5 | Dec 2012 | Panda 23 |
| Oct 2011 | Panda 3.0 | Mar 2013 | Panda update |
| Nov 2011 | Panda 3.1 | Jul 2013 | Panda update |
| Jan 2012 | Panda 3.2 | May 2014 | Panda 4.0 |
| Feb 2012 | Panda 3.3 | Sep 2014 | Panda 4.1 |
| Mar 2012 | Panda 3.4 | Jul 2015 | Panda 4.2 |
Twenty-four dated Panda refreshes in four and a half years, and the rows covering version ranges account for more than one release each. The versioning stopped when Panda was absorbed into core ranking systems, and the underlying question it asked — is this page genuinely useful to the person who landed on it — was inherited by the helpful content system and then by core itself. The lineage runs unbroken from Panda 1.0 in 2011 to the quality assessment inside the May 2026 core update.
Also in this family: MayDay (April/May 2010), which reshaped long-tail query results, and the Quality Update, nicknamed Phantom by the industry, in May 2015.
The link family: Penguin and link spam
| Date | Update |
| Oct 2007 | Paid links / PageRank enforcement changes |
| 24 Apr 2012 | Original Penguin |
| May 2012 | Penguin refresh |
| Oct 2012 | Penguin refresh |
| May 2013 | Penguin 2.0 |
| Oct 2013 | Penguin 2.1 |
| Oct 2014 | Penguin 3.0 |
| Nov 2014 | Penguin 3.1 |
| Dec 2014 | Penguin Everflux |
| 23 Sep 2016 | Penguin 4.0 |
| Jul 2021 | Link Spam Update |
| Dec 2022 | December 2022 Link Spam Update |
Penguin 4.0 in September 2016 is the pivot point. Google made Penguin granular and real-time and integrated it into core ranking systems. Before that, a site hit by Penguin waited months for the next refresh even after cleaning up its link profile. Afterwards, devaluation and re-evaluation happened continuously.
The subsequent link spam updates in 2021 and 2022 are a different mechanism — spam-detection systems neutralising manipulative links at scale rather than applying a site-wide filter. Practically, this shifted the response from "disavow aggressively" to "stop acquiring links that will simply be ignored."
The local family
| Date | Update | Status |
| Feb 2012 | Venice | Localised organic results |
| 24 Jul 2014 | Pigeon | Confirmed local ranking change |
| Dec 2014 | Pigeon expansion | Confirmed |
| Sep 2016 | Possum | Industry-named |
| Aug 2017 | Hawk | Industry-named |
| Nov 2021 | November Local Search Update | Confirmed |
Local ranking runs on its own systems, weighing relevance, distance, and prominence. That is why a business can lose map pack visibility during a period of no core update activity at all, and why local performance should be tracked on a separate line from organic web performance.
The language-understanding family
This is the most consequential family for anyone writing content today.
| Date | System | What it changed |
| Aug/Sep 2013 | Hummingbird | Shifted ranking toward query meaning and intent rather than keyword matching |
| Oct 2015 | RankBrain | Machine-learning interpretation of unfamiliar queries and concepts |
| 25 Oct 2019 | BERT | Contextual language understanding, initially a fraction of English queries |
| Oct 2020 | BERT expansion | Extended to nearly all English queries |
| Feb 2021 | Passage ranking | Ranking of individual passages within a longer page |
These are systems, not updates — they were introduced once and have run since. Together they are the reason exact-match keyword density stopped working: Google stopped matching strings and started matching meaning. Passage ranking is the underrated one. It means a well-structured section deep inside a long article can rank on its own, which is a direct argument for clear heading hierarchy and self-contained sections — the same structure that makes content extractable by AI answer engines.
