If your traffic dropped in late August, test it against this update first, and then against the two other things that happened the same month. Google's third spam update of 2026 came and went in under three days. The August 2026 spam update started August 18 and completed August 21, globally and in all languages, according to the Search Status Dashboard, and two days and sixteen hours is unusually brisk for a ranking update.
Brisk sounds gentle. Then the measurement came in and showed this one cut much deeper than it looked. What follows is how we'd work out whether it cut you.
What the update was, and what it wasn't
August was thick with update rumors, so it's worth being precise. This was a spam update, meaning targeted enforcement of Google's published spam policies. Those cover scaled content abuse (mass-produced AI filler included), site reputation abuse and expired domain abuse, along with link spam. It was not a core update. No core update was confirmed in August, or in September so far, and the August 25 helpful content update that turned up in some coverage never got dashboard corroboration. We'd treat it as noise.
The part we'd actually pay attention to came ten days after the rollout finished, when Google added a new spam policy section for back button hijacking, the trick where a site traps your back button to re-serve ads. Policy additions are the closest thing Google gives you to a roadmap. This one says enforcement is moving past content spam and into manipulative site behavior.
How hard it actually hit
The SE Ranking data shared with Search Engine Land is the clearest severity measurement anyone has published. During the update window, 16.71% of URLs that had been ranking in Google's top 10 fell beyond position 100 for the same keyword, against 9.2% in a comparable July baseline. That's an 82% increase in top-10 pages effectively vanishing, which made a top-10 URL about 1.8 times more likely to disappear from the top 100 than in a normal week.
Two details matter when you read your own charts. The volatility reached every one of the 20 industries tracked, with top-10 movement running from 74.64% in real estate to 85.55% in fashion and beauty, so no niche gets to assume it was skipped. And SE Ranking ran no URL-level analysis of which tactics got hit. Nobody can honestly tell you "it targeted X," and anyone selling you a confident list of what this update punished is extrapolating.
We have a hypothesis about the niche spread, clearly labeled as one. YMYL categories like real estate and healthcare sat among the steadiest, while affiliate-heavy consumer niches churned hardest, which is the opposite of what update folklore predicts.
The least exotic explanation is that years of aggressive quality enforcement left YMYL with less spam standing to remove. If that's right, the update's energy went where the easy-money content lives. For a B2B SaaS team, that reads as a warning label on thin programmatic pages and a mild all-clear for product content that's actually useful.
How often does ChatGPT mention your brand?
Most founders have no idea. The answer might surprise you.
The diagnosis: five checks in order
We'd never blame this update on a hunch, because a late-August drop had at least three plausible causes this year. The same month brought the spam update, expanding AI Overviews, and AI Mode default tests. So run these checks in sequence before you conclude anything.
Start by lining up the dates. Pull daily clicks for the pages that dropped and find the inflection point. A spam-update hit starts sharply, somewhere between August 18 and 21, and a drift that began earlier or slid gradually through the month points elsewhere.
The second check does the most diagnostic work: separate positions from clicks. In Search Console, compare average position and CTR for the declining queries. If positions went down, you're looking at a ranking event and this update is your suspect. If positions held while CTR sagged, the click died above you. That's the AI Overview squeeze, and no spam remediation will fix it.
Next, map the loss pattern. Update hits usually cluster in a content type, a template or a directory of programmatic pages, so losses spread evenly across unrelated pages suggest measurement or seasonality before they suggest a penalty.
Only then audit against the actual policies, reading your losing pages the way the policy reads them. Look for mass-generated content with no added value, guest sections rented out for rankings, and links you paid for in any currency. Be harder on yourself than your agency will be, because the scaled content policy doesn't care that the filler was expensive.
And before you rewrite a single page, rule out the boring explanations: tracking changes, cookie banners, a migration, a robots edit. Every quarter, some percentage of "the update hit us" turns out to be a tag manager deploy.
If the drop is in visits from ChatGPT, Copilot or DuckDuckGo rather than Google, the order changes. Our AI search outages checklist covers status pages, the Bing dependency and crawler blocks.
A worked example of the positions-versus-clicks read
Check two is where most diagnoses get made, so here it is with numbers. Say your comparison page averaged position 4 in July, with a 6% CTR and 900 monthly clicks. September's report shows position 4.2, a CTR of 3.1% and clicks around 460. The position held while the click-through halved, so that page wasn't hit by the spam update. It's being answered above the fold, and the way back is earning the citation inside the answer rather than a content audit.
The other shape looks nothing like that. Position 4 becomes position 38, the CTR looks normal for the new position, and clicks collapse in proportion. That's a ranking event dated to the update window, and the spam-policy audit from check four is your next hour. We'd run this read per page cluster, because a real site usually shows both shapes at once, and averaging them produces a diagnosis that fits neither.
15 hours a week manually. Or 15 minutes with RankControl.
Track citations, monitor competitors, and fix content gaps across every AI search engine. Automatically.
If the update hit you
We wish the recovery news were more exciting. Algorithmic spam actions carry no reconsideration request, so there's no form and nobody to appeal to. Recovery means removing or properly rebuilding whatever tripped the policy, then demonstrating quality while Google re-evaluates, and the visible movement typically arrives at later updates rather than next Tuesday. The trap is remediating with a trick. Swapping one shortcut for a newer shortcut is how sites become permanent update casualties.
This update's scaled-content enforcement revived the AI content question, so a quick reality check. Google's position has stayed consistent: AI-assisted content is fine, and mass-produced low-value content isn't, whoever produced it. What gets judged is whether each page would justify its existence to a human reader, whatever tool wrote it, and every update applies that test more literally.
The bigger pattern worth acting on
Step back one level and the August spam update stops looking like a Google story. In the same fortnight, ChatGPT's GPT-5.6 rollout reshuffled its citation sources away from forums and toward official domains. Two unrelated engines moved in one direction: they demoted low-trust content and consolidated on sources they can verify. We think that's the defining pattern of this quarter, and it means quality remediation now pays off on two scoreboards at once.
It also means diagnosis is turning into a permanent discipline instead of a scramble after each update. Position tracking catches Google's moves, but it says nothing about AI engines that reshuffled in the same week with no dashboard incident to point at. Teams that track citations per engine weekly had both stories in their data within days, dated and separable. Do the five checks once for August, then decide whether you want to run them from memory again in November, because on the current cadence something will make you.
AI search traffic grew 835% this year. Is your content ready?
RankControl generates 26 content formats optimized for ChatGPT, Claude, and Perplexity. Published on your domain, matched to your brand.
