Google’s August 2026 Spam Update: What UK SMEs Should Check (and What to Ignore)
Google finished rolling out the August 2026 spam update last week. If your rankings wobbled, or a Search Console chart looked ugly in mid-August, this is the practical version: what actually shipped, what did not, and what a UK SME should do with half an hour in Search Console.
Two things to get straight before you change a single page. This was a spam update, not a core update. And a confirmed Search Console logging error sat in the week immediately before the rollout. Mix those two up and you will “fix” a reporting glitch, or ignore a real policy problem.
What actually shipped (and what didn’t)
Google’s Search Status Dashboard recorded the August 2026 spam update as an incident affecting Ranking. It started on Tuesday 18 August 2026 at 09:27 Pacific (17:27 UK time) and completed on Friday 21 August at 01:49 Pacific (09:49 UK time). Google’s ranking history lists the duration as 2 days, 16 hours. Google said it applied globally and to all languages.
That is the whole official launch note. There was no Search Central blog post, no new spam category, and no figure for the share of queries affected. Google labelled it a spam update, not a core update. The 2026 core updates Google has published so far were in March and May. Different systems, different recovery path. Do not treat this as a “wait for the next core update and hope” event.
Google’s standing explanation is on Search Central: automated spam detection runs all the time, and a named spam update is a notable improvement to those systems (Google cites SpamBrain). Sites that move after a spam update should review Google’s spam policies. Sites that violate them “may rank lower in results or not appear in results at all.”
What did not ship: a link-spam update, a core update, a new policy document, or a recrawl button that undoes the damage. If someone is selling an emergency backlink package or a “priority reindex” as the fix, they are not reading the same dashboard as Google.
Don’t mix this with a Search Console reporting glitch
Before you assume every dip from mid-August is the spam update, read Google’s data-anomalies page. Google confirmed a logging error that caused a decrease in impressions on the Generative AI performance report in Search for data from 13 to 17 August 2026. Google’s wording is blunt: the issue affects data logging only. A separate logging error on 13 August under-reported clicks and impressions in the Discover performance report, including Generative AI features in Discover for properties that have them. Same caveat: logging only.
That window sits immediately before the spam update started. If Generative AI impressions fell from 13 August and your standard Search performance report moved from 18 August, those are not the same event. The first is Google failing to count some impressions. The second can be a real ranking change. Google did not say the standard Search performance report was part of that logging error. Treat the Generative AI report from 13–17 August as unreliable, and cross-check the rest against Google Analytics organic sessions or server logs.
Annotate 13–17 August as “GSC AI/Discover logging”, annotate 18–21 August as “spam update rollout”, and only then decide whether a page actually lost visibility.
What Google said it targets vs what it does not
Google did not publish a target list for this update. Anyone claiming it “went after AI content” or “hit local SEO” is guessing. What Google did say is narrower, and more useful.
It is a normal spam update for all languages and locations. It penalises some techniques that break Google’s spam policies. Google also confirmed to Search Engine Roundtable that this update does not target link spam, does not target the site reputation abuse policy, and does not target some other policies. That matters because the recovery advice for a link-spam update is different: Google’s spam-updates document says that when systems strip the effect of spammy links, any ranking benefit those links previously gave you is gone, and you cannot get it back. That is not this update. Do not start a panic disavow.
Review the live policy list, not a rumour. Google has not said which policies this rollout enforced. The two that most often catch ordinary UK business sites are scaled content abuse and doorway abuse.
Scaled content abuse is when many pages are generated primarily to manipulate rankings rather than help users. Google is explicit that the method does not matter. Using generative AI tools to produce many pages without adding value is listed as an example. So is scraping, stitching content from other pages, running multiple sites to hide the scale, and publishing pages that barely make sense but contain search keywords. Thin, AI-farmed content is not a separate named update. It sits inside that policy.
Doorway abuse is when pages are created to rank for specific, similar queries and then dump the user onto an intermediate page that is less useful than the actual destination. Google’s examples include multiple domains or pages aimed at specific regions or cities that funnel everyone to one page, and substantially similar pages that look more like a search-results list than a site you can browse.
Cloaking, keyword stuffing, scraping and thin affiliation are on the same list. Link spam and site reputation abuse are still policies — they just were not what this particular update was aimed at.
The UK B2B traps that actually get hit
Most UK SME sites that get into trouble after a spam update are not running a dark-pattern network. They are doing a well-meaning version of “more pages, more keywords”.
The first trap is the town-name template. A service business serving Lancashire and Greater Manchester publishes “accountants in Preston”, “accountants in Bolton”, “accountants in Blackburn”, each with the same three paragraphs and a swapped place name. That is textbook doorway abuse if those pages exist to rank for near-identical queries and then send everyone to one contact form. It is also a poor substitute for real local SEO. One strong location page with unique proof — case studies, coverage, Google Business Profile, local content — beats twenty clones. We have written about doing that properly in Local SEO Strategies for North West Businesses.