The spam, legal, and layout family
| Date | Update | Target |
| Jan 2012 | Page Layout / Top Heavy | Excessive above-the-fold advertising |
| Aug 2012 | Pirate / DMCA | Sites with high copyright removal request volumes |
| 28 Sep 2012 | Exact Match Domain (EMD) | Low-quality exact-match domains |
| Oct 2012 | Page Layout / Top Heavy | Ad-heavy layouts |
| Jun 2013 | Payday Loan | Heavily spammed query verticals |
| Feb 2014 | Page Layout refresh | Ad-heavy layouts |
| May 2014 | Payday Loan 2.0 | Spammy sites |
| Jun 2014 | Payday Loan 3.0 | Spammy queries |
| Oct 2014 | Pirate 2.0 | Copyright infringement |
| Jan 2011 | Scraper-related changes | Content scraping |
The mobile and experience family
| Date | Update |
| 21 Apr 2015 | Mobile-Friendly Update ("Mobilegeddon") |
| Apr 2016 | Mobile-Friendly Update #2 |
| Jun 2021 | Page Experience / Core Web Vitals rollout |
| 22 Feb 2022 | Desktop Page Experience Update |
Infrastructure, not ranking
| Date | Change | What it actually was |
| Dec 2005 – Mar 2006 | Big Daddy | Crawling and indexing infrastructure overhaul |
| 8 Jun 2010 | Caffeine | Indexing infrastructure rebuild enabling far fresher results |
Both are routinely listed as ranking updates. Neither was. Caffeine rebuilt how Google indexed the web so that new content could surface in minutes rather than days — enormous in effect, but the mechanism was indexing throughput, not ranking logic. Filing them correctly matters, because "my rankings dropped" and "my pages aren't getting indexed" are different problems with different fixes.
The early years: 2003–2010
| Date | Update | Date | Update |
| Feb 2003 | Boston | May/Jun 2005 | Bourbon |
| Apr 2003 | Cassandra | Sep 2005 | Gilligan |
| May 2003 | Dominic | Oct/Nov 2005 | Jagger |
| Jun 2003 | Esmeralda | Dec 2005 | Big Daddy begins |
| Jul 2003 | Fritz | Oct 2007 | Paid links enforcement |
| Nov 2003 | Florida | Mar/Apr 2008 | Dewey |
| Jan 2004 | Austin | Feb 2009 | Vince |
| Feb 2004 | Brandy | Aug 2010 | Brand-related update |
| Feb 2005 | Allegra | Dec 2010 | Merchant quality / negative reviews |
Florida in November 2003 is the one that still gets referenced, because it landed weeks before the holiday shopping season and eliminated a generation of keyword-stuffing tactics overnight. Vince in 2009 is the other notable entry: it shifted weight toward established brands for competitive commercial queries, and the underlying preference for recognisable, trusted entities has never gone away — it just became a component of quality assessment rather than a discrete update.
Treat the exact dates in this early section as approximate. Pre-2011 update naming and dating were reconstructed by the SEO community from observed volatility, not published by Google.
2016–2020 confirmed and industry-named updates
| Date | Update | Note |
| Jan 2016 | Core ranking update | |
| Mar 2017 | Fred | Industry nickname for confirmed ranking changes |
| Dec 2017 | Mid-December ranking/relevancy changes | |
| Mar 2018 | March broad core update | |
| Apr 2018 | April broad core update | |
| 1 Aug 2018 | August broad core update | Industry nickname: "Medic" |
| Mar 2019 | March 2019 core update | |
| 3 Jun 2019 | June 2019 core update | |
| 24 Sep 2019 | September 2019 core update | |
| 13 Jan 2020 | January 2020 core update | |
| 4 May 2020 | May 2020 core update | |
| 3 Dec 2020 | December 2020 core update |
"Medic" is worth a specific correction. Google never named an algorithm "Medic." The nickname came from the SEO industry's observation that health and medical sites appeared disproportionately affected by the August 2018 broad core update. The label stuck, and with it a persistent myth that Google runs a health-specific algorithm. It does not. What it does run is quality assessment that weighs expertise more heavily for topics where bad information causes real harm — which produces the same visible pattern through a different mechanism.
Ranking systems currently worth tracking
Updates come and go. Systems persist. These are the components actually running.
| System / family | Primary role | SEO importance |
| Core ranking systems | Broad relevance, quality, and usefulness assessment | Critical |
| Spam systems (SpamBrain) | Detecting and neutralising spam and manipulation | Critical |
| Reviews system | Rewarding substantive, original review content | High for review content |
| RankBrain | Machine-learning interpretation of queries and concepts | High |
| BERT | Language and context understanding | High |
| Neural matching | Concept matching between queries and pages | High |
| Passage ranking | Ranking relevant passages within longer pages | High |
| PageRank / link analysis | Link-based authority and relevance signals | High |
| Freshness systems | Prioritising recent content where the query benefits | Important |
| Exact-match domain systems | Preventing undue advantage for low-quality EMDs | Important |
| Original content systems | Rewarding original reporting | High |
| Local ranking systems | Relevance, distance, prominence | Critical for local SEO |
And the ones that have been retired or absorbed:
| System | Current status |
| Panda | Integrated into core ranking systems |
| Penguin | Integrated into core; real-time since 2016 |
| Hummingbird | Historical major ranking-system improvement |
| Standalone helpful content system | Integrated into core ranking systems, March 2024 |
If a strategy document you're working from still describes "recovering from Panda" or "waiting for the next Penguin refresh," it predates 2016 in its thinking, whatever date is on the cover.
