The September 2026 Spam Update: What’s Confirmed So Far

Google’s own ranking-updates dashboard lists a new entry: the September 2026 spam update, started 24 September 2026. It’s a spam update, not a core update, which matters because the two are officially distinct and get treated differently in how sites recover from them. As of this writing it’s about a day into rolling out, and Google’s own wording says the rollout may take up to two weeks to complete.

What Google actually said

The statement, relayed by trade press covering the status dashboard, reads: “Released the September 2026 spam update, which applies globally and to all languages. The rollout may take up to two weeks to complete… This is a normal spam update, and it will roll out for all languages and locations.” That’s the whole of it. No named policy, no named tactic, no niche called out. That’s standard practice for spam update announcements: Google’s own spam policies documentation describes the categories it enforces against in general terms, but individual update announcements have never confirmed which of those categories a specific update is enforcing, and I wouldn’t expect that to change here.

There’s no dedicated Search Central blog post about it yet, and nothing has landed in the official documentation changelog either. The dashboard entry and Google’s Search Central account are, so far, the only official channels carrying it.

Why the two-week window is the part worth noticing

Compare it with the other spam updates Google ran this year. March 2026 finished in 19.5 hours. June 2026 took 2 days 1 hour. August 2026 took 2 days 16 hours. A rollout window measured in weeks rather than hours or days is new for a 2026 spam update, and it’s the single most concrete fact in this whole story: whatever this update is enforcing, Google expects it to take meaningfully longer to fully apply than anything it shipped earlier this year.

That has a practical consequence. A ranking change on day one or two of a multi-week rollout isn’t the final picture, and neither is the absence of one. Judging this update against day-one data is judging it too early either way, in both directions: a site that drops early could still recover before the rollout finishes, and a site that looks untouched now isn’t necessarily clear.

What isn’t known yet, and shouldn’t be claimed

This update is roughly a day old. At that stage, there’s essentially no accumulated impact data: no confirmed niche breakdown, no volatility tracker readings tied specifically to this rollout, and no on-the-record reaction yet from the practitioners who usually publish detailed case studies after Google’s updates. Any post claiming to already know which niches this one hit hardest is either talking about a different update or guessing. The site reputation abuse policy change from August is a useful example of how long it actually takes for the real picture to become clear: the detailed analysis came weeks after the announcement, not days.

What to actually do while it rolls out

• Don’t make reactive changes based on a single day’s ranking movement. Google itself is telling you the rollout isn’t finished.
• Check Search Console’s Manual Actions report regardless of what you see in rankings; a spam update and a manual action are different mechanisms, and it’s worth ruling one out while you’re looking.
• If you’ve had quality problems flagged before, this is a reasonable moment to revisit them, since spam updates tend to reinforce standing quality signals rather than introduce new ones from scratch.
• Keep your own log of what you see day to day. With a rollout this long, the sites best placed to understand what happened to them will be the ones with their own before-and-after data, not ones relying on other people’s case studies written about a different site.

I’ll follow this one as more of the rollout completes. For now, the honest position is: it’s real, it’s a spam update, it started 24 September, and it’s going to take longer than anything Google has shipped this year to finish rolling out. Everything past that is still unconfirmed, and worth treating as such until it isn’t.