In short
Google ranks pages with many automated ranking systems, not one algorithm, and several times a year it changes them broadly in a core update. A core update does not target your site; it re-weighs everyone, so a drop usually means other pages now look more useful, not that you were punished. Spam is handled separately: by automated systems and, when a human reviewer confirms a policy violation, by a manual action that you can see in Search Console and lift with a reconsideration request.
How Google ranks: systems, not one algorithm
People say the algorithm as if it were one formula. Google describes something different: a set of automated systems that look at many signals and work together. Its ranking guide says it covers some of “our more notable ranking systems”, and that some of them “are part of our core ranking systems, which are the underlying technologies that produce search results in response to queries.” Others handle specific needs, such as crisis information or local news.
Two points from that guide shape how you should think about any ranking change:
- Ranking happens per page. “Our ranking systems are designed to work on the page level, using a variety of signals and systems to understand how to rank individual pages.”
- The site still matters. “Site-wide signals and classifiers are also used and contribute to our understanding of pages.”
So a strong page on a weak site can still rank, and a weak page on a strong site can still fail. Google does not publish weights for any system, and no single system decides a position on its own. If you want the bigger picture first, the lesson From crawl to ranking: the full pipeline shows where ranking sits after crawling and indexing.
What Google officially says
“Having some good site-wide signals does not mean that all content from a site will always rank highly, just as having some poor site-wide signals does not mean all the content from a site will rank poorly.”
Google Search Central — A guide to Google Search ranking systems“Several times a year, Google makes significant, broad changes to our search algorithms and systems.”
Google Search Central — Google Search's core updates and your website. Google adds: “These changes are broad in nature, and don't target specific sites or individual web pages.”“In general, most sites don't need to worry about core updates and may not even realize one has happened.”
Google Search Central — Google Search's core updates and your website“We detect policy-violating practices both through automated systems and, as needed, human review that can result in a manual action.”
Google Search Central — Spam policies for Google web search. Also: “Sites that violate our policies may rank lower in results or not appear in results at all.”“Our algorithms are extremely good at detecting spam, and in most cases we automatically discover it and remove it from our search results.”
Search Console Help — Manual actions reportThe ranking systems Google names
Google's ranking guide lists the systems below. The list is not complete (Google calls them its “more notable” systems), but it is the official vocabulary. When an article talks about a system that is not on this list, treat it as someone's theory.
What each one does, in Google's words where the wording matters:
| System | What it does |
|---|---|
| BERT | “is an AI system Google uses that allows us to understand how combinations of words express different meanings and intent.” |
| Crisis information systems | Show hotlines and trusted information for personal crisis searches, and SOS Alerts during natural disasters and wide-spread crises. |
| Deduplication systems | “our systems show only the most relevant results to avoid unhelpful duplication.” A page shown as a featured snippet is not repeated further down the first page. |
| Exact match domain system | “works to ensure we don't give too much credit for content hosted under domains designed to exactly match particular queries.” |
| Freshness systems | Several “query deserves freshness” systems that show recent content when a search expects it, for example news after an earthquake. |
| Link analysis systems and PageRank | Use how pages link to each other to understand what pages are about. Of PageRank, Google says: “How PageRank works has evolved a lot since then, and it continues to be part of our core ranking systems.” |
| Local news systems | Identify and surface local news sources, for example in Top stories and Local news features. |
| MUM | An AI system that understands and generates language. “It's not currently used for general ranking in Search but rather for some specific applications”. |
| Neural matching | “is an AI system that Google uses to understand representations of concepts in queries and pages and match them to one another.” |
| Original content systems | “We have systems to help ensure we are showing original content prominently in search results, including original reporting, ahead of those who merely cite it.” |
| Removal-based demotion systems | A significant volume of valid removals against a site (copyright, defamation, counterfeit goods, court orders, personal information) is used to demote other content from that site. |
| Passage ranking system | “is an AI system we use to identify individual sections or "passages" of a web page to better understand how relevant a page is to a search.” |
| RankBrain | “is an AI system that helps us understand how words are related to concepts.” It lets Google return relevant pages that do not contain every word of the query. |
| Reliable information systems | Surface more authoritative pages, demote low-quality content, elevate quality journalism, and show content advisories when reliable results are lacking. |
| Reviews system | “The reviews system aims to better reward high quality reviews, content that provides insightful analysis and original research, and is written by experts or enthusiasts who know the topic well.” |
| Site diversity system | “we generally won't show more than two web page listings from the same site in our top results”. “Site diversity generally treats subdomains as part of a root domain.” |
| Spam detection systems | A range of systems, including SpamBrain, that deal with content and behaviour violating the spam policies. “These systems are constantly updated to keep up with the latest ways that the spam threat evolves.” |
Retired systems: helpful content, Panda, Penguin, Hummingbird
Google keeps a separate “Retired systems” list and explains it plainly: “The following systems are noted for historical purposes. They've either been incorporated into successor systems or made part of our core ranking systems.” Retired does not mean the idea was dropped. It means the work is now done inside the core systems and no longer has its own name or its own update.