How to diagnose whether an update hit your site
Most update diagnoses go wrong in the first five minutes, because someone sees a traffic drop, checks a volatility tracker, sees red, and concludes an update caused it. Work the sequence instead.
Step 1 — Establish the exact date and time of the change. Use Search Console's Performance report with a day-level view. Pinpoint the day clicks and impressions diverged. A drop that starts mid-week and mid-day is often a tracking, tagging, or site issue rather than an update.
Step 2 — Check whether an update was actually running. Compare your date against Google's published ranking-updates history. A confirmed update has a documented start date and a documented completion date. If your drop began outside every announced rollout window, an algorithm update is not your first hypothesis.
Step 3 — Rule out the non-algorithmic explanations first. In rough order of frequency: a site migration or redesign, a robots.txt or noindex change, a server or CDN issue, an analytics tracking break, seasonality, a manual action notice in Search Console, or a competitor's change rather than your own. Manual actions in particular appear in Search Console and require a reconsideration request, not content improvement.
Step 4 — Separate the surfaces. Web Search, Discover, Google News, image results, and the local map pack move independently and are governed by different systems. The February 2026 Discover update is a clean example: a site heavily dependent on Discover traffic could see a substantial swing while its web Search performance stayed flat. Segment before you diagnose.
Step 5 — Wait for the rollout to complete. This is where the duration data becomes practical. With a median core rollout of 14.5 days, a report written on day four describes a system mid-deployment. Rankings genuinely do fluctuate and partially reverse during rollouts. Google's own guidance is to wait for completion before assessing.
Step 6 — Classify what hit you. Core and spam updates require completely different responses:
| Signal | Points to a core update | Points to a spam update |
| Pattern of change | Broad, gradual shifts across many pages | Sharper drops, often to specific sections or page types |
| Content affected | Whole topics or site areas | Pages using a tactic — thin scaled content, purchased links, expired domain reuse |
| What Google says | No penalty; nothing is "wrong" | A spam policy has been violated |
| Correct response | Improve genuine usefulness and quality | Remove the violating practice and comply with policy |
| Recovery timing | Typically at a subsequent core update | Once the violation is removed and re-crawled |
Step 7 — Analyse at the query and page level, not the sitewide level. Sitewide traffic is an average, and averages hide the actual story. Export query-level data for the period before and after the rollout and look for which intents moved. A site can lose informational queries while holding transactional ones — a completely different diagnosis from losing everything evenly.
Recovery paths by update type
| What happened | What to actually do | Realistic timeline |
| Core update decline | Audit content against Google's self-assessment questions. Improve depth, originality, and first-hand experience. Fix genuine quality problems rather than tweaking on-page elements. | Reassessment typically comes with a subsequent core update. There is no guaranteed recovery. |
| Spam update decline | Identify the specific policy violated — scaled content abuse, link spam, expired domain abuse, site reputation abuse. Remove it completely. | Once removed and re-crawled, though full re-evaluation can take a further update cycle. |
| Manual action | Fix the violation, then file a reconsideration request in Search Console. | Depends on review turnaround. This is the only category where you can directly petition Google. |
| Site reputation abuse | Remove or fully separate third-party content published to exploit the host domain's reputation. | Enforcement has been both manual and policy-driven; treat it as a compliance issue, not a ranking one. |
| Reviews-related decline | Add genuine first-hand evaluation evidence: original testing, comparisons, measurements, images. | Continuous since 2023 — there is no announced update to wait for. |
| Discover decline | Reassess against Discover-specific content policies. Do not apply web Search core-update thinking. | Discover traffic is inherently volatile; a single update is not always the explanation. |
One honest note that most recovery guides avoid: Google states plainly that a decline during a core update does not mean a page violated anything, and that there is no guaranteed path back. Some sites drop because competing content improved, not because theirs got worse. Ranking is relative.
Common mistakes when working with update history
Treating volatility trackers as confirmation. Third-party sensors measure SERP fluctuation. Fluctuation happens constantly, including on days with no update at all. A spike is a prompt to investigate, never evidence of a named update.
Reacting during the rollout. Making significant changes on day three of a fourteen-day rollout means you will never know whether the recovery came from your change or from the rollout completing. Diagnose during, act after.
