Google's September 2026 Spam Update Has Finished: Was Your Site Hit, and What Now?

Google's September 2026 spam update finished on Thursday, 8 October. Today is Sunday, 11 October. Two full days have passed since then, a Friday and a Saturday. The agency's verdict has arrived anyway: "Google updated its algorithm. That's why traffic is down." Two days can't carry a verdict in either direction.

Nobody can tell you yet what it did to your site. All you can take today is a first reading; the verdict comes around 25 October. Until then, fill in one sheet and change nothing on the site.

Valid as of 11 October 2026.

What did Google announce about the September 2026 spam update, and when did it finish?

The update began on 24 September 2026 and finished on 8 October. On 24 September at 09:15 Pacific Daylight Time (PDT), Google wrote on the incident page of its Google Search Status Dashboard: "Released the September 2026 spam update, which applies globally and to all languages." The same entry added: "The rollout may take up to two weeks to complete." On 8 October at 01:37 PDT the same page added: "The rollout was complete as of October 8, 2026."

What matters to you is how long it ran, because those are the days whose numbers you can't read. The dashboard's history page records 13 days and 16 hours for this one. The March spam update took 19 hours and 30 minutes, June's took 2 days and 1 hour, and August's 2 days and 16 hours. That makes this the longest of the four spam updates in 2026, which is why the "Do not judge" window in the sheet below runs two weeks.

What did the update target? What Google has not said

Google has not said what the update targeted. The two announcements leave three questions open: which policies or tactics the update acted on, why it ran for two weeks, and whether it had any measured effect on Arabic-language results.

So any sentence that starts "the new Google update targeted..." is the speaker's opinion. They aren't quoting Google.

Does it apply to Arabic sites and stores in Saudi Arabia and Jordan?

In scope, yes: the update "applies globally and to all languages," in the words of Google's 24 September announcement. Effect is a separate question. Nobody has published a measurement specific to Saudi Arabia, Jordan or Arabic: not Google, and no source I trust.

My guess, and it is only a guess, is that most stores and companies here weren't touched. A store that sells its own products, writes its own pages and never bought links gives these systems nothing to detect. Those with a reason to worry usually know it already: hundreds of pages generated in one batch, links bought in bulk, or a cheap SEO package where nobody asked what was actually done.

How do you tell from Search Console whether your site was hit? The before-and-after sheet

Compare two equal periods that both sit outside the rollout days, one page group at a time. A page group is a set of pages of one type whose numbers are read together: category pages, product pages, articles. The site total hides this: articles can lose half their clicks while sales hold, while categories can lose a tenth and sales feel it.

In Search Console, open Performance and check that the search type filter reads Web. Click the date filter above the chart, choose the Compare tab and set the two periods to custom dates: 9–22 October against 10–23 September. If your account doesn't offer custom dates there, read each period on its own and write its number down. Copy the clicks total for each period. To isolate a group, add a page filter ("URLs containing") or a query filter ("Queries containing"). You work out the difference yourself: (after − baseline) ÷ baseline × 100. Once you filter, the group rows won't add up to the site total. That's fine as long as each group is compared with itself.

Page group	How to isolate it	Baseline clicks: 10–23 September	Rollout days: 24 September–8 October	After clicks: 9–22 October	Difference %	Seasonal event in the baseline?	Organic orders or leads (before → after)	Decision
Brand-name queries	Query filter, Custom (regex) option: the name in Arabic, then |, then the name in Latin script, in one box		Do not judge
Category pages	Page filter: the part that repeats in every category URL		Do not judge
Product pages	Page filter: the part that repeats in every product URL		Do not judge
Articles and blog	Page filter: the blog path		Do not judge

Open two products and two categories and look at the URLs. If there is a fixed folder such as /products/, type it in. If there isn't, ask your agency or developer for a Custom (regex) expression that matches your URL pattern.

I call it the before-and-after sheet, and it gets read on three dates. Each date trails the period it reads, because Search Console's help page on data discrepancies says: "Collected data is usually available in 2–3 days."

1. Today, 11 October. Fill in the whole baseline column. The "after" column may show only 9 October today; 10 October may be partial or missing, and the report doesn't keep Riyadh or Amman time. The same help page says: "In all views (except the 24-hour view), the Performance report tracks and labels daily data in Pacific Time (PT)." 9 and 10 October are weekend days in Saudi Arabia and Jordan, so if you compare at all, compare with the baseline's Fridays and Saturdays (11–12 and 18–19 September), not with a 14-day average.
2. 18 October. One full week, 9–15 October, against 17–23 September. Still a note, not a decision.
3. 25 October. Two weeks against two weeks. Before reading, look at the last day shown on the chart. If 22 October isn't there, the period is short and you'll see a drop that doesn't exist. Wait a day or two, then write the decision column.

The seasonal column is for Saudi National Day, 23 September. Retailers' offers for it usually start days earlier, so a Saudi store's baseline clicks may be higher than usual. If you write "yes", run the same comparison for last year: 10–23 September 2025 against 9–22 October 2025. If clicks fell by a similar share then, what you're seeing is the season. First check on the Search Status Dashboard whether an update was rolling out in those days of 2025, because that weakens last year's comparison. A Jordanian store that doesn't sell in Saudi Arabia writes "no" and skips this.