- Helpful content system. Announced in 2022 as the Helpful Content Update, it aimed to show original content written for people rather than for search traffic. “In March 2024, it evolved and became part of our core ranking systems”. Since then there is no separate helpful content update to wait for; that assessment happens in core ranking and moves with core updates.
- Panda. Announced in 2011 to favour high-quality, original content. It “evolved and became part of our core ranking systems in 2015.”
- Penguin. “This was a system designed to combat link spam.” Announced in 2012, “it was integrated into our core ranking systems in 2016.”
- Hummingbird. “This was a major improvement to our overall ranking systems made in August 2013.”
In practice: if someone tells you your site was hit by Panda, or that you need to recover from the helpful content update today, translate it into current terms. The question is how your pages perform in core ranking, and the answer shows up around core updates.
What a core update is
A core update is one of those broad, several-times-a-year changes to Google's systems. Google gives notice when one happens on its list of ranking updates. Google's own explanation is that “As content on the web changes, we assess and update our systems to keep pace, as a whole.”
Google explains it with a restaurant list. You wrote a list of your 20 favourite restaurants years ago; new places have opened, you have changed your mind about others, so the list changes. Google's conclusion: “restaurants that move down aren't necessarily "bad"; there are just other restaurants that make your top 20.”
Three consequences follow:
- Nothing gets repaired in a core update. There is no specific violation to undo. Your pages were compared again with everything else that answers the same queries.
- Gains and losses are relative. You can lose positions without changing anything, because a competitor improved or the systems now value something your pages lack.
- Not every update is announced. “We're continually making updates to our search algorithms, including smaller core updates.” Google adds: “These updates are not announced because they aren't widely noticeable”.
How an update rolls out and where to see it
An update does not switch on at one moment. It rolls out over a period of days to weeks, and positions can move up and down during that time. Two Google places tell you what is going on:
- Google Search ranking updates list. The developers.google.com address for this list now redirects to the Ranking history of the Search Status Dashboard. It lists each announced update with its start date and how long the rollout took.
- Search Status Dashboard (
status.search.google.com). Its front page shows the current state of four services, Crawling, Indexing, Ranking and Serving, over recent days, and marks incidents. The Ranking history goes back several years and also offers an RSS feed and a JSON history.
The history shows that core updates are not the only announced type. You will see spam updates (changes to spam detection), past helpful content and reviews updates (from before those systems were folded in), Discover updates, and occasional incidents where ranking itself had a problem. Rollout lengths vary a lot from one update to the next, so read the dates for the specific update rather than assuming a typical length.
Spam policies: the full list
Google defines spam as “techniques used to deceive users or manipulate our Search systems into featuring content prominently”, which now explicitly includes “attempting to manipulate generative AI responses in Google Search.” The policies apply to every kind of content Google finds on the web, including Google's own properties. The list is not a loophole map: “Our policies cover common spam practices, but Google may act against any type of spam practices we detect.”
| Policy | Violation, for example | Not a violation |
|---|---|---|
| Cloaking | Different content for Googlebot than for users, to manipulate rankings | A paywall where Google sees the full content and Flexible Sampling guidance is followed |
| Doorway abuse | Near-identical city pages that all funnel to one page | A clear, browsable hierarchy of genuinely different pages |
| Expired domain abuse | A bought charity domain repurposed for casino or affiliate content | Buying a domain and building a real site on it |
| Hacked content | Injected code, pages, hidden links or redirects added by an attacker | The site owner is a victim; clean up and secure the site |
| Hidden text and link abuse | White text on white, text off-screen, font size 0, a link on one hyphen | Accordions, tabs, sliders, tooltips, screen-reader-only text |
| Keyword stuffing | Blocks of city names, phone number lists, unnatural repetition | Using the words your topic naturally needs |
| Link spam | Paid links, excessive link exchanges, automated links, optimised anchors in guest posts | Paid or sponsored links marked rel="sponsored" or rel="nofollow" |
| Machine-generated traffic | Automated queries to Google, including scraping results for rank checking without permission | Using official, permitted interfaces |
| Malicious practices | Malware, unwanted software, back button hijacking | Normal downloads that do what users expect |
| Misleading functionality | A fake generator or tool that only leads to ads | A tool that really delivers the promised function |
| Scaled content abuse | Many low-value pages from AI, scraping, synonymising or stitching, made to rank | Producing content at scale that genuinely helps users, by any method |
| Scraping | Republishing others' content with no added value or only synonyms swapped | Quoting with attribution and adding your own analysis |
| Site reputation policy | Third-party pages placed on a strong host mainly to borrow its ranking signals | Wire services, syndicated news, forums, editorial columns, integrated affiliate content |
| Sneaky redirects | Mobile users sent to a spam domain, desktop users see the normal page | Site moves, merging pages, redirect after login |
| Thin affiliation | Product descriptions copied from the merchant with affiliate links and nothing else | Original reviews, testing, comparisons, price information |
| User-generated spam | Spam posts, comments, profiles or uploads added by users | A moderated community with real user content |
A few policies deserve a closer look because they catch legitimate sites:
- Scaled content abuse “is when many pages are generated for the primary purpose of manipulating search rankings and not helping users.” Google says this applies “no matter how it's created”, so the issue is value, not whether a person or an AI tool wrote it.