Applying core-update advice to a spam problem. Rewriting content will not resolve a link spam issue. Cleaning links will not resolve thin scaled content. Misclassification wastes months.
Chasing industry nicknames. "Fred," "Phantom," "Possum," "Hawk," and "Medic" are useful shorthand among practitioners. They have no official documentation and no confirmed rollout windows. Building a remediation plan around them means building on inference.
Ignoring the surface distinction. Pooling Discover, News, image, local, and web Search traffic into one number guarantees a wrong diagnosis eventually.
Assuming quiet means stable. 2025 had four announced updates. It was not a quiet year for ranking changes — it was a quiet year for announcements. Systems absorbed into core change continuously without a named rollout.
Waiting for a specific update family to return. There are no more standalone reviews updates. There is no standalone helpful content system. Waiting for either is waiting for something that has been retired.
Build your own update-tracking database
If you manage SEO across multiple sites, keep this data in a structured format rather than a folder of blog bookmarks. The following schema captures what actually matters at diagnosis time.
| Field | Guidance |
| Year | 2026 |
| Start date | ISO format: 2026-08-18 |
| Completion date | 2026-08-21 |
| Update name | August 2026 Spam Update |
| Update type | Core / Spam / Reviews / Policy / Indexing / Industry-named |
| Status | Completed / In progress |
| Rollout duration | 2 days, 16 hours |
| Google confirmed? | Yes / No |
| Official or industry name | Official / Industry |
| Search surface | Web Search / Discover / News / Images / Local |
| Location scope | Global, or specific regions if stated |
| Language scope | As specified by Google |
| Primary target | Search spam, content quality, links, reviews |
| Content impact | High / Medium / Low |
| Link impact | High / Medium / Low |
| Local SEO impact | High / Medium / Low |
| Technical SEO impact | High / Medium / Low |
| Recovery guidance | Link to the relevant Google documentation |
| Official Google source | Ranking updates history / Search Central |
| Secondary source | Reputable SEO publication, context only |
| Verification level | Google-confirmed / Industry-observed |
| SEO notes | Your own analysis and client impact |
| Action required | Audit / Monitor / None |
Two fields carry more weight than the rest. Verification level stops speculation from hardening into fact over time — a claim recorded as "industry-observed" in 2026 should still read as industry-observed in 2029. Search surface prevents the single most common analytical error, which is comparing a Discover-driven traffic swing against a web Search core update.
Expert tips
Annotate your analytics with rollout windows, not rollout start dates. Draw the band from start to completion. A vertical line at the announcement date visually attributes the entire subsequent period to the update, which is exactly the wrong reading.
Build your reporting calendar around the median rollout, not the announcement. With core updates completing in roughly 14 days, a client report issued a week after an announcement will describe a partial deployment and probably require retraction.
Adjust your spam-update response speed for 2026 conditions. When spam rollouts complete in under two days, mid-rollout adjustment is no longer possible. Compliance has to be continuous rather than reactive.
Check for overlapping rollouts before assigning cause. October 2023 ran a spam update and a core update one day apart. March 2024 launched both on the same day. When rollouts overlap, single-cause attribution is not available and you should say so in the report rather than pick one.
Use passage ranking deliberately. Because Google can rank an individual section of a page, structure long articles as a series of genuinely self-contained answers with descriptive headings. This same structure is what makes content extractable by AI answer engines — one structural decision serving two retrieval systems.
Read Google's policy pages, not just the update announcements. The material changes of 2024 were largely in the written spam policies — scaled content abuse, site reputation abuse, expired domain abuse. Those definitions determine what a future spam update will act on.
Watch the composition of the calendar, not just the count. A year with three spam updates and two core updates, as 2026 has been so far, signals a different enforcement posture than a year weighted toward core.
Update-response checklist
Before an update lands
☐ Search Console verified, with Performance data exported monthly for baselines
☐ Traffic segmented by surface: web Search, Discover, News, images, local
☐ Rank tracking segmented by intent, not just by head keyword
☐ Content inventory current, with thin and outdated pages already identified
☐ Link profile reviewed, with any paid or manipulative acquisition stopped
☐ Site audited against Google's written spam policies
☐ Annotations enabled in analytics for deployments and content changes
When an update is announced
☐ Record the name, type, start date, and affected surface
☐ Note the expected duration based on that update type's historical median
☐ Take a clean pre-rollout baseline snapshot
☐ Freeze significant site changes for the duration where practical
☐ Brief stakeholders that assessment comes after completion
After the rollout completes
☐ Compare a full post-completion period against a matched pre-rollout period
☐ Segment results by surface, page type, and query intent
☐ Classify the change: core, spam, policy, technical, or unrelated
☐ Rule out non-algorithmic causes before concluding
☐ Match the response to the classification
☐ Log the outcome in your update database for the next cycle
FAQ Section
What is a Google algorithm update?