The second trap is the AI blog farm. A weekly article that explains something your team actually knows is not spam. Fifty posts a month generated to cover every long-tail variant of “what is X”, with no first-hand detail and no named author who could defend the page, is scaled content abuse waiting to be measured. Google’s policy is about volume, intent and value, not about whether a tool touched the draft. If you want the compounding version of content, it is consistency with substance — see The Power of Consistent Content for SEO Success.
The third trap is the extra domain. A company buys yourservice-manchester.co.uk and yourservice-liverpool.co.uk, slaps a slightly different homepage on each, and points them all at the same phone number. Google lists “multiple websites with slight variations to the URL and home page” as doorway abuse. One proper site is the job. If the site itself is the bottleneck — slow, hard to edit, or a template that encourages thin extra URLs — that is a web development problem, not a content calendar problem.
The fourth trap is the footer city list: a block of towns the page is “trying to rank for”, or the same phrase repeated until it sounds unnatural. Google still lists that as keyword stuffing. None of this means “stop writing about locations” or “never use AI”. It means: if a page would embarrass you in a sales meeting, it should not be in the index.
A 30-minute UK check in Search Console
You do not need a full audit to see whether this update is relevant. You need a clean date range and a bit of discipline.
Set the Search results performance report to the United Kingdom, Search type Web. Compare 18–27 August 2026 with the previous 28 days. Look at clicks and average position, not just impressions. Ignore the Generative AI performance report for 13–17 August; Google has already told you that chart under-counted impressions.
If you have Google Analytics or another session source, put the same dates side by side. A GSC AI-impressions dip with stable organic sessions is the logging error. A drop in both clicks and sessions from 18 August is worth investigating as ranking.
Open Pages. Sort by the largest click or impression change. You are looking for clusters: a folder of location URLs, a run of thin blog posts, or near-duplicate service pages falling together. One important URL moving is often competition or seasonality. Twenty similar URLs moving at once is a pattern.
Check Queries the same way. If you lost a family of “service + town” queries and your branded and core-service queries held, that points at doorway-style pages rather than a whole-site demotion.
Open Manual actions. A named spam update is automated. A manual action is a different problem. Do not skip the check, but do not expect to find one.
Glance at Pages indexing. A pile of “Crawled — currently not indexed” is useful context, not proof you were hit. It tells you where Google already had a low opinion of a cluster.
Export the page table. Highlight anything published at scale in the last 12–18 months that a customer would not miss. Those are your candidates for consolidation, noindex, or a proper rewrite.
Thirty minutes should tell you one of three things: you were not hit; a specific cluster was hit; or the chart you panicked about was the mid-August logging error. Stop there if the data is messy. Guessing is how sites get worse.
If you were hit
Google’s recovery advice is short and, unfortunately, accurate. Review the spam policies and make sure the site complies. Making changes may help if Google’s automated systems learn, over a period of months, that the site now complies. There is no recrawl shortcut. Requesting indexing on a handful of URLs does not reset a site-level spam classification. Nor does publishing three new posts to “show you are active”.
Work the pattern, not the symptom.
Map the URLs that lost impressions together. If they share a template, a generator, or a location-swap, treat them as one problem.
Keep the pages that have unique proof: original process, case studies, named expertise, genuine local detail. Merge near-duplicates onto the strongest URL and redirect the rest. Delete or noindex pages that exist only to occupy a keyword.
Stop the factory. If a process is still generating thin pages, switching off the tap matters more than rewriting last month’s output. Google is looking at whether the site complies going forward.
Do not disavow your backlinks as a first move. Google said this update does not target link spam. Do not buy a new link package either. Both are the wrong lever.
Fix the site so remaining pages can convert the traffic you still have. That is ordinary website optimisation for lead generation, and it is still worth doing while systems re-learn the site.
Give it months, not a fortnight. Spam updates also get periodic refreshes, which is when cleaned-up sites sometimes move. If you only tidy the copy and leave 200 doorway URLs live, do not expect a refresh to rescue you.
If you were not hit
Do nothing dramatic. That is the whole section, and it is the one most agencies skip because it does not sell a project.
Do not start a volume content programme because a competitor dropped. Do not clone their town pages “while there is a gap”. Do not rip up a working information architecture. The update is a reminder of the line Google has already published, not an invitation to rebuild a healthy site.
What is worth doing, quietly: keep publishing pages a sales director would stand behind; keep one website, not a cluster of microsites; keep local pages genuine if you use them; keep the technical basics in good order so Google can crawl what you actually want indexed. If organic traffic is stable, spend the energy on converting it rather than on manufacturing more URLs.
What to do next
If last week’s charts are still sitting on your desk, the next step is a clean read of Search Console against Google’s spam policies, not another tool that promises to name the algorithm. We will do that with you. XL Marketing’s SEO work starts with a free SEO audit: what moved, whether it was the August spam update, the mid-August logging error, or something you were going to have to fix anyway, and what is actually worth changing.
If the site itself is making thin extra URLs too easy to publish, we can look at that on the web development side in the same conversation. If you already know you want a wider view of content, search and lead generation together, start from digital marketing or get in touch.
Google’s August 2026 spam update is finished. Your job is not to guess what it “went after”. It is to make sure the pages you keep would still make sense if a customer landed on them tomorrow — and to ignore the noise that was never a ranking change in the first place.