- Link spam does not ban paid links. “It's not a violation of our policies to have such links as long as they are qualified with a rel="nofollow" or rel="sponsored" attribute value to the <a> tag.”
- Expired domain abuse “is where an expired domain name is purchased and repurposed primarily to manipulate search rankings by hosting content that provides little to no value to users.”
- Thin affiliation: “Not every site that participates in an affiliate program is a thin affiliate.” Added value is what separates the two.
- Site reputation policy: “Having third-party content alone isn't inconsistent with the site reputation policy”. It becomes a problem when the content is there mainly because of the host's ranking signals. Outside the European Economic Area such pages can get a manual action; inside the EEA they may instead be ranked separately from the main domain.
- Sneaky redirects: “there are many legitimate, non-spam reasons to redirect one URL to another.” The test is intent to deceive.
The same page ends with Other practices that can lead to demotion or removal: legal removals (a significant volume of valid copyright, defamation, counterfeit or court-ordered removals), personal information removals (for example sites with exploitative removal practices), policy circumvention (moving violating content to new subdomains, subdirectories or sites) and scam and fraud (impersonating businesses, fake customer support).
Manual actions
A manual action is the one case where a person at Google is directly involved. “Google issues a manual action against a site when a human reviewer at Google has determined that pages on the site are not compliant with Google's spam policies.” It is not a secret: “If your site is affected by a manual action, we will notify you in the Manual actions report and in the Search Console message center.”
The Manual actions report shows a count at the top. “If your site has no manual actions, you'll see a green check mark and an appropriate message.” Each action lists the affected URL patterns, which can be a directory such as https://example.com/real-estate/* or the whole site.
The types Google lists in the report help are: back button hijacking; site abused with third-party spam; user-generated spam; spammy free host; structured data issue; unnatural links to your site; unnatural links from your site; thin content with little or no added value; cloaked images; hidden text and/or keyword stuffing; major spam problems; AMP content mismatch; sneaky mobile redirects; a set of News and Discover policy actions; and site reputation policy.
How a manual action is lifted
- Read the action and its affected pages in the report.
- Fix every affected page. Google is strict about this: “Fixing the issue on just some pages will not earn you a partial return to search results.” If there are several actions, fix all of them.
- Make sure Google can reach the pages: no login, no paywall, not blocked by robots.txt or
noindex. Test with URL Inspection. - Request review in the report. A good request “Explains the exact quality issue on your site.” It “Describes the steps you've taken to fix the issue.” And it “Documents the outcome of your efforts.”
- Wait for the decision. “Most reconsideration reviews can take several days or weeks”, and link-related requests can take longer. Do not resubmit while a request is pending.
If you bought a site that already had a manual action, fix the issues and say in the request that you recently acquired it. Unnatural-links actions are the case where a disavow file can play a role; the lesson Link audits and disavow covers that.
Manual actions and security issues are separate reports. The Security Issues report “lists indications that your site was hacked, or behavior on your site that could potentially harm a visitor or their computer”. Those pages can show a warning label in results or a browser warning page, which manual actions normally do not.
Algorithmic change or manual action: how to tell
Most traffic drops are algorithmic. A manual action is rarer, always visible in Search Console, and has a clear way out. Knowing which one you face decides everything you do next.
| Algorithmic change (core update, spam system, other systems) | Manual action | |
|---|---|---|
| Who decides | Automated systems | A human reviewer at Google |
| Notification | None for your site; at most an announced update on the dashboard | Manual actions report and Search Console message |
| Scope | Many sites move at once; inside your site, some pages more than others | Your site: listed URL patterns or the whole site |
| Cause | Relative usefulness, or spam detected automatically | A confirmed violation of a named spam policy |
| How to fix | Improve the site as a whole for users; remove spam practices | Fix all affected pages, then request review |
| How it ends | Systems re-assess over time; often visible at a later core update; no guarantee | Google revokes the action after a successful review |
A quick check, in this order:
- Manual actions report. Any action listed? If yes, that is your first job.
