Google’s September spam update took almost two weeks to finish, far longer than any spam update earlier this year, and many site owners spent that stretch wondering whether a traffic dip was their fault. The rollout is over now, which makes this a good time to check your own numbers calmly instead of reacting to forum chatter. In this article, we will explore what Google said, how the timing compares with earlier 2026 updates, what trackers recorded, and a practical way to check your site.
The Rollout From 24 September to 8 October
Start of the Rollout
Google released the update on Thursday, 24 September 2026. Search Engine Journal quoted Google’s Search Status Dashboard as giving a start of 9:15 am Pacific time, while Search Engine Roundtable put it at about 12:00 pm Eastern time. Those two timestamps sit roughly a quarter of an hour apart once you convert between zones, so the gap looks like a rounding habit rather than a real disagreement. Google said the change was global and applied to all languages, and it warned at the outset that the rollout “may take up to two weeks to complete.”
Finish Line
Google’s closing note said the rollout was finished on 8 October 2026, and Search Engine Roundtable recorded the completion at 4:37 am Eastern that morning. Counting from the start to that moment gives roughly 13.7 days, which is where the “nearly 14 days” in the headlines comes from. That is simple arithmetic on the reported times, and it describes the rollout window rather than how long any single site felt an effect. Google labelled it a “normal spam update,” which suggests it was filed as routine maintenance of its spam systems and not as a special event.
Most of the coverage comes from the United States and uses American time zones, so if you work in Asia, Australia or Europe, the local date of your first visible change may differ by a day from the dates above. Keep that in mind when you line up your own charts with the published timeline. A Thursday afternoon launch in New York is already Friday in much of the Asia-Pacific region.
Long Windows Complicate Cause and Effect
A rollout that lasts two weeks is awkward for anyone trying to work out cause and effect. When an update lands in a few hours, a sharp change in traffic can be tied to a single date. When it unfolds over fourteen days, a site can wobble up, down, and back again, and the final position may not be visible until the rollout is long over. Google’s own advice reflects this, which is why it generally suggests waiting for a rollout to finish before drawing conclusions.
It also means that the first week of October is a poor moment to judge a site that moved in late September. Some of the movement may have been undone by the time the update completed, and some may only have begun. If you made edits in the middle of the window, you have probably muddied your own evidence, so note the dates of those edits now while you still remember them.
Four Spam Updates in 2026 Compared
This was the fourth confirmed spam update of the year, following ones in March, June and August, according to Search Engine Journal. What stands out is not the count but the pace of each rollout. Earlier rounds finished quickly, and the September one stretched out far beyond them, which is why so many people kept asking whether it was still running.
| Update | Type | Reported length of rollout |
| March 2026 | Spam | About 19.5 hours |
| May 2026 | Core | 11 days 21 hours |
| June 2026 | Spam | About 2 days 1 hour |
| August 2026 | Spam | About 2 days 16 hours (18 to 21 August) |
| September 2026 | Spam | About 13.7 days (24 September to 8 October) |
The spam figures come from Search Engine Journal, and the core update figure comes from Relevant Audience. By simple arithmetic, the September rollout lasted about five times as long as the August one and well over ten times as long as March’s. A third-party summary of the August round said it was aimed at sites that mass-produce low-value pages, though that is one source’s reading and Google’s own wording was less specific.
The table also shows something useful about timing. The one core update on the list, in May, ran for nearly twelve days, so a long rollout is normal for a broad ranking change. A spam update that behaves more like a core update in length is unusual, and it is the reason some commentators started comparing the two. Duration alone does not tell you how many sites were affected, and Google has not offered that number.
For a small publisher, the practical lesson is to stop treating each spam update as a short, sharp event. Plan for a window of a couple of weeks in which results may keep shifting, and schedule your review for after the all-clear, not during it. A business that sells seasonally should also write down which promotions overlapped with the window, because campaigns, email sends, and price changes can all push traffic in directions that look like search effects.
