Google's Search Status Dashboard lists a "September 2026 spam update" starting on 24 September 2026, five weeks after an August spam update that began on 18 August and ran for two days and sixteen hours. A store that watched its organic traffic move in the last week of September will want to know whether this is why, and the useful answer starts with what Google says a spam update is rather than with what commentators say it targets.
This is Google's own description of the update, the spam policies among its sixteen that an ordinary online store can break without intending to, what to check if traffic moved after the 24th, and what Google says about getting back.
Google: "SpamBrain is our AI-based spam-prevention system. From time-to-time, we improve that system to make it better at spotting spam and to help ensure it catches new types of spam." When it does, "we refer to this as a spam update" and the date goes on the dashboard. That is the whole of what Google has said about September's: the name and the start date. It has not said which techniques this one is better at catching, and nothing here guesses.
The consequence of being caught is stated on the spam policies page: "Sites that violate our policies may rank lower in results or not appear in results at all." And the advice to a site that saw a change is the same every time: review the spam policies and make sure the site complies with them.
So a spam update is not an algorithm change in the sense a core update is. It does not re-weigh what makes a page good. It gets better at identifying pages built to manipulate rankings, and a site that was not doing that should see nothing. Most stores are in that position. The ones that are not usually got there by accident, through a supplier, a plugin, a content tool or a previous agency.
Organic traffic moves for many reasons, and an update with a date on it attracts the blame for all of them. Before assuming the spam update is the cause, rule out the others that happen in the same weeks.
A manual action. Search Console's Manual actions report shows whether a person at Google has applied an action to the site, which is separate from any automated update and comes with its own reconsideration process. If the report is empty, the cause is not a manual action.
Pages dropping out of the index. Search Console's page indexing report shows pages excluded and why. A rise in pages marked as not indexed, blocked by robots.txt or marked noindex usually traces to a change on the site rather than to Google.
A change on the site. A new theme, a platform migration, changed addresses without redirects, or a robots.txt edit in the same weeks can move traffic on its own. The site's own change log is the first place to look.
The season. Compare the same weeks last year. Many stores see organic traffic shift in late September and October for reasons that repeat every year.
A change in how results look. Fewer clicks with steady impressions can mean the page still ranks while more of the answer appears on the results page itself. That is a different problem, with different fixes.
Tracking. A change to the analytics setup or consent banner can make traffic appear to fall when it has not. Search Console's clicks, which do not depend on the store's tracking, are the check.
Only once these are ruled out does the spam policy review below become the likely explanation.
Google lists sixteen spam policies. Several are about hacking, malware and cloaking and are not accidents. These are the ones a store can walk into.
Scaled content abuse. "Many pages generated for the primary purpose of manipulating search rankings and not helping users." A store that generated a landing page for every keyword variation, a blog post for every search term in a tool export, or a collection page for every colour of every product with the same paragraph on each, has built exactly this, and generative tools make it a morning's work.
Thin affiliate pages. Affiliate content where "product descriptions and reviews are copied directly from the original merchant without any original content". The store-side version is the reverse: product pages carrying the supplier's paragraph, identical to forty other retailers'.
Product descriptions and supplier feeds is the longer argument for why that copy was never ranking anyway.
Scraped content. Content taken "from other sites, often through automated means" and hosted "with the purpose of manipulating search rankings". Review widgets that import reviews written elsewhere, specification tables lifted from a manufacturer, and blog sections fed by a feed all qualify.
Doorway abuse. Pages created to rank for specific queries that funnel users to another page rather than answering there. City pages for a store that ships everywhere, size or colour pages that all redirect to one product, and search-result pages indexed as landing pages are the common forms.
Why Shopify collection pages do not rank covers the related problem of collections with nothing on them.
Reputation, Domains, Hidden Text and Links
Site reputation abuse. "Publishing third-party content mainly because of that host site's already-established ranking signals." A store that lets a partner publish coupon pages, guest posts or an unrelated blog on its domain for the traffic is the host in that sentence.
Expired domain abuse. Buying an expired domain and repurposing it "primarily to manipulate search rankings by hosting content that provides little to no value". A brand that bought a lapsed domain for its history rather than its name should read that twice.
Keyword stuffing, hidden text and user-generated spam. Filling a page "with keywords or numbers in an attempt to manipulate rankings", text placed where humans will not see it, and spammy content added through reviews, Q&A or comments. The last one is a moderation job the store may not know it has.
