Google began rolling out its September 2026 spam update on Sept. 24. The rollout started at 9:15 a.m. PDT. It runs up to two weeks — the longest window Google has given any spam update this year. The update is the fourth of 2026, after March, June, and August.

Prior updates each carried "a few days" windows. None took longer than two days and 16 hours. Google has not said what this update targets. It applies globally and across all languages, per the Search Status Dashboard.

What changes for sites this time?

The longer rollout means early volatility won't settle quickly. Ranking swings in the first few days may not reflect the full effect. Sites that see traffic shifts should check their dates against Google's window. Google says affected sites should review its spam policies.

The takeaway: wait for the rollout to finish before acting on early data.

How does the new Search Console filter work?

Google is adding a multimodal filter to the Performance report. It tracks clicks on pages reached through image-based queries. Covered inputs include Google Lens, Circle to Search on Android, image uploads, and Chrome's right-click "Search this image" feature. The rollout is global. Query data is missing — multimodal searches mostly use images, not text.

Ecommerce and image-heavy publishers gain page-level multimodal traffic split by country and device, but they still miss the queries that triggered each click. A product page with multimodal clicks won't show which image the person used.

The takeaway: you can see where people land, not what they searched with.

Why did pages vanish over a casino canonical?

Google's John Mueller responded to a Reddit thread about pages dropping from Google's index. The site owner said Google treated a casino betting page as the canonical. The wrong canonical affected pages about companies and suppliers. Another Redditor found a third-party canonical indexed under a generic JavaScript error title.

Mueller said Googlebot may have crawled an error page. He pointed to Search Console's live URL test as a diagnostic. In a second reply, he said each possible outcome leaves the page "not showing in search for normal content." He runs automated tests before pushing his own sites live. He suggested fetching critical pages on a schedule "so that you can fix them before they become stable problems for search engines to pick up."

Mueller treats prevention as cheaper than diagnosis, since once a broken canonical stabilizes in Google's index, untangling the cross-domain mess becomes a far longer cleanup job for any publisher caught in the middle.

The takeaway: catch a broken page before the crawler does.

How steep did publisher search traffic fall?

Chartbeat's 2026 Publisher Playbook reports Google Search fell 40.2% year over year. The data compares July 2025 to July 2026. The previous year saw a 21.9% decline. Google Discover's drop widened from 6.6% to 34.3%. The playbook does not specify the measurement basis — pageviews or each channel's share of pageviews.

The network skews toward news and media publishers. Direct, internal, and dark social traffic grew their share over the same period. Chartbeat frames the growing channels as ones media companies have "greater control over," a soft argument for audiences that don't depend on Google's algorithm.

The takeaway: your July comparison matters more than the network average.

What did the AI Mode click experiment find?

University of Pennsylvania and Northeastern University researchers ran a field experiment. They tracked 1,100 U.S. participants through a three-day baseline and seven days of treatment. They used a browser extension to record search behavior.

People assigned to AI Mode clicked through to other websites 18.8 percentage points less often than people using regular Google Search. They also gave lower trust and satisfaction ratings. AI Mode handled 0.6% of participants' baseline searches. That low baseline led the authors to call the setup a forced, full-adoption scenario.

Google's Liz Reid wrote in August 2025 that organic click volume from Search to websites had been "relatively stable" year over year. This experiment measured a different case — searches redirected to AI Mode. The 18.8-point drop shows what full adoption looked like in this sample.

It's a controlled forced-adoption test, not a forecast for what happens when real users mix AI Mode with regular Search in natural proportions.

The takeaway: forced AI Mode use cuts clicks hard in this sample, but the conditions don't match real-world adoption.

What should you watch next?

The spam update's two-week window closes around Oct. 8 if it runs the full term. Search Console multimodal data will keep accumulating; expect query coverage to remain absent. Chartbeat's playbook leaves measurement scope unclear, so watch for clarification on pageviews vs. share. The AI Mode experiment is a preprint, not a published study, so peer review could shift its claims. As Search Console splits traffic into more views, more of the explaining lands in the records you keep.