Spam Updates and Core Updates Are Different Tools
Goals Compared
A core update is a broad adjustment to the systems Google uses to rank helpful, relevant results. It can move sites up as easily as down, and nothing needs to be wrong with a page for it to lose ground to a competitor whose content is now judged a better match. A spam update, by contrast, improves the automated systems that detect content and behavior that break Google’s published spam policies. Pages caught by those systems tend to be demoted or, in severe cases, dropped from results.
Neither system works on a calendar that site owners can predict. Google runs core updates and spam updates several times a year and announces them on its Search Status Dashboard, usually only when they begin and again when they end. Spam detection also runs continuously in the background, so a site can be affected outside any named update.
That difference matters when you diagnose a change. A fall during a core update often calls for a hard look at overall quality and usefulness compared with competitors. A fall during a spam update points you toward a narrower question: is there anything on the site that looks like it was built to manipulate rankings rather than to help visitors?
Recovery Paths
Google’s guidance for both types is that improvement takes time, but the mechanics differ a little. After a spam update, Google says affected sites should review its spam policies and fix whatever violates them. It also says recovery can take many months and that periodic refreshes of the systems will occur. Making changes may help if the automated systems learn over several months that the site now complies. There is no request form and no promised date.
One more caution is worth stating. Search Engine Roundtable noted that some people described the volatility during this rollout as feeling like a core update, although its author said they were not certain. Treat that as an observation, not a classification. Google called this a spam update, and the safest working assumption is that its spam systems were the ones adjusting.
Scope: What Google Would and Would Not Say
Google said the update targets sites that violate some of its search spam policies, and it was explicit that the update does not cover every policy. That wording is easy to skim past, but it carries real information. Some spam practices you might have read about, such as certain kinds of link manipulation, were not the focus this time. Search Engine Roundtable reports that the update does not target link spam, so a backlink clean-up is an unlikely first fix if your traffic moved in this window.
That also shapes how you spend your time. A site owner who sees a dip and immediately pays for a link audit may spend days on something unrelated to the cause. Starting with content, which is where Google’s description of the update points, is the more sensible first move, and a link review can follow later if the content check turns up nothing.
Be careful about reading too much into that. Google’s statement is only that this update was not aimed at link spam, and it is not a claim that links stopped mattering or that link policies were relaxed. Separate systems handle other kinds of abuse, and nothing in the coverage suggests any change to those.
Google would not say what share of search queries was affected. That is typical of its announcements, and it means nobody outside the company can say how widely the update landed. If you see a percentage quoted on social media or in a vendor newsletter, treat it as someone’s estimate, since there is no official figure behind it.
Google also said nothing about policy changes with this update. In practice, that means the rules you should check are the ones already published, not new ones. If a site was hit, the likely explanation is that something it was doing already crossed an existing line, and detection improved.
There is one more thing to keep in mind about the audience for these facts. Most of the reporting comes from American search news outlets, and the update applied to every language and country. A site in another market, such as a regional news publisher in South Asia or a shop in Germany, should assume the same systems apply, but local result pages can look quite different, and a tracker built around US search results says little about them.
Volatility the Tracking Tools Recorded
Search Engine Roundtable documented ranking turbulence in several windows during the rollout: 25 to 27 September, a spike on 30 September, and another stretch from 4 to 6 October. It pointed to charts from tracking tools including Mozcast, Semrush and Sistrix. The write-up described those charts only, without numeric scores, so nobody can say from the coverage how strong the swings were compared with past updates.
Those tools measure how much results move for a fixed set of tracked searches, usually concentrated on markets and keywords that suit their customers. They are a handy weather report, but they are not a verdict on your site. A hot day on a tracker can reflect shuffling among pages you have never competed with, and a quiet day can hide large moves in a niche the tool does not sample.