- Security Issues report. Hacked content can look like a ranking drop.
- Search Status Dashboard. Does the start of your drop fall inside an announced update's dates?
- Performance report. Compare the right weeks, then split by search type (Web, Images, Video, News), pages and queries.
- Your own changes. Releases, migrations,
noindex, robots.txt or redirect changes around the same dates.
This is only the triage. The full method is in the lesson Why traffic dropped: a diagnostic method.
What to do and not do after an update
Google's core updates page gives a short, specific procedure. The lesson After a core update walks through it in detail; here is the core of it.
- Wait for the end. “Confirm that the core update has finished rolling out.” Note the start and end dates on the dashboard.
- Wait another week. “We recommend waiting at least a full week after a core update completes before analyzing your site in Search Console.” Then compare a week after with a week before the update started.
- Size the drop. For a small drop (Google's example is position 2 to 4): “There's no need to take drastic action”, and in fact “we recommend avoiding making changes to content that's already performing well”. For a large drop (Google's example is position 4 to 29), do a deeper assessment.
- Assess the site as a whole. Use Google's self-assessment questions, be objective, ask people unaffiliated with the site, and look hardest at the pages that lost most.
- No quick fixes. “Avoid doing "quick fix" changes (like removing some page element because you heard it was bad for SEO).”
- Delete only as a last resort. “Deleting content is a last resort, and only to be considered if you think the content can't be salvaged.”
Then set expectations. Google says “some changes can take effect in a few days, but it could take several months for our systems to learn and confirm that the site as a whole is now producing helpful, reliable, people-first content in the long term.” And: “If it's been a few months and you still haven't seen any effect, that could mean waiting until the next core update.” Finally, Google is honest that “there's no guarantee that changes you make to your website will result in noticeable impact in search results”.
Rule out technical causes before blaming an update
OurSEO Tools runs a 69-point checklist and crawls up to 1,000 pages, recording status code, canonical, indexability and load time per URL, so you can check whether a drop has a technical cause on your side.
Common mistakes
- ✕Changing things mid-rolloutPositions are still moving, so you cannot tell what your change did.
- ✕Comparing the wrong datesA week during the rollout, or a period that mixes seasons, hides the real effect.
- ✕Calling every drop a penaltyA core update is not a penalty. A penalty in Google's sense is a manual action, and you would see it in Search Console.
- ✕Blaming the update for your own bugA
noindex, robots.txt rule or broken redirect released the same week looks like an update hit. - ✕Mass-deleting contentGoogle calls deletion a last resort; improving or merging usually comes first.
- ✕Disavowing links after a core updateA disavow file belongs to unnatural-links problems, not to core-update drops.
- ✕Chasing retired systemsThere is no separate Panda, Penguin or helpful content update to recover from any more.
- ✕Partial clean-up before a reconsideration requestFixing only some affected pages does not earn a partial return.
FAQ
Is a core update a penalty?
No. Google says core updates “don't target specific sites or individual web pages.” A drop means your pages were re-assessed against others. A penalty in the strict sense is a manual action, which appears in the Manual actions report.
Where do I see when a core update started and ended?
On the Search Status Dashboard, in the Ranking history. Each announced update has a start date and a rollout length; the old ranking updates list on developers.google.com now redirects there.
Does the helpful content system still exist?
Not as a separate system. Google lists it under retired systems: “In March 2024, it evolved and became part of our core ranking systems”. The same goals are now assessed in core ranking.
How long does recovery take?
Google gives a range, not a number: some changes take effect in a few days, a site-wide re-assessment can take several months, and sometimes you wait for the next core update. There is no guarantee of recovery.
Will Google tell me if an algorithm demoted my site?
No. Algorithmic changes come with no message for your site. Only manual actions and security issues are reported to you in Search Console.
Is AI-generated content against the spam policies?
Not by itself. Scaled content abuse applies “no matter how it's created”: the problem is many low-value pages made mainly to rank, whether written by people, tools or AI.
Sources
- Google Search Central — A guide to Google Search ranking systems
- Google Search Central — Google Search's core updates and your website
- Google Search Central — Spam policies for Google web search
- Google Search ranking updates (Search Status Dashboard, Ranking history)
- Search Console Help — Manual actions report
- Google Search Status Dashboard
Quotes checked against the official pages on 30 October 2026.