The August 2026 Spam Update Knocked 82% More Sites Out of the Top 100. Here's How to Tell If You Were One.

Between August 18 and August 21, 2026, a top 10 ranking URL was about 1.8 times more likely to vanish past position 100 than it was two weeks earlier. That is SE Ranking's number, reported by Search Engine Land: 16.71% of URLs that had been sitting in the top 10 fell out of the top 100 entirely during the update window, against 9.2% during a comparable five day stretch in late July when Google had no confirmed update running. That is an 82% increase in top 10 URLs falling off the map.

They looked at 100,000 US keywords across 20 industries, comparing August 17 to August 22 against July 26 to July 31.

So if your traffic fell over in that window, you are not imagining it. But "my traffic fell in August" and "I got hit by the august 2026 google spam update" are two different claims, and most people I talk to have not actually done the work to tell them apart. That matters, because the fix for one is nothing like the fix for the other.

Here is how I check.

First, Pin Down the Window

The update started August 18, 2026 at 9:27 a.m. Pacific and finished August 21. Two days and sixteen hours, global, all languages. Google logged it on the Search Status Dashboard as a ranking incident, which is where you should verify the timestamps yourself rather than trusting my summary or anybody's blog post, mine included.

Two days and sixteen hours is short. That is useful, because it gives you a tight edge to look for. A rollout that takes two weeks smears the drop across a fortnight and makes attribution genuinely hard. This one does not. If your decline started on the 10th, or the 26th, the update is not your explanation and you need to stop reading news posts and start looking at your own site.

How to Actually Compare the Dates in Search Console

Open Search Console, Performance, Search results. Set the date range to Custom, Compare. Compare August 12 to August 17 against August 22 to August 27. Six days each side, both fully outside the rollout, both ending on the same weekday.

That last part is the one people get wrong. Search traffic has a weekly shape. B2B collapses on weekends, ecommerce does the opposite. If you compare a Monday-to-Friday block against a block that includes a weekend, you have manufactured a drop that has nothing to do with Google. Same number of days, same weekdays, both sides clear of the 18th through the 21st.

Then, and this is the step almost nobody does, run the same comparison one year back. July and August are seasonally strange for a lot of industries. If your August always dips 20% against your July, and this year it dipped 24%, you have a seasonality story with a small update on top, not a spam update catastrophe.

Now switch the comparison to the Pages tab and sort by click difference. This is where the diagnosis actually happens.

What a Spam Hit Looks Like Versus a Core Update

Core updates are reassessments. Google is rescoring quality and relevance across the board, and the shape it leaves behind is usually a slope: a lot of pages losing a bit, positions sliding from 4 to 9 rather than 4 to nowhere, some pages actually gaining. It reads like a re-sort.

Spam updates leave a cliff. That SE Ranking figure is measuring exactly this: not gentle slippage but URLs leaving the top 100. When a page goes from position 6 to position 140, that is not Google deciding a competitor answered the query slightly better. That is a suppression signal.

The second tell is the pattern across your site. A core update tends to hit unevenly, by topic or page type or query intent. A spam update, when it lands, often flattens a whole section or a whole domain at once, because the classification is being applied to a pattern of behaviour rather than to individual page quality. Glenn Gabe's case studies from this update include a site with 1.5 million URLs impacted site-wide and a thin Amazon affiliate setup losing 14,000 queries. Not a page here and a page there. The whole thing.

Third tell: check your impressions against your clicks. If impressions held roughly steady and clicks fell, you probably have a click-through problem, an AI Overview eating your result, or a SERP layout change. If impressions fell off the same cliff as clicks, you were removed from the results, not out-competed inside them.

Check What This Update Actually Targeted, Because It Is Narrower Than the Headlines

This is where I want to correct something I saw repeated a lot in the days after the rollout, including in a couple of drafts of my own notes.

The August 2026 spam update was not primarily a link spam update. Both link spam and site reputation abuse were reported as not the main focus this time, since those are handled by other systems on their own schedule. What Gabe documented instead was scaled content abuse, programmatic content published at volume, AI-generated content stacked with other spam signals, thin affiliate pages, and sites running malicious redirects to risky domains.

So if you spent the last two weeks frantically disavowing links because you assumed a spam update means a link update, you may have spent that time on the wrong problem. Toxic links are still worth cleaning up, and I have gone through that exercise enough times to have written up what the spam networks actually look like, but they are unlikely to be your August story.

The other thing worth knowing: since Google rewrote its spam policy definition in May 2026 to cover attempts to manipulate generative AI responses, spam enforcement is no longer only about the ten blue links. Gabe notes sites can drop across surfaces including AI Overviews and AI Mode. Which means if you have been running tactics specifically designed to get a brand mentioned inside AI answers, that behaviour now sits inside the same policy that produced this update. Google's spam team has been enforcing at this cadence for a while now, and I wrote about an earlier instalment of it when back button hijacking became a spam signal.

