Google’s August 2026 spam update: what changed in the EEA?

Google’s August 2026 spam update began on August 18 and completed on August 21; the global, all-language rollout lasted 2 days and 16 hours. Southern, Search Engine Journal (2026)

Released the August 2026 spam update, which applies globally and to all languages. The rollout may take a few days to complete.

— Google Search Status Dashboard, quoted by Search Engine Roundtable (2026)

This was the third spam update of 2026. The previous ones landed in March and June; the March update completed in 19 hours and 30 minutes, while the June update took 2 days and 1 hour. The August rollout followed the same recent pattern of short, multi-day releases. Southern, Search Engine Journal (2026)

Google announced no new spam-policy type or companion blog post with the rollout. The August 28 site reputation policy announcement is separate: it changes the regional effect of manual actions rather than describing the algorithmic rollout. Google Search Central (2026)

Google doesn't publicly list which techniques a given spam update targets. Barry Schwartz of Search Engine Roundtable wrote that this update does not target link spam or the site reputation abuse policy; that detail does not appear in Google's release note (Schwartz, Search Engine Roundtable (2026)). It should therefore be read as independent industry reporting rather than an official Google statement.

That leaves the rest of Google's spam policies. One relevant heading to assess in cosmetics and ecommerce is scaled content abuse.

Scaled content abuse is when many pages are generated for the primary purpose of manipulating search rankings and not helping users.

— Google Search Central, Spam Policies (2026)

The first example Google gives under that heading is immediately familiar: "Using generative AI tools or other similar tools to generate many pages without adding value for users" (Google Search Central (2026)). The test isn't how a page was produced, but whether it adds value for the reader.

A beauty brand's direct risk depends on whether its pages are produced at scale primarily to manipulate rankings without adding value for users; having product and category pages is not, by itself, evidence of low risk. Google Search Central (2026)

Under Google's scaled-content-abuse criteria, two examples that may raise risk are distributing AI-written product descriptions across hundreds of variant pages on the same skeleton and generating a separate, near-duplicate indexable page for every shade or size. In both cases, the assessment depends on whether the pages are produced primarily to manipulate rankings without adding value for users. Google Search Central (2026)

Google's test is not page count or the presence of particular elements; it is whether pages are produced primarily to manipulate rankings without adding value for users. Google Search Central (2026) In Soleach's editorial assessment, a shade-specific INCI list, skin-tone imagery and usage notes can be signs of added user value; changing only the product name does not establish such value by itself. In our audit of 216 beauty brand websites we found most brands left their content structure at the platform default — which is also where this problem tends to accumulate.

Even when a cosmetics brand's own site does not change, the visibility of a third-party page describing it—such as a "top 10 dark spot creams" or "which serum actually works" article—can change. For an example of such a page being selected as a source in an AI summary, see the AI answer citing a listicle instead of your site.

Within Google's spam policies, the heading covering this kind of publishing is site reputation abuse.

Site reputation abuse is a tactic where third-party content is published on a host site mainly because of that host's already-established ranking signals, which it has earned primarily from its first-party content.

— Google Search Central, Spam Policies (2026)

Per Schwartz's reporting, the August update doesn't target that particular policy (Schwartz, Search Engine Roundtable (2026)). Those publishers' "best X" lists may still meet Google's scaled-content-abuse criteria if many pages are produced primarily to manipulate rankings without adding value for users. Google Search Central (2026)

The practical consequence (our inference): your referral and brand-search traffic can fall while nothing changes on your site at all. What dropped isn't you — it's the page that describes you. That's why the data to review after 18 August isn't only Search Console, but your referral sources.

From August 30, 2026, the effect of a site reputation policy manual action depends on the searcher’s region: it directly affects the relevant site section in results outside the EEA, while its impact does not apply to results inside the EEA. Google Search Central (2026)

Google also says that, inside the EEA, the affected section may be separated in its systems and may rank independently from the rest of the site over time. Saying ‘there is no penalty in the EEA’ is therefore incomplete: the regional manual-action impact does not apply, but the third-party section may no longer benefit from the main domain in the same way.

For searchers outside the EEA, the previous approach continues and the manual action directly affects only the relevant section; the rest of the site should remain unaffected. A beauty brand selling globally should therefore examine country or region segments rather than one average ranking chart.

No. Google’s site reputation policy remains in force; what changes is how a manual action under that policy affects Search results in the EEA. Google’s 2024 explanation also says the policy does not cover all third-party content, only content hosted without close oversight and intended to manipulate Search rankings. Google Search Central (2024)

An editorially reviewed product test on a beauty publication is not the same as a series of coupon pages created only to benefit from the publisher’s authority. The assessment should examine who owns the content, how much oversight the publisher exercises, and whether the page is intended to manipulate Search rankings.