The pattern of separate bursts instead of one smooth rise is also normal for a long rollout. Systems of this size are typically deployed in stages, and the visible effect arrives in waves as different data centers and signals pick up the change. If your traffic dropped on a date that does not match one of the reported waves, that does not rule the update out, but it should make you look for other causes as well, such as a technical fault, a seasonal shift or a change you made yourself.
A practical habit is to keep a short list of non-search explanations beside the search ones. Did a plugin update change your templates? Did a host migration alter response times? Did a major source of referral traffic, such as a social post or newsletter feature, end? Checking these takes an hour and can save weeks spent auditing content that was never the problem.
Reading Your Own Search Console Data
Choose a Fair Comparison Window
Open the performance report in Search Console and compare a period after the rollout with the weeks before 24 September. A fair comparison uses equal lengths and the same weekday mix, because traffic on many sites swings between weekdays and weekends. Because the rollout ended on 8 October, effects may continue to settle for some time, so look at the period after it ended and again a couple of weeks later.
Do not forget seasonality. Autumn changes behavior for many topics, from school and travel to shopping, and a drop that matches last year’s pattern is not evidence of a penalty. If you have data from the same weeks a year ago, compare those too.
Segment Before Concluding
A sitewide total hides the pattern that matters. Filter by page type, such as articles, product pages, category pages and location pages, and see whether one group fell while the others held. Then split by country, because a global update can still hit one market harder than another, and by site section, such as a blog, a forum or a directory area.
If the whole site moved evenly, a technical issue, a tracking change or a seasonal swing is more likely than a spam classification. If one template or one folder fell sharply while the rest stayed flat, that is the pattern worth investigating. It narrows your audit to a specific set of pages instead of the entire site.
Clicks Beside Impressions
Look at clicks and impressions side by side. If impressions held but clicks fell, your pages are still showing and something else, such as a changed results page or lower positions, is shaping behavior. If both fell together for a defined group of pages, that group lost visibility. Check the Manual actions report too, because a manual action is a human decision that appears in Search Console, while an algorithmic change does not send a message.
If you run an online shop, add one more cut: separate your product pages from your informational content. Shops often rely on thin supplier descriptions across hundreds of items, and an update aimed at low-value pages can treat those very differently from a well-developed buying guide. Publishers with affiliate links should do something similar, comparing pages that offer original testing or first-hand comparison with pages that simply list products.
Finally, annotate. Write down the date, what you saw, and what you changed, in a spreadsheet or a note attached to your analytics. Records like these are the difference between a reasoned recovery and months of guessing.
Policies to Review and Mistakes to Avoid
Rules Worth Rereading
Google publishes its spam policies on Search Central, and the wording can change, so read the current version rather than relying on a summary. From general knowledge, the policies that most often matter to ordinary publishers cover scaled content abuse, which means producing large volumes of pages mainly to rank, whether by hand, automation or a mix. Expired domain abuse concerns buying an old domain to borrow its reputation for unrelated content. Site reputation abuse is about hosting third-party material on a trusted site mainly to exploit its ranking signals. Cloaking means showing search engines something different from what visitors see, and thin affiliate content means pages that add little beyond a manufacturer’s description and a purchase link.
Hold these loosely. This update did not announce a rule change, and Google did not say which policies it focused on beyond saying it did not cover them all. Use the list as an audit checklist, not as a diagnosis.
Route to Recovery According to Google
The path Google describes is to fix the problem and then wait. Changes that bring a site into line may help if its automated systems learn over several months that the site complies, and periodic refreshes will occur on a schedule nobody outside Google knows. You cannot ask for one. That timeline is frustrating, but it also means that rushed, half-considered edits give you little advantage over careful ones.
It helps to separate two ideas that often get blurred. Fixing a problem is something you control and can finish in days or weeks. Having Google recognize the fix is something you do not control, and it depends on the next refresh of its systems. Judge your own work by whether the pages are now better and compliant, and judge the outcome later, when the systems have had a chance to look again.
