Google's September 2026 spam update began on September 24 and finished on October 8. It applied globally and across all languages. For a site that lost traffic during that period, the timing is a reason to investigate, but it does not establish that the update caused the decline or identify a particular policy violation.

The August spam update ran from August 18 to 21, putting two rollouts roughly five weeks apart. A monthly comparison can easily combine their effects with a migration, a reporting change or weaker demand. Start with a dated record of what happened, then test the explanations against the affected pages.

Establish the timeline before choosing a remedy

The official September incident record (opens in a new tab) gives a start of September 24 at 09:15 and an end of October 8 at 01:00, US/Pacific. Google initially estimated up to two weeks. Its history page summarizes the duration as 13 days and 16 hours; the incident timestamps produce an exact interval of 13 days, 15 hours and 45 minutes.

That distinction matters mainly when aligning detailed records. The completion message was posted at 01:37 PDT, later than the recorded end. Do not use the message timestamp as a separate ranking event. For everyday diagnosis, marking September 24 through October 8 is usually more useful than implying minute-level certainty about when a particular site changed.

Google's history also lists spam updates beginning March 24 and June 24, lasting 19 hours and 30 minutes, and two days and one hour respectively. These are separate events, not evidence of continuous punishment. As of October 10, only two days have passed since September's completion, so there is little post-rollout evidence to evaluate.

Record what changed on the website and in measurement

Create a timeline with three categories: confirmed Google rollouts, website changes and measurement changes. Add the first sustained decline, rather than selecting the lowest single day. Include template releases, URL changes, consent updates and tracking deployments that could affect the comparison.

Suppose, as a hypothetical example, an analytics decline starts immediately after a consent-platform change while Search Console clicks remain broadly stable. That pattern justifies checking measurement first. It does not prove rankings were unaffected, but it gives the team a specific explanation to test before rewriting content.

Search Console and on-site analytics measure different events, so their totals need not match. Compare direction and timing while preserving the distinction between search clicks and recorded sessions. Also annotate any change in report coverage, including newly available search-type data, rather than silently treating every reporting period as equivalent.

A useful SEO audit should preserve these competing explanations until evidence rules them out. Labeling the problem “spam update damage” at the beginning can steer every later decision toward the wrong remedy.

Find the affected pages and markets

Break down the decline by page, country, device and search type. Then create your own page-type groups, such as product pages, location pages or editorial guides. Page type is an analytical grouping you define; it is not a native Search Console dimension.

Look at both absolute losses and percentages. A small page can lose most of its clicks without explaining much of the business impact. Conversely, a modest decline on a major landing page can account for a large part of the total. Prioritize the groups that materially change the picture.

For an Israeli company targeting overseas customers, the relevant country is where the search originated, not where the marketing team sits. A decline confined to one market deserves a different investigation from a similar decline across every language and country. Examine local demand, availability and the pages serving that market.

Compare equivalent weekdays and, where useful, the previous year's pattern. Israeli holidays may affect local demand, but they should not be used as a default explanation for an English-language campaign abroad. Treat seasonality as a hypothesis supported by the market's data.

Rule out technical failures and separate manual actions

Check Search Console's Manual Actions and Security Issues reports. A manual action is different from an automated ranking change. An empty Manual Actions report does not establish that every page is healthy, and a reconsideration request is not a general appeal against algorithmic traffic loss.

Next, inspect representative affected URLs alongside unaffected ones. A focused technical SEO investigation can examine server responses, accidental indexing restrictions, canonical selection and redirects. Compare the current behavior with the deployment log, rather than assuming the website remained unchanged during the rollout.

If one template declined, inspect what its pages share. If the whole site declined, check shared infrastructure and broader explanations. These patterns guide the investigation; neither pattern independently proves the cause. Fix a confirmed outage or accidental restriction promptly instead of waiting for more ranking data.

Preserve copies of the evidence before making changes. A screenshot, response sample or export is more useful than a later recollection that “something looked wrong.” It also lets developers and content editors work from the same problem statement.

Review actual spam policies, not guessed update targets

The September announcement did not identify a particular tactic, industry or content format as its target. Do not turn a traffic chart into a claim that Google targeted all AI-assisted content, affiliate sites or a specific business model.

Review publishing practices against the actual policies. Ask whether affected page sets offer useful original information, whether pages were created mainly to manipulate rankings, and whether links or third-party content raise specific concerns. Record the evidence behind each concern. A policy label should follow the facts, not replace them.

Backlink review is appropriate when there is a concrete reason, such as known link schemes or an unnatural-links manual action. Review the site's link-building practices and available records before acting on a third-party toxicity score.

Google says most sites do not need the disavow tool. Its stated threshold combines a considerable number of spammy, artificial or low-quality links with a manual action caused, or likely to be caused, by those links. Incorrect use can harm performance. A spam-update announcement alone does not meet that threshold.

Make controlled changes and set review dates

Separate urgent fixes from longer-term improvements. Repair confirmed technical defects, then group content or structural work around a clear diagnosis. Changing URLs, templates, copy, internal links and a disavow file simultaneously makes it harder to understand what helped or introduced a new problem.

Keep a record of the affected pages, the reason for each change and the next review date. In SEO consulting, that record should support decisions rather than become a promise of recovery. Google's guidance explains that some changes can take days and others months to have an effect, with no guarantee of a noticeable improvement.

A useful progress report can still be concrete: the indexing error was fixed, a duplicated page set was consolidated, or unsupported claims were corrected. Those are verifiable outcomes. Ranking recovery remains something to observe over time, not a deadline to invent.