A Google algorithm update is a change to one or more of the ranking systems Google uses to order search results. Google runs thousands of changes each year and formally announces only a small number of them, giving those a name, a start date, and a published rollout duration. Announced updates fall into distinct classes — core, spam, and surface-specific updates such as Discover — and each class calls for a different response.
How many Google algorithm updates are there in 2026?
Six named updates were confirmed as of 29 August 2026: the February 2026 Discover update, the March 2026 spam update, the March 2026 core update, the May 2026 core update, the June 2026 spam update, and the August 2026 spam update. That is already more than the four announced across all of 2025.
How long does a Google core update take to roll out?
Across the 16 confirmed core rollouts from November 2021 through May 2026, the mean duration is 17.4 days and the median is 14.5 days. Excluding the 45-day March 2024 update, the mean falls to 15.5 days. Roughly two weeks is the realistic expectation, though the two 2026 core updates completed in about 12 days each.
What was the longest Google core update?
The March 2024 core update, at 45 days — the rollout in which Google absorbed the standalone helpful content system into its core ranking systems. The shortest was the December 2024 core update at 6 days and 4 hours.
What is the difference between a core update and a spam update?
A core update is a broad revision to how Google assesses relevance and quality across all queries; it is not a penalty, and a decline does not mean a site did anything wrong. A spam update improves the systems that detect manipulative practices; a decline does indicate a spam policy was violated. Core recovery comes through genuine quality improvement, usually reassessed at a later core update. Spam recovery comes through removing the violating practice.
Is the helpful content update still a separate thing?
No. The standalone helpful content system was integrated into Google's core ranking systems in March 2024. There are no separately announced helpful content updates any more, and sites affected by helpful content signals are now reassessed as part of broad core updates rather than at a dedicated refresh.
Are there still Google reviews updates?
Not as announced events. The November 2023 reviews update was the last one Google announced separately. Reviews improvements now run continuously without individual announcements, so there is no scheduled update to wait for after improving review content.
Was "Medic" a real Google algorithm?
No. "Medic" is a nickname the SEO industry gave the August 2018 broad core update because health and medical sites appeared disproportionately affected. Google never named an algorithm "Medic" and does not run a health-specific ranking algorithm. The same is true of "Fred," "Phantom," "Possum," and "Hawk" — all industry labels, not official names.
How do I know if a Google algorithm update affected my site?
Pinpoint the exact day traffic changed in Search Console, check that date against Google's published ranking-updates history, rule out non-algorithmic causes such as migrations, indexing changes, tracking breaks, and manual actions, then segment by surface and query intent. Wait until the rollout is confirmed complete before drawing conclusions — a median core rollout runs about two weeks, and rankings fluctuate throughout.
Can you recover from a Google core update?
Sometimes, but there is no guaranteed path. Google's guidance is that a core update decline does not indicate a violation and that improvements may not be reflected until a subsequent core update. Some declines occur because competing content improved rather than because the affected site got worse.
Were Caffeine and Big Daddy algorithm updates?
Neither was a ranking update. Big Daddy (2005–2006) and Caffeine (2010) were crawling and indexing infrastructure changes. They had large effects on what appeared in search results, but through indexing capability rather than ranking logic — which is why "my rankings dropped" and "my pages aren't indexed" call for different diagnoses.
What is a Google ranking incident?
A ranking incident is a service issue Google logs on its Search Status Dashboard under the Ranking heading, titled "Ranking is experiencing an ongoing issue." Unlike a named update, it has no update name and no guidance, because nothing was deliberately shipped — it is a malfunction. Two are on record: 5 October 2023 (26 days) and 15 August 2024 (4 days, 11 hours). Both overlapped named rollouts, so any traffic diagnosis in those windows has more than one candidate cause.
What Google ranking systems are running right now?
Core ranking systems, spam systems including SpamBrain, the reviews system, RankBrain, BERT, neural matching, passage ranking, PageRank and link analysis, freshness systems, exact-match domain systems, original content systems, and local ranking systems. Panda, Penguin, and the standalone helpful content system have all been absorbed into core.