Errors That Make Things Worse
The most common mistake is mass-deleting pages without a reason. Removing a thousand pages because they are old or thin might help, or it might remove the pages that earned your traffic. Decide page by page, or at least group by group, and write down the reasoning. Another mistake is making large changes while the rollout is still running, because you then cannot tell which change caused which result.
Avoid panic rewrites of content that has performed well for years. Avoid chasing a link clean-up when the update was not about links.
A related temptation is to blame the most recent change you made. If you updated a theme in the week before the update, it is natural to suspect it, but without a before-and-after comparison by page group, you are guessing. Start with the segmented data, and only then ask which changes could plausibly explain the pattern you found. And be wary of anyone who promises guaranteed recovery, since nobody outside Google controls when its systems refresh. A fix made in good faith and documented clearly is the realistic goal.
A Calm Order of Work for the Next Month
Begin by confirming that your drop is real and dated. Check Search Console for the period around 24 September to 8 October, compare it with the weeks before, and rule out tracking glitches, outages, and seasonal patterns. If nothing moved beyond normal variation, you can stop there and simply keep an eye on the numbers.
If there is a clear drop in one segment, audit that segment first. Read a sample of the affected pages as a stranger would, and ask whether each one exists to help a visitor or to catch a search. Look for patterns such as pages that differ only by a place name, near-duplicate articles, thin product descriptions copied from suppliers, or sections that carry little original value. A content audit is more likely to be useful than a backlink audit this time, given what is known about the update.
Then fix in batches you can track. Improve or merge pages that deserve to exist, remove or consolidate those that do not, and keep a simple log with the date of each batch. Resist the urge to change everything at once, because a messy edit history makes it hard to judge what worked.
For sites with many pages, prioritize by traffic lost and by how clearly a page fits a problem pattern. A template that produced several hundred near-identical location pages deserves attention before a single oddly worded article. For a small blog with fifty posts, a careful read-through over a weekend may be enough, and you may find that most of your content is perfectly sound and that the dip was ordinary noise.
After that, wait. Google’s own advice implies that the response may come months later, not days, and that a periodic refresh is what reveals whether you were right. Meanwhile, keep publishing useful work, keep your technical basics clean, and check back monthly.
If your traffic did not move at all, resist the temptation to relax completely. A site that escaped this round has not been cleared of anything, and a spam update is a good prompt for a light review of the riskiest parts of your content anyway. Look at your oldest templated pages and your thinnest sections, and ask whether you would be comfortable explaining their purpose to a visitor. That quiet habit is cheaper than a recovery project later. This guidance is general, and anyone with a serious business at stake should consider a qualified search professional who is open about methods.
Frequently Asked Questions
1. How Long Did the September 2026 Spam Update Take to Roll Out?
It began on 24 September 2026 and Google said it was complete on 8 October 2026, which works out to roughly 13.7 days. Google had warned it could take up to two weeks.
2. When Should I Check Search Console for Effects?
Compare the period after 8 October with the weeks before 24 September, using equal lengths and matching weekdays. Check again a couple of weeks later, because effects can keep settling after a rollout ends.
3. Can a Site Recover After a Spam Update?
Google says recovery is possible but can take many months. Fix anything that breaks its spam policies, keep records, and wait for its systems to refresh, which it says happens periodically.
4. Does This Update Target Link Spam?
According to Search Engine Roundtable, no. That does not mean links are irrelevant, only that this round was not aimed at them, so a content review is a better first step.
5. Is a Spam Update the Same as a Core Update?
No. A core update adjusts how Google ranks helpful results broadly, while a spam update improves detection of content that breaks its spam policies. Some observers said this one felt like a core update, but Google called it a normal spam update.
6. Should I Delete Thin Pages Straight Away?
Not without reasoning. Review them in groups, improve or merge the ones that deserve to exist, and record why you removed any others, since deleting good pages by mistake can cost you traffic.