Rule Out the Boring Explanations First

Before you conclude anything, kill the mundane causes. I have watched people write a three-page recovery plan for what turned out to be a broken robots.txt.

Did you deploy anything between the 15th and the 22nd? Check your git log, not your memory. A CMS update, a plugin, a theme change, a migration, a canonical tag that started pointing somewhere clever.

Is your analytics still recording? Consent banner changes, a tag manager container that got republished, a GA4 property that quietly stopped receiving events. If Search Console shows steady clicks and your analytics shows a cliff, the cliff is in your measurement.

Check Search Console's coverage and indexing reports for a jump in excluded pages. Check the Manual actions report too, though I will save you the suspense: spam updates are algorithmic and do not produce a manual action. If you have a manual action, that is a separate human review of your site and a reconsideration request is the right move. If you do not, a reconsideration request does nothing at all.

And check whether your rank tracker changed anything: location, device, tracked keyword set. A tracking config change can produce a beautifully convincing fake drop.

Why the Recovery Path Is Different, and Why This Is the Uncomfortable Part

With a core update, the honest advice really is close to "improve the site and wait for the next one." Google reassesses broadly on its own schedule, and sites do come back when the underlying quality changed.

Spam updates do not work like that, and I think this is the single most misunderstood thing about them.

A spam update is not a rescoring you can be lucky in. It is an enforcement pass. The system got better at recognising a pattern that violates a stated policy, and it found yours. Waiting does not help, because the next run will apply the same, probably better, classifier to the same unchanged site and reach the same conclusion. If the cause is still there, the next update hits you again.

There is also no re-score button. Fixing the violation does not start a clock. Google's systems have to crawl and observe compliance over time, and the reported recovery timeline for spam demotions is measured in months, not weeks. That is a genuinely bad piece of news to deliver to a client, and I have never found a way to make it sound better than it is.

The other honest bit: I cannot tell you whether removing 40,000 programmatically generated pages will bring your rankings back, because nobody can. Sites that clean up do recover. Some do not, or recover to a fraction of where they were. If your traffic model depended on content Google now classifies as scaled abuse, the recovery target may simply be lower than the peak.

The Checklist

Work through this in order. Do not skip to the fixing.

1. Confirm the update dates on Google's Search Status Dashboard yourself. August 18 to August 21, 2026.
2. Search Console, Performance, Compare: August 12-17 against August 22-27. Same weekdays, same length, both clear of the window.
3. Run the same comparison against the previous year to rule out seasonality.
4. Compare impressions to clicks. Both fell together, or only clicks fell?
5. Sort the Pages tab by click difference. Note whether losses are concentrated in one section or spread across the domain.
6. Pick your ten worst-hit URLs and check their actual current position. Slid a few places, or gone past 100?
7. Check your deploy log and CMS change history for August 15 to 22.
8. Verify analytics and rank tracker configuration did not change.
9. Check Search Console for indexing changes and for a manual action.
10. If it still points to the update: inventory your programmatic and AI-generated pages, your affiliate templates, and any redirect logic on the site. Against Google's spam policies, honestly, page type by page type.
11. Write down what you changed and the date you changed it. Months from now you will need that timeline to interpret whatever happens next.

Steps 1 through 9 take about an hour if your Search Console access is already set up. Most of the panicked emails I get would be answered by step 4.

What I Would Do on Monday

If the comparison shows a cliff, clustered by page type, with impressions gone and no manual action, treat it as a spam classification and go find the pattern. It will usually be embarrassingly obvious once you are looking for it: a template that generated 12,000 near-identical location pages, a bulk-published AI content run from spring, an affiliate section nobody has opened in a year.

If the comparison shows a slope, spread across topics, with some pages up, that is core update territory or ordinary competitive movement, and the work is quality, not cleanup.

If the comparison shows nothing much at all once you correct for weekdays and seasonality, congratulations, go back to work. That happens more often than the panic suggests.

I run this diagnostic as the first pass of any audit now, and most of it is mechanical enough that I have handed chunks of the analysis to AI while keeping the judgment calls for myself. The routine checks that would have caught a lot of this earlier live in my monthly SEO maintenance checklist, and if you are still publishing while you sort this out, the pre-publish checklist is the version that stops you adding to the pile.

I am one person in Augsburg working mostly with US clients, so I see a small sample, not the whole index. But the pattern in that sample is consistent: the sites that came through August fine were not the ones with the best links. They were the ones that never had a programmatic content problem to begin with.

If you have gone through the checklist and you still cannot tell what happened, that is a reasonable thing to want a second pair of eyes on. That is roughly what an audit is, and you can find me at wunderlandmedia.com.