Key Takeaways
- Google's published ranking-updates history begins in late 2021. Everything before that is a mix of Google-confirmed milestones and industry-assigned nicknames, and the two should never be recorded the same way.
- Seven classes of change get called "updates": core, spam, reviews and helpful content, policy and enforcement, indexing and infrastructure, industry-named volatility, and ranking incidents. Each requires a different response.
- Across 16 confirmed core rollouts from 2021 to 2026, the median duration is 14.5 days and the mean is 17.4 days. Assess after completion, not during.
- The March 2024 core update ran 45 days — the longest on record — because it absorbed the standalone helpful content system into core ranking systems.
- Spam update rollouts have compressed sharply: a 13.9-day mean in 2024–2025 against a 1.8-day mean in 2026, with the March 2026 spam update completing in 19 hours and 30 minutes.
- Named updates per year fell from 11 in 2021 to 4 in 2025 before rising to 6 through August 2026. The drop reflects consolidation into core, not a slowdown in engineering.
- Panda, Penguin, and the helpful content system are all part of core ranking systems now. Recovery arrives at a subsequent core update rather than at a dedicated refresh.
- The November 2023 reviews update was the last announced reviews update; reviews improvements are continuous.
- Caffeine and Big Daddy were indexing and infrastructure changes, not ranking updates.
- "Medic," "Fred," "Phantom," "Possum," and "Hawk" are SEO-industry nicknames. Google never used any of them.
- Google's dashboard also logs ranking incidents — service issues, not updates. Two are on record, and both overlapped a named rollout.
- In October 2023, three dashboard entries ran concurrently: a core update, a spam update, and a ranking incident. Single-cause attribution is not available in that window.
- Google published impact percentages for older updates and stopped after 2021. Penguin touched about 3% of queries; Panda 11.8% of US results.
- Discover, News, images, local, and web Search move independently. Segment by surface before diagnosing anything.
Conclusion
The value of an algorithm update history isn't the list. Anyone can publish a list. The value is in knowing which entries describe the same underlying event, which ones describe something Google never actually confirmed, and which ones changed how the game works rather than just shuffling results for a fortnight.
Read this way, twenty-three years of updates tell a coherent story. Google began with crude, easily gamed signals and a handful of hand-tuned changes with codenames. It spent the 2010s building specialised filters — Panda for content quality, Penguin for links, Pigeon for local, Payday and Pirate for specific abuse patterns — each versioned, refreshed, and announced separately. Then it spent the 2020s dismantling that architecture, absorbing the specialised systems into a single continuously running core and reserving named announcements for broad core revisions and spam enforcement.
The rollout data captures the endpoint of that arc. Core updates now deploy in about twelve days, down from a peak above twenty. Spam updates deploy in under two. The calendar has fewer names on it and each name covers more ground. For anyone doing the work, that means the old rhythm of watching for a specific filter's refresh has been replaced by something less legible and more demanding: build content that survives continuous evaluation, because there is no longer a scheduled moment when your fixes get reconsidered.
Not sure whether the last core update actually cost you traffic — or whether something else did?
An algorithm impact assessment answers that specifically. We pull your Search Console data across the full rollout window, segment it by surface, page type, and query intent, check it against the confirmed update record, and rule out the migrations, indexing errors, and tracking breaks that get misattributed to Google every month. You get a written diagnosis with the update class identified and a prioritised remediation plan — or a clear statement that no update was responsible, which is just as useful.
How to choose the best SEO agency in Canada for algorithm recovery
Search for the best SEO agency in Canada and you'll get directories and ranked listicles, none of which tell you whether a firm can actually diagnose an algorithm hit. Three questions separate the ones that can. Ask how they would rule out migrations, indexing changes and tracking breaks before blaming a Google update. Ask what they do when two rollouts overlap and single-cause attribution isn't available. Ask what they would tell you if the answer turned out to be that no update was responsible at all. A firm that answers all three well is worth your time, wherever it happens to sit in a directory.
Seeing your traffic slide and not sure why?
Whether you run a business in Canada, the United States or Europe, a sudden drop in website traffic often traces back to a Google algorithm update — and what you should do about it depends entirely on which kind of update hit you. Connect with 4GoodHosting, the best SEO agency in Canada, or email us or call @ 1 866 708 4678 for a free audit. We'll check your traffic against the confirmed update record, work out whether a core update, a spam update, a ranking incident or something else entirely was responsible, and tell you what's worth doing about it.