To tell whether a difference is large, use your own site as the yardstick. Take the four weeks before the rollout: 27 August–2 September, 3–9, 10–16 and 17–23 September. The largest gap between one week and the next is that group's normal variation. A Saudi store leaves the last week out, because National Day is in it. That yardstick is my rule, not Google's.

For the orders column, a store looks at orders and revenue from organic search; how to read them in GA4 is covered in the technical SEO guide for online stores. Fill it in once for the whole site unless you already use GA4's landing-page report, and leave the brand-query row blank. Don't match the two reports day by day, because they may date days in different time zones; compare total with total. A company that sells through leads asks its sales team how many quote requests and WhatsApp conversations came from search in each period. If nobody asks customers where they came from, that problem was there before any update.

Then the decision, one of three:

• Ignore. The difference sits inside normal variation and orders haven't moved.
• Wait. Clicks fell and the reading is earlier than 25 October; or there's a seasonal event in the baseline; or clicks fell and orders are steady.
• Act. In the 25 October reading the drop is still there in a group other than brand-name queries, it's larger than normal variation, organic orders or leads fell with it, and you've ruled out a technical cause, a manual action and the season.

The same holds if the numbers went up: credit nobody before 25 October.

What should you do this week if clicks fell, and what if nothing moved?

If clicks fell, this week is for ruling out the other causes Google's debugging guide lists: technical issues, security issues, spam issues (where it tells you to check the Manual Actions report), seasonality and changing interests, and site moves, including URL changes.

Ask your team what changed on the site since 10 September: a new theme, categories rearranged, pages deleted after a season. Run the 30-minute Search Console check to rule out indexing. Then open Security & Manual Actions in Search Console and look at both reports, as Google's debugging guide asks. The distinction, as I'd put it: a manual action is a human reviewer's decision and shows in that report, while an update to the automated systems leaves nothing there.

If you have ruled those out and the drop is still there in the 25 October reading, then Google's own guidance applies. Its page "Google Search spam updates and your site" on Google Search Central, last updated 10 December 2025 (UTC), says: "Sites that see a change after a spam update should review our spam policies to ensure they are complying with those." On timing, the same page says: "Making changes may help a site improve if our automated systems learn over a period of months that the site complies." The unit is months, so don't believe anyone who promises you two weeks.

The policies page, last updated 28 August 2026 (UTC), is a list you can read in one sitting: Scaled content abuse, Link spam, Doorway abuse, Keyword stuffing, Thin affiliation, and others. It tells you where to look, not what this update targeted. One thing you can do this week without touching the site: ask the agency for a written list of every link it bought and every page it generated.

If nothing moved, the update gives you nothing to do. Carry on with what you were working on.

What should you ignore after the update?

Ignore anything that asks you for a decision before the sheet is complete. Three things in particular:

1. Volatility charts. They typically track a sample of queries the tool vendor picked, and that sample may not include a single query your customers type.
2. "Ranking recovery" offers. Someone who doesn't know what the update targeted can't price the cure. What I wrote about the offers to refuse in AI search applies here unchanged.
3. Disavowing links and deleting pages in a hurry. You don't yet know whether those pages or links were the cause, and a deletion isn't undone with one click. Google has not said this was a link spam update, and even for those, the same spam updates page says: "In the case of a link spam update (an update that specifically deals with link spam), making changes might not generate an improvement." Two exceptions: a manual action over links gets handled the way that report says, and hacked pages listed under Security Issues get cleaned today.

Why I read orders before rankings

Because a ranking screenshot doesn't pay a salary. My habit is to read the sales sheet before any report, and the orders column is that habit applied to search. Put its figure in the weekly revenue sheet next to the ad numbers, and you can see how much of your revenue organic search accounts for.

Reading rankings alone makes people rush the repair. My view: the worst response to this update is to change 10 things in a week nobody can measure. If the numbers improve afterwards, you won't know which change helped. If they get worse, you won't know whether it was the update or your own changes. And White Friday, the region's late-November sales season, falls on 27 November, 47 days away. A site taken apart now goes into the season still being put back together.

At Argan Package in Saudi Arabia, where I was Digital Marketing Manager, organic sales rose 15% within 4 months after Search Console and technical SEO issues were fixed. That result has nothing to do with this update. I mention it because it was measured in sales over months, not in one keyword's position over a week. The details are in the Merchant Center and SEO case study.

Today: the sheet, two dates, and one line to the agency

Fill in the baseline column today, with the orders or leads from organic search beside it. Put two dates in the calendar: 18 October for the one-week reading and 25 October for the decision. And answer the agency's message with one line: send the comparison for each page group on these dates, once the last day of the period shows in the report, with the orders.

If 25 October leaves you with an "act" decision and you want someone to read Search Console and the site with you and rank what to review first, that's what my technical SEO audit is for. Request a consultation, and send the sheet filled in.

If you want your team to do this reading without waiting for the agency, see corporate training programs.