Link spam. Links created "primarily for the purpose of manipulating search rankings". Google's note on this one is the harshest on the page: when the effect of spammy links is removed, "any ranking benefit the links may have previously generated for your site is lost" and "cannot be regained". A store that paid for links in 2023 and ranked on them will not get that ranking back by removing them; it will have to earn the position again.
First, whether it moved at all in the organic line, by page type, in Search Console, rather than in a total that includes paid, social and the AI channels. A drop concentrated on one page type (the blog, the collections, the location pages) points at a policy; a drop across everything points elsewhere.
The Shopify SEO guide for a store that has to rank is the baseline for what each page type should be doing.
Second, whether anything on the site matches a policy above, honestly. The pages most stores find are the ones nobody remembers making: the programmatic landing pages, the imported blog, the partner's section. Third, whether any of those pages earn their place by being useful to a shopper, which is Google's test in every policy's wording. A page that helps users is not spam however it was made; a page made for rankings is spam however good it looks.
Then fix or remove. Remove what exists only to rank. Rewrite what a shopper would read if it said something. Add structured data to what remains, because
structured data for online stores is how a search engine confirms a product page is a product page. And do not expect the graph to turn the next week.
A clean up done in a hurry can remove the pages that were earning traffic along with the ones that were not. A few rules keep it precise.
List before acting. Export every page of the type under suspicion with its clicks and impressions from Search Console for the last few months. Pages that earned real traffic from real searches are evidence of usefulness, and deserve a closer look before removal.
Remove what exists only to rank. A page with no purpose for a shopper should be removed and answer as not found. That tells Google plainly the page has gone.
Redirect only to a true equivalent. Redirect a removed page only when another page answers the same need, such as an old colour page to the product it described. Redirecting hundreds of removed pages to the home page does not carry anything across; Google tends to treat such redirects as pages that no longer exist.
Rewrite what is worth keeping. A collection page with a real purpose and a copied paragraph is worth rewriting rather than removing. The test is whether a shopper arriving on it would find something useful there.
Tidy the signals. Remove deleted pages from the sitemap and from internal links, so the site stops pointing at pages it has retired.
Keep a record. Which pages were removed, redirected or rewritten, and when. The months Google says recovery takes are long enough that nobody will remember otherwise, and the record shows what to read when the new year's traffic comes in.
Google's sentence on timing is the one to plan around: changes to comply "may help a site improve if our automated systems learn over a period of months that the site complies with our spam policies." Months, not days, and learn rather than reindex. A store that cleaned up in October should read its organic line in the new year, not at the end of the month, and should spend the interval on the thing a spam update does not touch, which is making the pages that remain the most useful answer to the query.
eCommerce SEO against marketplace optimisation explains why that work is different from the listing work on Amazon, and why it compounds.
Most stores that trip a spam policy did so through a decision nobody reviewed: a tool switched on, a partner given a section, a batch of pages generated on a quiet afternoon. A few standing rules stop it happening again.
New page types are a decision. Any new kind of page, a location page, a comparison page, a generated landing page, is approved by whoever owns the store's search performance before it is built, with a sentence on who it is for.
Generated content is reviewed before it is published. Content tools are useful for drafts. A page goes live only when a person has checked it says something true and useful that the store's other pages do not.
Partners do not publish on the domain. Coupon sections, guest posts and partner content hosted for traffic are the pattern Google's site reputation policy describes. If a partner's content belongs anywhere, it belongs on the partner's site.
Reviews and questions are moderated. User submitted content is the store's responsibility once it is on the page. Remove spam and irrelevant links as they arrive.
No paid links. Links bought for ranking are both a policy breach and, as Google says, a benefit that is lost when caught. Links earned by being worth citing are the only kind worth having.
A quarterly look at what is indexed. Search Console's indexing report shows every page Google holds. A quarterly read finds the pages nobody remembers making before an update finds them first.
We audit a store against the sixteen policies in Google's wording rather than in a tool's score, find the pages nobody remembers making, and decide page type by page type whether to remove, rewrite or keep, then build the content that earns the rankings back over the months Google says it takes. That is part of our
eCommerce SEO work. If your organic traffic moved in the last week of September, the first hour of that audit usually says whether this update is the reason.