This update changes the answer in our existing August 2026 spam guide by adding the EEA distinction. To connect policy work with visibility, apply the content-ownership and source-integrity steps in our GEO guide.

Beauty brands should first inventory every page published on third-party domains on their behalf, then document editorial oversight, content ownership, and linking arrangements. Search Console’s Manual Actions report shows both current actions and the site’s manual-action history. Google Search Console Help (2026)

• List sponsored product guides, coupon directories, marketplace subpages, and advertorial content by domain and accountable team.
• Keep contracts, approval flows, and revision logs showing whether the publisher applies a real review and correction process.
• Do not leave Search Console alerts only with technical SEO; name a shared response owner across legal, content, ecommerce, and market teams.
• Monitor visibility inside and outside the EEA separately; do not interpret one country’s data as a global result.

Google says site owners will continue to be notified in Search Console when a manual action is applied. After remediation, an action believed to be wrong can be challenged through a reconsideration request; eligible sites may also have access to mediation. Google Search Central (2026)

International SEO reporting should separate visibility inside and outside the EEA for the same page, because Google explicitly says a page may have a manual action while the resulting change affects only results shown to users outside the EEA. Google Search Central (2026)

The first comparison layer should include country-level impressions, clicks, average position, and the affected URL directory. The second should add publisher domain, content type, and campaign owner. This prevents relative visibility inside the EEA and manual-action impact outside it from disappearing into one global average.

Do not label a sudden rank-tracking difference as a policy effect by default; compare Search Console, analytics, and country-level SERP samples. For another measurement-layer change, see our google.com/goto SEO measurement guide.

The May 15, 2026 clarification matters because Google explicitly confirmed that its spam policies also cover attempts to manipulate generative AI responses in Search. Schwartz, Search Engine Land (2026) The current definition reads:

In the context of Google Search, spam refers to techniques used to deceive users or manipulate our Search systems into featuring content prominently, such as attempting to manipulate Search systems into ranking content highly or attempting to manipulate generative AI responses in Google Search.

— Google Search Central, Spam Policies (2026)

Before the change, the sentence stopped at "ranking content highly"; the clause about manipulating generative AI responses was added in May (Schwartz, Search Engine Land (2026)). Aggressive tactics aimed at appearing in AI Overviews and AI Mode are now explicitly within policy scope.

The June spam update followed that clarification; Google hasn't said whether the August update specifically targets that behaviour (Southern, Search Engine Journal (2026)). That's an open question rather than a claimed link — but an open question that shapes tactic selection for any brand running GEO work.

Google's traffic-drop debugging guide advises against reading a single day's data and recommends ruling out data anomalies first (Google Search Central (2025)). The order should be:

• Check the Search Console Data Anomalies page first — the drop may stem from a change in data processing or a logging error (Google Search Central (2025)).
• Extend the Performance report date range to 16 months; seasonality is strong in cosmetics, and the drop may line up with the same period last year.
• Sort the Pages table by click difference: is the drop site-wide or confined to a group of pages? Those point to different problems.
• Open the Manual Actions report. If an action exists, investigate the issue Google identified; if none exists, examine algorithmic updates alongside technical issues, security issues, seasonality, and data anomalies (Google Search Central (2025)).
• Look beyond Search Console: check which referring sites lost traffic. The health of publishers sending you traffic is independent of the health of your own site.

There's also a measurement trap in this window. Google logged on its data anomalies page that impressions in the Generative AI performance report in Search have been dropping since 13 August 2026, and stated the issue affects data logging only. John Mueller said it is "just a logging issue and not representative of visibility changes in Search" (Schwartz, Search Engine Roundtable (2026)). Don't confuse that drop with the spam update — the dates sit next to each other but the causes are separate.

Google's spam updates documentation frames recovery as a learning process rather than a timeline: making changes may help a site improve if its automated systems learn over a period of months that the site complies with the spam policies (Google Search Central (2025)).

The same document notes one exception: with link spam updates, making changes might not produce an improvement, because once the systems remove the effect of spammy links, any ranking benefit those links previously generated is lost (Google Search Central (2025)).

The practical reading for a beauty brand: don't build your September season campaigns on the assumption that organic data will bounce back quickly after 18 August. While organic is uncertain, shifting demand capture toward paid isn't a change of strategy — it's a temporary rebalancing.